{
  "model.audio_tower.layers.5.lconv1d.linear_end.input_max": {
    "atom_offset": 1,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 3,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.920890835113823e-4,
    "track": "int8"
  },
  "model.language_model.layers.15.self_attn.q_proj.weight": {
    "atom_offset": 65540,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0034910186659544706,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_start.input_min": {
    "atom_offset": 114693,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.mlp.down_proj.weight": {
    "atom_offset": 114695,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004152312874794006,
    "track": "int8"
  },
  "model.language_model.layers.15.post_attention_layernorm.weight": {
    "atom_offset": 409608,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 409657,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.self_attn.k_proj.weight": {
    "atom_offset": 409659,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.652128554880619e-4,
    "track": "int8"
  },
  "model.language_model.layers.22.mlp.down_proj.weight": {
    "atom_offset": 415804,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.00390625,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.down_proj.output_max": {
    "atom_offset": 710717,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.self_attn.k_proj.weight": {
    "atom_offset": 710719,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.786694189533591e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.q_proj.output_max": {
    "atom_offset": 716864,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.o_proj.input_max": {
    "atom_offset": 716866,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.self_attn.v_proj.weight": {
    "atom_offset": 716868,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.6329048927873373e-4,
    "track": "int8"
  },
  "model.language_model.layers.5.post_per_layer_input_norm.weight": {
    "atom_offset": 723013,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.post_attention_layernorm.weight": {
    "atom_offset": 723062,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.q_proj.output_min": {
    "atom_offset": 723087,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.q_proj.input_max": {
    "atom_offset": 723089,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 723091,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 723124,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.self_attn.q_proj.weight": {
    "atom_offset": 723126,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004890501964837313,
    "track": "int8"
  },
  "model.language_model.layers.28.post_per_layer_input_norm.weight": {
    "atom_offset": 772279,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.per_layer_input_gate.weight": {
    "atom_offset": 772328,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 784617,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_end.linear.weight": {
    "atom_offset": 784619,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0011841781670227647,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 801004,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.post_attention_layernorm.weight": {
    "atom_offset": 801006,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.self_attn.v_proj.weight": {
    "atom_offset": 801055,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.6329048927873373e-4,
    "track": "int8"
  },
  "model.language_model.layers.6.layer_scalar": {
    "atom_offset": 807200,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.o_proj.input_max": {
    "atom_offset": 807202,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.v_proj.input_min": {
    "atom_offset": 807204,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.post_feedforward_layernorm.weight": {
    "atom_offset": 807206,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.self_attn.o_proj.weight": {
    "atom_offset": 807231,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0023375984746962786,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.q_proj.linear.weight": {
    "atom_offset": 856384,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0030450294725596905,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.down_proj.output_min": {
    "atom_offset": 865601,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.layer_scalar": {
    "atom_offset": 865603,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.v_proj.input_max": {
    "atom_offset": 865605,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.v_proj.output_max": {
    "atom_offset": 865607,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.o_proj.input_min": {
    "atom_offset": 865609,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.post.input_min": {
    "atom_offset": 865611,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.k_proj.output_max": {
    "atom_offset": 865613,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.self_attn.k_norm.weight": {
    "atom_offset": 865615,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 865624,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.q_proj.input_max": {
    "atom_offset": 865626,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.q_proj.input_min": {
    "atom_offset": 865628,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.mlp.up_proj.weight": {
    "atom_offset": 865630,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004828986246138811,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.v_proj.input_max": {
    "atom_offset": 1160543,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 1160545,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_end.input_min": {
    "atom_offset": 1160547,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.k_proj.output_min": {
    "atom_offset": 1160549,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.self_attn.k_norm.weight": {
    "atom_offset": 1160551,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.up_proj.input_min": {
    "atom_offset": 1160560,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.self_attn.v_proj.weight": {
    "atom_offset": 1160562,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.002245324896648526,
    "track": "int8"
  },
  "model.language_model.layers.21.input_layernorm.weight": {
    "atom_offset": 1166707,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.k_proj.input_max": {
    "atom_offset": 1166756,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.mlp.gate_proj.weight": {
    "atom_offset": 1166758,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0031526819802820683,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 1314215,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.v_proj.output_min": {
    "atom_offset": 1314248,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.v_proj.linear.weight": {
    "atom_offset": 1314250,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002583661349490285,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.input_layernorm.weight": {
    "atom_offset": 1323467,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.self_attn.v_proj.weight": {
    "atom_offset": 1323492,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 0.0028143455274403095,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 1335781,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.per_layer_projection.weight": {
    "atom_offset": 1335814,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.o_proj.output_min": {
    "atom_offset": 1348103,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.o_proj.output_min": {
    "atom_offset": 1348105,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.input_layernorm.weight": {
    "atom_offset": 1348107,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.self_attn.q_norm.weight": {
    "atom_offset": 1348156,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.post_feedforward_layernorm.weight": {
    "atom_offset": 1348165,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 1348214,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.conv_norm.weight": {
    "atom_offset": 1348216,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.norm_pre_attn.weight": {
    "atom_offset": 1348249,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2775590419769287,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.q_norm.weight": {
    "atom_offset": 1348266,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.relative_k_proj.weight": {
    "atom_offset": 1348269,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014456200879067183,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.down_proj.linear.weight": {
    "atom_offset": 1364654,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0014609990175813437,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_end.input_max": {
    "atom_offset": 1401519,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.post.input_min": {
    "atom_offset": 1401521,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.k_proj.input_min": {
    "atom_offset": 1401523,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.k_proj.input_max": {
    "atom_offset": 1401525,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 1401527,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.574126357212663e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 1467064,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.k_norm.weight": {
    "atom_offset": 1467066,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.self_attn.q_proj.weight": {
    "atom_offset": 1467069,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.0038447342813014984,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 1565374,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 4.5752336154691875e-4,
    "track": "int8"
  },
  "model.language_model.layers.27.per_layer_projection.weight": {
    "atom_offset": 1630911,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.self_attn.v_proj.weight": {
    "atom_offset": 1643200,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.002229945966973901,
    "track": "int8"
  },
  "model.language_model.layers.0.mlp.down_proj.weight": {
    "atom_offset": 1649345,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004336860030889511,
    "track": "int8"
  },
  "model.language_model.layers.4.per_layer_input_gate.weight": {
    "atom_offset": 1796802,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.down_proj.output_min": {
    "atom_offset": 1809091,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 1809093,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.mlp.down_proj.weight": {
    "atom_offset": 1809095,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0054133860394358635,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.down_proj.input_max": {
    "atom_offset": 2104008,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.v_proj.output_min": {
    "atom_offset": 2104010,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.mlp.up_proj.weight": {
    "atom_offset": 2104012,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0038447342813014984,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.k_proj.output_max": {
    "atom_offset": 2251469,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.post_feedforward_layernorm.weight": {
    "atom_offset": 2251471,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_start.output_min": {
    "atom_offset": 2251520,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.pre_feedforward_layernorm.weight": {
    "atom_offset": 2251522,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.o_proj.input_max": {
    "atom_offset": 2251571,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.v_proj.input_min": {
    "atom_offset": 2251573,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.k_proj.input_min": {
    "atom_offset": 2251575,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.self_attn.q_proj.weight": {
    "atom_offset": 2251577,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.00346026080660522,
    "track": "int8"
  },
  "model.language_model.layers.12.post_feedforward_layernorm.weight": {
    "atom_offset": 2300730,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.post.output_max": {
    "atom_offset": 2300779,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.self_attn.o_proj.weight": {
    "atom_offset": 2300781,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.007135826628655195,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 2349934,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.535679033026099e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.o_proj.linear.weight": {
    "atom_offset": 2415471,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002798966597765684,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_start.input_max": {
    "atom_offset": 2424688,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.up_proj.input_max": {
    "atom_offset": 2424690,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.mlp.gate_proj.weight": {
    "atom_offset": 2424692,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004828986246138811,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.o_proj.input_max": {
    "atom_offset": 2719605,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.post_per_layer_input_norm.weight": {
    "atom_offset": 2719607,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.input_layernorm.weight": {
    "atom_offset": 2719656,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_end.input_max": {
    "atom_offset": 2719705,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.o_proj.output_min": {
    "atom_offset": 2719707,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.v_proj.input_max": {
    "atom_offset": 2719709,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.q_proj.output_min": {
    "atom_offset": 2719711,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.self_attn.k_proj.weight": {
    "atom_offset": 2719713,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.594457568600774e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 2732002,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.v_proj.output_max": {
    "atom_offset": 2732035,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 2732037,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 2732039,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.per_layer_input_gate.weight": {
    "atom_offset": 2732041,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.v_proj.output_max": {
    "atom_offset": 2744330,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_start.output_min": {
    "atom_offset": 2744332,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.o_proj.linear.weight": {
    "atom_offset": 2744334,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.00239911419339478,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.v_proj.output_min": {
    "atom_offset": 2753551,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.input_layernorm.weight": {
    "atom_offset": 2753553,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.per_layer_model_projection.weight": {
    "atom_offset": 2753578,
    "byte_length": 27525184,
    "num_atoms": 430081,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.self_attn.k_norm.weight": {
    "atom_offset": 3183659,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.input_layernorm.weight": {
    "atom_offset": 3183668,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.down_proj.linear.weight": {
    "atom_offset": 3183717,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0016839936142787337,
    "track": "int8"
  },
  "model.language_model.layers.2.self_attn.k_proj.weight": {
    "atom_offset": 3220582,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0026913140900433064,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.down_proj.input_min": {
    "atom_offset": 3226727,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.input_layernorm.weight": {
    "atom_offset": 3226729,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.input_layernorm.weight": {
    "atom_offset": 3226778,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.self_attn.k_norm.weight": {
    "atom_offset": 3226803,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.self_attn.v_proj.weight": {
    "atom_offset": 3226812,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.7097995411604643e-4,
    "track": "int8"
  },
  "model.language_model.layers.22.post_feedforward_layernorm.weight": {
    "atom_offset": 3239101,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.up_proj.output_min": {
    "atom_offset": 3239150,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.mlp.down_proj.weight": {
    "atom_offset": 3239152,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005505659617483616,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_start.output_max": {
    "atom_offset": 3534065,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.norm_pre_attn.weight": {
    "atom_offset": 3534067,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2303149551153183,
    "track": "int8"
  },
  "model.language_model.layers.23.layer_scalar": {
    "atom_offset": 3534084,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.q_proj.input_max": {
    "atom_offset": 3534086,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.k_proj.linear.weight": {
    "atom_offset": 3534088,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0034756397362798452,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.o_proj.input_min": {
    "atom_offset": 3543305,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.self_attn.o_proj.weight": {
    "atom_offset": 3543307,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.0026759349275380373,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 3641612,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.post_attention_layernorm.weight": {
    "atom_offset": 3641614,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 3641663,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.norm_post_attn.weight": {
    "atom_offset": 3641665,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.17519685626029968,
    "track": "int8"
  },
  "model.language_model.layers.13.layer_scalar": {
    "atom_offset": 3641682,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.post.linear.weight": {
    "atom_offset": 3641684,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.727546664886177e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 3658069,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 3658071,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.o_proj.output_max": {
    "atom_offset": 3658073,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.self_attn.k_proj.weight": {
    "atom_offset": 3658075,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.671352216973901e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.post.input_max": {
    "atom_offset": 3664220,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.gate_proj.output_min": {
    "atom_offset": 3664222,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.post_attention_layernorm.weight": {
    "atom_offset": 3664224,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.self_attn.q_proj.weight": {
    "atom_offset": 3664273,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.003937007859349251,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.q_proj.input_max": {
    "atom_offset": 3762578,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.o_proj.output_max": {
    "atom_offset": 3762580,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.up_proj.input_max": {
    "atom_offset": 3762582,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.post_attention_layernorm.weight": {
    "atom_offset": 3762584,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.post_feedforward_layernorm.weight": {
    "atom_offset": 3762633,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.self_attn.k_norm.weight": {
    "atom_offset": 3762682,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.mlp.gate_proj.weight": {
    "atom_offset": 3762691,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006859005894511938,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.k_proj.linear.weight": {
    "atom_offset": 4057604,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.381520747207105e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_end.output_max": {
    "atom_offset": 4073989,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.k_proj.input_min": {
    "atom_offset": 4073991,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 4073993,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_end.output_max": {
    "atom_offset": 4073995,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.self_attn.o_proj.weight": {
    "atom_offset": 4073997,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.005351869855076075,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.q_proj.output_min": {
    "atom_offset": 4172302,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.self_attn.v_proj.weight": {
    "atom_offset": 4172304,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.479115305002779e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.q_proj.input_min": {
    "atom_offset": 4178449,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.layer_scalar": {
    "atom_offset": 4178451,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.o_proj.input_max": {
    "atom_offset": 4178453,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 4178455,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 4178457,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.0233759842813015,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.q_proj.output_min": {
    "atom_offset": 4178538,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.self_attn.q_norm.weight": {
    "atom_offset": 4178540,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.mlp.up_proj.weight": {
    "atom_offset": 4178549,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.00818159431219101,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 4473462,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.05167322978377342,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.down_proj.linear.weight": {
    "atom_offset": 4473543,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.001876230351626873,
    "track": "int8"
  },
  "model.language_model.layers.32.input_layernorm.weight": {
    "atom_offset": 4510408,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.down_proj.output_min": {
    "atom_offset": 4510457,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 4510459,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 4510461,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.post.output_min": {
    "atom_offset": 4510494,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.post_attention_layernorm.weight": {
    "atom_offset": 4510496,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_start.output_max": {
    "atom_offset": 4510545,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.v_proj.input_max": {
    "atom_offset": 4510547,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.input_layernorm.weight": {
    "atom_offset": 4510549,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 4510598,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.113127456046641e-4,
    "track": "int8"
  },
  "model.language_model.layers.5.mlp.down_proj.weight": {
    "atom_offset": 4576135,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.005105806980282068,
    "track": "int8"
  },
  "model.language_model.layers.14.self_attn.q_norm.weight": {
    "atom_offset": 4723592,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.o_proj.input_min": {
    "atom_offset": 4723609,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 4723611,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.per_layer_input_gate.weight": {
    "atom_offset": 4723613,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.gate_proj.output_max": {
    "atom_offset": 4735902,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.self_attn.v_proj.weight": {
    "atom_offset": 4735904,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.6905758790671825e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.v_proj.output_min": {
    "atom_offset": 4748193,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.gate_proj.input_min": {
    "atom_offset": 4748195,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.per_layer_input_gate.weight": {
    "atom_offset": 4748197,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 4760486,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 9.188915137201548e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.q_norm.weight": {
    "atom_offset": 4826023,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.post_attention_layernorm.weight": {
    "atom_offset": 4826026,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 4826051,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 8.035494829528034e-4,
    "track": "int8"
  },
  "model.language_model.layers.27.input_layernorm.weight": {
    "atom_offset": 4891588,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.down_proj.output_max": {
    "atom_offset": 4891637,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 4891639,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 4891641,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.mlp.gate_proj.weight": {
    "atom_offset": 4891643,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006582185160368681,
    "track": "int8"
  },
  "model.language_model.layers.0.self_attn.q_norm.weight": {
    "atom_offset": 5186556,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 5186565,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_end.output_min": {
    "atom_offset": 5186598,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.q_proj.output_min": {
    "atom_offset": 5186600,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 5186602,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.down_proj.input_max": {
    "atom_offset": 5186604,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 5186606,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 5186639,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 5186641,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.q_proj.input_min": {
    "atom_offset": 5186643,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.pre_feedforward_layernorm.weight": {
    "atom_offset": 5186645,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.gate_proj.output_min": {
    "atom_offset": 5186694,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.norm_out.weight": {
    "atom_offset": 5186696,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.30314961075782776,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 5186713,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 5186746,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.305364076979458e-4,
    "track": "int8"
  },
  "model.language_model.layers.20.self_attn.k_norm.weight": {
    "atom_offset": 5252283,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 5252292,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.08513779193162918,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 5252373,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.post.input_min": {
    "atom_offset": 5252375,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.up_proj.output_min": {
    "atom_offset": 5252377,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.output_proj.bias": {
    "atom_offset": 5252379,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.self_attn.v_proj.weight": {
    "atom_offset": 5252428,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.004736712668091059,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.v_proj.input_max": {
    "atom_offset": 5258573,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.mlp.up_proj.weight": {
    "atom_offset": 5258575,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004029281437397003,
    "track": "int8"
  },
  "model.language_model.layers.6.self_attn.q_norm.weight": {
    "atom_offset": 5553488,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.q_norm.weight": {
    "atom_offset": 5553497,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.pre_feedforward_layernorm.weight": {
    "atom_offset": 5553500,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 5553525,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 5553558,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.mlp.down_proj.weight": {
    "atom_offset": 5553560,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.005198080558329821,
    "track": "int8"
  },
  "model.language_model.layers.4.self_attn.q_norm.weight": {
    "atom_offset": 5701017,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.mlp.gate_proj.weight": {
    "atom_offset": 5701034,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004367617890238762,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 5995947,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.k_norm.weight": {
    "atom_offset": 5995949,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.post_feedforward_layernorm.weight": {
    "atom_offset": 5995952,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.post.output_max": {
    "atom_offset": 5996001,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.post.input_max": {
    "atom_offset": 5996003,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 5996005,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.gate_proj.output_max": {
    "atom_offset": 5996007,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.post_per_layer_input_norm.weight": {
    "atom_offset": 5996009,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.q_proj.input_min": {
    "atom_offset": 5996058,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.q_proj.input_max": {
    "atom_offset": 5996060,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.conv_norm.weight": {
    "atom_offset": 5996062,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.k_proj.output_min": {
    "atom_offset": 5996095,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.mlp.up_proj.weight": {
    "atom_offset": 5996097,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0037370817735791206,
    "track": "int8"
  },
  "model.language_model.layers.1.self_attn.o_proj.weight": {
    "atom_offset": 6143554,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.002660555997863412,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.o_proj.input_min": {
    "atom_offset": 6192707,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.o_proj.linear.weight": {
    "atom_offset": 6192709,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003429502947255969,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.q_proj.linear.weight": {
    "atom_offset": 6201926,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0024606299120932817,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.o_proj.input_max": {
    "atom_offset": 6211143,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 6211145,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.034202754497528076,
    "track": "int8"
  },
  "model.language_model.layers.30.per_layer_input_gate.weight": {
    "atom_offset": 6211226,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.patch_embedder.position_embedding_table": {
    "atom_offset": 6223515,
    "byte_length": 31457344,
    "num_atoms": 491521,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 6715036,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.self_attn.k_norm.weight": {
    "atom_offset": 6715038,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.k_proj.output_max": {
    "atom_offset": 6715047,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.post_feedforward_layernorm.weight": {
    "atom_offset": 6715049,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.down_proj.input_min": {
    "atom_offset": 6715098,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.q_proj.input_max": {
    "atom_offset": 6715100,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.pre_feedforward_layernorm.weight": {
    "atom_offset": 6715102,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 6715151,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.q_proj.input_min": {
    "atom_offset": 6715153,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.k_proj.linear.weight": {
    "atom_offset": 6715155,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.689099340699613e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.q_proj.linear.weight": {
    "atom_offset": 6731540,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.004244586452841759,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 6740757,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.0214074794203043,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.up_proj.input_min": {
    "atom_offset": 6740838,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.norm_pre_attn.weight": {
    "atom_offset": 6740840,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2303149551153183,
    "track": "int8"
  },
  "model.language_model.layers.33.per_layer_input_gate.weight": {
    "atom_offset": 6740857,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.pre_feedforward_layernorm.weight": {
    "atom_offset": 6753146,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.gate_proj.linear.weight": {
    "atom_offset": 6753195,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0026759349275380373,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 6790060,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.o_proj.input_min": {
    "atom_offset": 6790062,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 6790064,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.228100439533591e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.gate_proj.output_max": {
    "atom_offset": 6855601,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 6855603,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.mlp.up_proj.weight": {
    "atom_offset": 6855605,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004429134074598551,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.k_proj.input_max": {
    "atom_offset": 7003062,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.self_attn.o_proj.weight": {
    "atom_offset": 7003064,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004183070734143257,
    "track": "int8"
  },
  "model.language_model.layers.20.post_feedforward_layernorm.weight": {
    "atom_offset": 7052217,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.up_proj.input_min": {
    "atom_offset": 7052266,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.v_proj.output_min": {
    "atom_offset": 7052268,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_start.input_max": {
    "atom_offset": 7052270,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.up_proj.input_max": {
    "atom_offset": 7052272,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.q_proj.output_max": {
    "atom_offset": 7052274,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.mlp.up_proj.weight": {
    "atom_offset": 7052276,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0028912401758134365,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_start.linear.weight": {
    "atom_offset": 7199733,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.0013456570450216532,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.v_proj.output_min": {
    "atom_offset": 7232502,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.post.input_max": {
    "atom_offset": 7232504,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.post_feedforward_layernorm.weight": {
    "atom_offset": 7232506,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.post.input_max": {
    "atom_offset": 7232555,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_end.output_min": {
    "atom_offset": 7232557,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.post.output_max": {
    "atom_offset": 7232559,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.self_attn.k_norm.weight": {
    "atom_offset": 7232561,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.per_layer_input_gate.weight": {
    "atom_offset": 7232570,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.v_proj.linear.weight": {
    "atom_offset": 7244859,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002229945966973901,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.gate_proj.output_max": {
    "atom_offset": 7254076,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.k_proj.input_max": {
    "atom_offset": 7254078,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.mlp.gate_proj.weight": {
    "atom_offset": 7254080,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005444143898785114,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 7548993,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.gate_proj.input_min": {
    "atom_offset": 7549026,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.self_attn.q_norm.weight": {
    "atom_offset": 7549028,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.per_layer_projection.weight": {
    "atom_offset": 7549037,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.pre_feedforward_layernorm.weight": {
    "atom_offset": 7561326,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.v_proj.linear.weight": {
    "atom_offset": 7561375,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.458784384652972e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.pre_feedforward_layernorm.weight": {
    "atom_offset": 7577760,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.k_proj.input_max": {
    "atom_offset": 7577785,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.post_feedforward_layernorm.weight": {
    "atom_offset": 7577787,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.k_proj.input_min": {
    "atom_offset": 7577836,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.per_dim_scale": {
    "atom_offset": 7577838,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.per_layer_input_gate.weight": {
    "atom_offset": 7577843,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.self_attn.k_proj.weight": {
    "atom_offset": 7590132,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.7097995411604643e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.down_proj.output_min": {
    "atom_offset": 7602421,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_end.linear.weight": {
    "atom_offset": 7602423,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.035125258378685e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.gate_proj.input_max": {
    "atom_offset": 7618808,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.v_proj.output_min": {
    "atom_offset": 7618810,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.k_proj.linear.weight": {
    "atom_offset": 7618812,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.381889736279845e-4,
    "track": "int8"
  },
  "model.language_model.layers.3.input_layernorm.weight": {
    "atom_offset": 7635197,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.up_proj.input_min": {
    "atom_offset": 7635246,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.up_proj.linear.weight": {
    "atom_offset": 7635248,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.00239911419339478,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 7672113,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.mlp.down_proj.weight": {
    "atom_offset": 7672146,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0037985974922776222,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.v_proj.input_max": {
    "atom_offset": 7819603,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.per_layer_input_gate.weight": {
    "atom_offset": 7819605,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.up_proj.linear.weight": {
    "atom_offset": 7831894,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002030019648373127,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.gate_proj.input_max": {
    "atom_offset": 7868759,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 7868761,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.post.output_max": {
    "atom_offset": 7868763,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_end.output_max": {
    "atom_offset": 7868765,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.conv_norm.weight": {
    "atom_offset": 7868767,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 7868800,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.down_proj.output_min": {
    "atom_offset": 7868802,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.q_proj.output_min": {
    "atom_offset": 7868804,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.k_proj.linear.weight": {
    "atom_offset": 7868806,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0031834400724619627,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.up_proj.output_min": {
    "atom_offset": 7878023,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.k_proj.linear.weight": {
    "atom_offset": 7878025,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0028912401758134365,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.v_proj.output_max": {
    "atom_offset": 7887242,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 7887244,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.727915653958917e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 7952781,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.input_layernorm.weight": {
    "atom_offset": 7952783,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.layer_scalar": {
    "atom_offset": 7952808,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.post.output_max": {
    "atom_offset": 7952810,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.self_attn.k_norm.weight": {
    "atom_offset": 7952812,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_end.output_min": {
    "atom_offset": 7952821,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.layer_scalar": {
    "atom_offset": 7952823,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 7952825,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.421075038611889e-4,
    "track": "int8"
  },
  "model.language_model.layers.33.mlp.up_proj.weight": {
    "atom_offset": 8018362,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.00479822838678956,
    "track": "int8"
  },
  "model.language_model.layers.13.pre_feedforward_layernorm.weight": {
    "atom_offset": 8313275,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.o_proj.output_min": {
    "atom_offset": 8313324,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.mlp.up_proj.weight": {
    "atom_offset": 8313326,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0031680609099566936,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.v_proj.linear.weight": {
    "atom_offset": 8460783,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.189653115347028e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.q_proj.linear.weight": {
    "atom_offset": 8477168,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_start.linear.weight": {
    "atom_offset": 8486385,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.0011303519131615758,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.up_proj.output_min": {
    "atom_offset": 8519154,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 8519156,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.up_proj.output_max": {
    "atom_offset": 8519158,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 8519160,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_start.output_max": {
    "atom_offset": 8519162,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.post_feedforward_layernorm.weight": {
    "atom_offset": 8519164,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.v_proj.input_min": {
    "atom_offset": 8519189,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.v_proj.linear.weight": {
    "atom_offset": 8519191,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.958969170227647e-4,
    "track": "int8"
  },
  "model.language_model.layers.34.layer_scalar": {
    "atom_offset": 8535576,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.q_proj.output_max": {
    "atom_offset": 8535578,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.v_proj.output_max": {
    "atom_offset": 8535580,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 8535582,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 8535584,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.03789370134472847,
    "track": "int8"
  },
  "model.language_model.layers.13.self_attn.o_proj.weight": {
    "atom_offset": 8535665,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004060039296746254,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.k_proj.output_max": {
    "atom_offset": 8584818,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.mlp.down_proj.weight": {
    "atom_offset": 8584820,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004705954808741808,
    "track": "int8"
  },
  "model.language_model.layers.24.mlp.up_proj.weight": {
    "atom_offset": 8732277,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004582923371344805,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.up_proj.input_min": {
    "atom_offset": 9027190,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 9027192,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 9027194,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.up_proj.linear.weight": {
    "atom_offset": 9027196,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0030911662615835667,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.gate_proj.output_min": {
    "atom_offset": 9064061,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 9064063,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 9064065,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.458784384652972e-4,
    "track": "int8"
  },
  "model.language_model.layers.26.self_attn.q_norm.weight": {
    "atom_offset": 9129602,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 9129611,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.113127456046641e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.q_proj.input_min": {
    "atom_offset": 9195148,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 9195150,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 9195152,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.o_proj.linear.weight": {
    "atom_offset": 9195154,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002798966597765684,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.q_proj.linear.weight": {
    "atom_offset": 9204371,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.536048022098839e-4,
    "track": "int8"
  },
  "model.language_model.layers.28.self_attn.v_proj.weight": {
    "atom_offset": 9220756,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.7097995411604643e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.q_proj.linear.weight": {
    "atom_offset": 9226901,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0029373769648373127,
    "track": "int8"
  },
  "model.language_model.layers.17.mlp.up_proj.weight": {
    "atom_offset": 9236118,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006735974457114935,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.post.input_max": {
    "atom_offset": 9531031,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.norm_post_attn.weight": {
    "atom_offset": 9531033,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.13385826349258423,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.q_proj.input_max": {
    "atom_offset": 9531050,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 9531052,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.k_proj.output_max": {
    "atom_offset": 9531054,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.k_proj.input_min": {
    "atom_offset": 9531056,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.pre_feedforward_layernorm.weight": {
    "atom_offset": 9531058,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 9531107,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.k_proj.linear.weight": {
    "atom_offset": 9531109,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.304995087906718e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.up_proj.output_max": {
    "atom_offset": 9547494,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 9547496,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.post.output_min": {
    "atom_offset": 9547498,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.mlp.down_proj.weight": {
    "atom_offset": 9547500,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004675196949392557,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.down_proj.output_min": {
    "atom_offset": 9694957,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 9694959,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.228469428606331e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.v_proj.output_min": {
    "atom_offset": 9760496,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.q_proj.output_min": {
    "atom_offset": 9760498,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.mlp.gate_proj.weight": {
    "atom_offset": 9760500,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004736712668091059,
    "track": "int8"
  },
  "model.language_model.layers.30.self_attn.q_norm.weight": {
    "atom_offset": 10055413,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.self_attn.q_norm.weight": {
    "atom_offset": 10055422,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 10055431,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.mlp.down_proj.weight": {
    "atom_offset": 10055433,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004336860030889511,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.q_proj.input_min": {
    "atom_offset": 10350346,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 10350348,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.layer_scalar": {
    "atom_offset": 10350350,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.v_proj.output_max": {
    "atom_offset": 10350352,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.post_feedforward_layernorm.weight": {
    "atom_offset": 10350354,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.q_proj.output_max": {
    "atom_offset": 10350379,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.post.output_max": {
    "atom_offset": 10350381,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 10350383,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.post_attention_layernorm.weight": {
    "atom_offset": 10350385,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.v_proj.input_max": {
    "atom_offset": 10350434,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.v_proj.input_min": {
    "atom_offset": 10350436,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.q_proj.output_min": {
    "atom_offset": 10350438,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 10350440,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.input_layernorm.weight": {
    "atom_offset": 10350442,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.k_proj.input_max": {
    "atom_offset": 10350491,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.conv_norm.weight": {
    "atom_offset": 10350493,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.gate_proj.input_min": {
    "atom_offset": 10350526,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.k_norm.weight": {
    "atom_offset": 10350528,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.up_proj.output_min": {
    "atom_offset": 10350531,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.q_proj.input_max": {
    "atom_offset": 10350533,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 10350535,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.input_layernorm.weight": {
    "atom_offset": 10350537,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 10350586,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.post_per_layer_input_norm.weight": {
    "atom_offset": 10350588,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.post.linear.weight": {
    "atom_offset": 10350637,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0012610728153958917,
    "track": "int8"
  },
  "model.language_model.layers.14.post_per_layer_input_norm.weight": {
    "atom_offset": 10367022,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.up_proj.linear.weight": {
    "atom_offset": 10367071,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0027066930197179317,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.k_proj.input_min": {
    "atom_offset": 10403936,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.down_proj.input_max": {
    "atom_offset": 10403938,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.v_proj.output_min": {
    "atom_offset": 10403940,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 10403942,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.gate_proj.input_max": {
    "atom_offset": 10403944,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.post_feedforward_layernorm.weight": {
    "atom_offset": 10403946,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.q_proj.output_max": {
    "atom_offset": 10403971,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.k_proj.input_min": {
    "atom_offset": 10403973,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 10403975,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.382627714425325e-4,
    "track": "int8"
  },
  "model.language_model.layers.24.post_attention_layernorm.weight": {
    "atom_offset": 10469512,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.o_proj.input_max": {
    "atom_offset": 10469561,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.up_proj.input_max": {
    "atom_offset": 10469563,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.up_proj.linear.weight": {
    "atom_offset": 10469565,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002245324896648526,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.pre_feedforward_layernorm.weight": {
    "atom_offset": 10506430,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.k_proj.linear.weight": {
    "atom_offset": 10506455,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003060408402234316,
    "track": "int8"
  },
  "model.language_model.layers.23.per_layer_projection.weight": {
    "atom_offset": 10515672,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.input_layernorm.weight": {
    "atom_offset": 10527961,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.self_attn.q_norm.weight": {
    "atom_offset": 10528010,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.pre_feedforward_layernorm.weight": {
    "atom_offset": 10528019,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.o_proj.output_max": {
    "atom_offset": 10528068,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.v_proj.linear.weight": {
    "atom_offset": 10528070,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.00173013040330261,
    "track": "int8"
  },
  "model.language_model.layers.29.mlp.down_proj.weight": {
    "atom_offset": 10544455,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004644439090043306,
    "track": "int8"
  },
  "model.language_model.embed_tokens_per_layer.weight": {
    "atom_offset": 10839368,
    "byte_length": 4697620544,
    "num_atoms": 73400321,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.mlp.up_proj.weight": {
    "atom_offset": 84239689,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0046136812306940556,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.down_proj.input_min": {
    "atom_offset": 84534602,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.post_per_layer_input_norm.weight": {
    "atom_offset": 84534604,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.q_proj.input_max": {
    "atom_offset": 84534653,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.k_proj.input_max": {
    "atom_offset": 84534655,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 84534657,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.v_proj.input_max": {
    "atom_offset": 84534659,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.gate_proj.input_max": {
    "atom_offset": 84534661,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.q_norm.weight": {
    "atom_offset": 84534663,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.mlp.down_proj.weight": {
    "atom_offset": 84534666,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004367617890238762,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.v_proj.output_min": {
    "atom_offset": 84829579,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.up_proj.output_max": {
    "atom_offset": 84829581,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.o_proj.linear.weight": {
    "atom_offset": 84829583,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0033987450879067183,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.up_proj.linear.weight": {
    "atom_offset": 84838800,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.00239911419339478,
    "track": "int8"
  },
  "model.language_model.layers.20.mlp.up_proj.weight": {
    "atom_offset": 84875665,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.language_model.layers.21.post_per_layer_input_norm.weight": {
    "atom_offset": 85170578,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.gate_proj.output_max": {
    "atom_offset": 85170627,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 85170629,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.88207452185452e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.v_proj.input_max": {
    "atom_offset": 85236166,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.self_attn.o_proj.weight": {
    "atom_offset": 85236168,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.004152312874794006,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.v_proj.input_min": {
    "atom_offset": 85334473,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.up_proj.input_max": {
    "atom_offset": 85334475,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.pre_feedforward_layernorm.weight": {
    "atom_offset": 85334477,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.o_proj.linear.weight": {
    "atom_offset": 85334526,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003352608298882842,
    "track": "int8"
  },
  "model.audio_tower.layers.11.norm_pre_attn.weight": {
    "atom_offset": 85343743,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.1683070808649063,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_start.input_min": {
    "atom_offset": 85343760,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.down_proj.output_min": {
    "atom_offset": 85343762,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.relative_k_proj.weight": {
    "atom_offset": 85343764,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014071727637201548,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.k_proj.output_min": {
    "atom_offset": 85360149,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.q_norm.weight": {
    "atom_offset": 85360151,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 85360154,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.k_norm.weight": {
    "atom_offset": 85360156,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.relative_k_proj.weight": {
    "atom_offset": 85360159,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.00219918810762465,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.pre_feedforward_layernorm.weight": {
    "atom_offset": 85376544,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.down_proj.input_min": {
    "atom_offset": 85376569,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.embed_audio.embedding_projection.weight": {
    "atom_offset": 85376571,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0021684300154447556,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.post.input_max": {
    "atom_offset": 85413436,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.gate_proj.input_min": {
    "atom_offset": 85413438,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.self_attn.k_norm.weight": {
    "atom_offset": 85413440,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.up_proj.input_max": {
    "atom_offset": 85413449,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.v_proj.input_min": {
    "atom_offset": 85413451,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.v_proj.output_max": {
    "atom_offset": 85413453,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 85413455,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.920890835113823e-4,
    "track": "int8"
  },
  "model.language_model.layers.2.mlp.gate_proj.weight": {
    "atom_offset": 85478992,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0036448079627007246,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 85626449,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.self_attn.q_norm.weight": {
    "atom_offset": 85626451,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.self_attn.k_proj.weight": {
    "atom_offset": 85626460,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 0.002429872052744031,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.k_proj.output_max": {
    "atom_offset": 85638749,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.o_proj.input_min": {
    "atom_offset": 85638751,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.k_proj.linear.weight": {
    "atom_offset": 85638753,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.497600697912276e-4,
    "track": "int8"
  },
  "model.language_model.layers.23.pre_feedforward_layernorm.weight": {
    "atom_offset": 85655138,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.gate_proj.output_max": {
    "atom_offset": 85655187,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.k_proj.output_max": {
    "atom_offset": 85655189,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.self_attn.k_norm.weight": {
    "atom_offset": 85655191,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.up_proj.output_max": {
    "atom_offset": 85655200,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.layer_scalar": {
    "atom_offset": 85655202,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.q_proj.input_min": {
    "atom_offset": 85655204,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 85655206,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.920890835113823e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.gate_proj.input_max": {
    "atom_offset": 85720743,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.post_per_layer_input_norm.weight": {
    "atom_offset": 85720745,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.v_proj.input_max": {
    "atom_offset": 85720794,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 85720796,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.down_proj.input_min": {
    "atom_offset": 85720829,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.o_proj.linear.weight": {
    "atom_offset": 85720831,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0042753443121910095,
    "track": "int8"
  },
  "model.language_model.layers.27.layer_scalar": {
    "atom_offset": 85730048,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 85730050,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.post_feedforward_layernorm.weight": {
    "atom_offset": 85730052,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 85730101,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 85730103,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.q_proj.input_min": {
    "atom_offset": 85730105,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.up_proj.output_min": {
    "atom_offset": 85730107,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.gate_proj.linear.weight": {
    "atom_offset": 85730109,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002798966597765684,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.pre_feedforward_layernorm.weight": {
    "atom_offset": 85766974,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.post_attention_layernorm.weight": {
    "atom_offset": 85766999,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.self_attn.v_proj.weight": {
    "atom_offset": 85767048,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.729023203253746e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.down_proj.linear.weight": {
    "atom_offset": 85773193,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0015763409901410341,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.q_proj.linear.weight": {
    "atom_offset": 85810058,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0011687992373481393,
    "track": "int8"
  },
  "model.language_model.layers.18.self_attn.k_proj.weight": {
    "atom_offset": 85826443,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.594457568600774e-4,
    "track": "int8"
  },
  "model.language_model.layers.14.post_attention_layernorm.weight": {
    "atom_offset": 85832588,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.o_proj.output_min": {
    "atom_offset": 85832637,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 85832639,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.112758466973901e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.k_proj.linear.weight": {
    "atom_offset": 85898176,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.881704950705171e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.o_proj.output_max": {
    "atom_offset": 85914561,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.q_proj.input_max": {
    "atom_offset": 85914563,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.self_attn.q_norm.weight": {
    "atom_offset": 85914565,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.post_feedforward_layernorm.weight": {
    "atom_offset": 85914574,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 85914599,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.q_proj.input_max": {
    "atom_offset": 85914601,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_end.input_max": {
    "atom_offset": 85914603,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 85914605,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.up_proj.output_max": {
    "atom_offset": 85914607,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.post_attention_layernorm.weight": {
    "atom_offset": 85914609,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.v_proj.linear.weight": {
    "atom_offset": 85914634,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.76636297814548e-4,
    "track": "int8"
  },
  "model.language_model.layers.30.layer_scalar": {
    "atom_offset": 85931019,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.input_layernorm.weight": {
    "atom_offset": 85931021,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.o_proj.linear.weight": {
    "atom_offset": 85931070,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0021376721560955048,
    "track": "int8"
  },
  "model.audio_tower.layers.5.norm_pre_attn.weight": {
    "atom_offset": 85940287,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2736220359802246,
    "track": "int8"
  },
  "model.language_model.layers.25.self_attn.o_proj.weight": {
    "atom_offset": 85940304,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004982775542885065,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.v_proj.output_min": {
    "atom_offset": 85989457,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.v_proj.output_min": {
    "atom_offset": 85989459,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 85989461,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_end.input_max": {
    "atom_offset": 85989463,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_start.linear.weight": {
    "atom_offset": 85989465,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.001253383350558579,
    "track": "int8"
  },
  "model.language_model.layers.5.mlp.up_proj.weight": {
    "atom_offset": 86022234,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.00366018689237535,
    "track": "int8"
  },
  "model.audio_tower.layers.9.norm_out.weight": {
    "atom_offset": 86169691,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.33070865273475647,
    "track": "int8"
  },
  "model.language_model.layers.12.self_attn.k_norm.weight": {
    "atom_offset": 86169708,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.post_feedforward_layernorm.weight": {
    "atom_offset": 86169717,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.self_attn.o_proj.weight": {
    "atom_offset": 86169766,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0028451033867895603,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.v_proj.input_min": {
    "atom_offset": 86218919,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.pre_feedforward_layernorm.weight": {
    "atom_offset": 86218921,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.input_layernorm.weight": {
    "atom_offset": 86218970,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.q_proj.linear.weight": {
    "atom_offset": 86218995,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0021222932264208794,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 86228212,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 4.229207697790116e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.v_proj.linear.weight": {
    "atom_offset": 86293749,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003106545191258192,
    "track": "int8"
  },
  "model.language_model.layers.13.mlp.down_proj.weight": {
    "atom_offset": 86302966,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004890501964837313,
    "track": "int8"
  },
  "model.audio_tower.layers.0.norm_out.weight": {
    "atom_offset": 86450423,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.18208661675453186,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.v_proj.input_min": {
    "atom_offset": 86450440,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_end.linear.weight": {
    "atom_offset": 86450442,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.112019906751812e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.v_proj.input_max": {
    "atom_offset": 86466827,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 86466829,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.k_proj.output_min": {
    "atom_offset": 86466862,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.post.input_min": {
    "atom_offset": 86466864,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.k_proj.linear.weight": {
    "atom_offset": 86466866,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.00390625,
    "track": "int8"
  },
  "model.language_model.layers.12.self_attn.o_proj.weight": {
    "atom_offset": 86476083,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0026913140900433064,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.post.output_max": {
    "atom_offset": 86525236,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.layer_scalar": {
    "atom_offset": 86525238,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.input_layernorm.weight": {
    "atom_offset": 86525240,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.q_proj.output_min": {
    "atom_offset": 86525289,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 86525291,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.q_proj.linear.weight": {
    "atom_offset": 86525293,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010688360780477524,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.k_norm.weight": {
    "atom_offset": 86541678,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.gate_proj.output_min": {
    "atom_offset": 86541681,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.pre_feedforward_layernorm.weight": {
    "atom_offset": 86541683,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.v_proj.output_max": {
    "atom_offset": 86541732,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.post_feedforward_layernorm.weight": {
    "atom_offset": 86541734,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.v_proj.linear.weight": {
    "atom_offset": 86541783,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0030757873319089413,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.post_feedforward_layernorm.weight": {
    "atom_offset": 86551000,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.o_proj.input_max": {
    "atom_offset": 86551025,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.post_attention_layernorm.weight": {
    "atom_offset": 86551027,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.v_proj.linear.weight": {
    "atom_offset": 86551076,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0022914616856724024,
    "track": "int8"
  },
  "model.language_model.layers.34.post_per_layer_input_norm.weight": {
    "atom_offset": 86560293,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.gate_proj.output_max": {
    "atom_offset": 86560342,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.q_norm.weight": {
    "atom_offset": 86560344,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.down_proj.output_max": {
    "atom_offset": 86560347,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.per_layer_projection.weight": {
    "atom_offset": 86560349,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.self_attn.q_norm.weight": {
    "atom_offset": 86572638,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.norm_post_attn.weight": {
    "atom_offset": 86572655,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.0974409431219101,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_start.output_min": {
    "atom_offset": 86572672,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.norm_post_attn.weight": {
    "atom_offset": 86572674,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.17716535925865173,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 86572691,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.gate_proj.input_min": {
    "atom_offset": 86572724,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.post.output_min": {
    "atom_offset": 86572726,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.post_attention_layernorm.weight": {
    "atom_offset": 86572728,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.down_proj.output_max": {
    "atom_offset": 86572777,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.norm_out.weight": {
    "atom_offset": 86572779,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.12992125749588013,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.v_proj.output_max": {
    "atom_offset": 86572796,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 86572798,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 0.001330277998931706,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.down_proj.output_min": {
    "atom_offset": 86638335,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.per_layer_projection.weight": {
    "atom_offset": 86638337,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.mlp.gate_proj.weight": {
    "atom_offset": 86650626,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.006889763753861189,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_start.input_min": {
    "atom_offset": 86798083,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.self_attn.q_proj.weight": {
    "atom_offset": 86798085,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004183070734143257,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.v_proj.input_min": {
    "atom_offset": 86847238,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.gate_proj.output_min": {
    "atom_offset": 86847240,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.post_feedforward_layernorm.weight": {
    "atom_offset": 86847242,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_start.output_max": {
    "atom_offset": 86847267,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 86847269,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 86847271,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_start.output_max": {
    "atom_offset": 86847273,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.self_attn.q_proj.weight": {
    "atom_offset": 86847275,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0036448079627007246,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.v_proj.input_max": {
    "atom_offset": 86896428,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.post_attention_layernorm.weight": {
    "atom_offset": 86896430,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 86896479,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.down_proj.output_max": {
    "atom_offset": 86896481,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.per_layer_projection.weight": {
    "atom_offset": 86896483,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.per_layer_projection.weight": {
    "atom_offset": 86908772,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.q_proj.linear.weight": {
    "atom_offset": 86921061,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.381151758134365e-4,
    "track": "int8"
  },
  "model.language_model.layers.25.per_layer_projection.weight": {
    "atom_offset": 86937446,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 86949735,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.post_attention_layernorm.weight": {
    "atom_offset": 86949737,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.up_proj.input_max": {
    "atom_offset": 86949786,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.q_proj.output_max": {
    "atom_offset": 86949788,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.k_proj.input_min": {
    "atom_offset": 86949790,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.up_proj.output_min": {
    "atom_offset": 86949792,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 86949794,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 4.671352216973901e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.input_layernorm.weight": {
    "atom_offset": 87015331,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.v_proj.output_max": {
    "atom_offset": 87015356,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.q_proj.output_max": {
    "atom_offset": 87015358,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.self_attn.k_norm.weight": {
    "atom_offset": 87015360,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_end.input_max": {
    "atom_offset": 87015377,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 87015379,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 87015412,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.99741649441421e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.up_proj.input_max": {
    "atom_offset": 87080949,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.norm_post_attn.weight": {
    "atom_offset": 87080951,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.14370079338550568,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.q_proj.input_max": {
    "atom_offset": 87080968,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 87080970,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.post.output_min": {
    "atom_offset": 87080972,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.v_proj.output_max": {
    "atom_offset": 87080974,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.self_attn.q_norm.weight": {
    "atom_offset": 87080976,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_end.output_max": {
    "atom_offset": 87080993,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.mlp.up_proj.weight": {
    "atom_offset": 87080995,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.002983513753861189,
    "track": "int8"
  },
  "model.language_model.layers.34.self_attn.q_proj.weight": {
    "atom_offset": 87375908,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.002660555997863412,
    "track": "int8"
  },
  "model.language_model.layers.10.self_attn.v_proj.weight": {
    "atom_offset": 87474213,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0028143455274403095,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.o_proj.input_min": {
    "atom_offset": 87480358,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 87480360,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 87480362,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_start.input_min": {
    "atom_offset": 87480364,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.q_proj.linear.weight": {
    "atom_offset": 87480366,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0050750491209328175,
    "track": "int8"
  },
  "model.language_model.layers.12.per_layer_input_gate.weight": {
    "atom_offset": 87489583,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.post.output_min": {
    "atom_offset": 87501872,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.q_proj.input_min": {
    "atom_offset": 87501874,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.post_attention_layernorm.weight": {
    "atom_offset": 87501876,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.k_proj.input_min": {
    "atom_offset": 87501925,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 87501927,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.post_per_layer_input_norm.weight": {
    "atom_offset": 87501929,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_end.input_max": {
    "atom_offset": 87501978,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 87501980,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 87502013,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.per_layer_input_gate.weight": {
    "atom_offset": 87502015,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 87514304,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.344180390238762e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.post_feedforward_layernorm.weight": {
    "atom_offset": 87579841,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.per_layer_input_gate.weight": {
    "atom_offset": 87579866,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.post_per_layer_input_norm.weight": {
    "atom_offset": 87592155,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_start.output_max": {
    "atom_offset": 87592204,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.post_attention_layernorm.weight": {
    "atom_offset": 87592206,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.post_per_layer_input_norm.weight": {
    "atom_offset": 87592231,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 87592280,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.norm_post_attn.weight": {
    "atom_offset": 87592282,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.06545275449752808,
    "track": "int8"
  },
  "model.language_model.layers.34.per_layer_input_gate.weight": {
    "atom_offset": 87592299,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.v_proj.linear.weight": {
    "atom_offset": 87604588,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0031526819802820683,
    "track": "int8"
  },
  "model.language_model.layers.32.post_per_layer_input_norm.weight": {
    "atom_offset": 87613805,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.self_attn.o_proj.weight": {
    "atom_offset": 87613854,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0031219241209328175,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 87663007,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.025959646329283714,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.q_proj.output_max": {
    "atom_offset": 87663088,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.k_proj.output_min": {
    "atom_offset": 87663090,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_start.output_min": {
    "atom_offset": 87663092,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.self_attn.q_norm.weight": {
    "atom_offset": 87663094,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.k_proj.input_max": {
    "atom_offset": 87663111,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 87663113,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 87663115,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.k_proj.linear.weight": {
    "atom_offset": 87663117,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0033987450879067183,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.k_proj.output_max": {
    "atom_offset": 87672334,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 87672336,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.k_proj.output_min": {
    "atom_offset": 87672338,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.up_proj.linear.weight": {
    "atom_offset": 87672340,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002276082755997777,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.up_proj.linear.weight": {
    "atom_offset": 87709205,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0023375984746962786,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.k_proj.output_max": {
    "atom_offset": 87746070,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.q_proj.output_max": {
    "atom_offset": 87746072,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.q_proj.linear.weight": {
    "atom_offset": 87746074,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.189653115347028e-4,
    "track": "int8"
  },
  "model.language_model.layers.7.mlp.gate_proj.weight": {
    "atom_offset": 87762459,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.005505659617483616,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_start.input_max": {
    "atom_offset": 87909916,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.per_layer_input_gate.weight": {
    "atom_offset": 87909918,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.layer_scalar": {
    "atom_offset": 87922207,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.self_attn.o_proj.weight": {
    "atom_offset": 87922209,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003506397595629096,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 87971362,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.v_proj.input_max": {
    "atom_offset": 87971395,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.up_proj.output_min": {
    "atom_offset": 87971397,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.self_attn.v_proj.weight": {
    "atom_offset": 87971399,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0023683563340455294,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_start.input_min": {
    "atom_offset": 87977544,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 87977546,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.subsample_conv_projection.input_proj_linear.weight": {
    "atom_offset": 87977548,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0029219980351626873,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 87993933,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.497231708839536e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.up_proj.input_max": {
    "atom_offset": 88059470,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 88059472,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.v_proj.output_max": {
    "atom_offset": 88059474,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.v_proj.linear.weight": {
    "atom_offset": 88059476,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003306471509858966,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.q_norm.weight": {
    "atom_offset": 88068693,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.layer_scalar": {
    "atom_offset": 88068696,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.mlp.gate_proj.weight": {
    "atom_offset": 88068698,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004736712668091059,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.down_proj.output_max": {
    "atom_offset": 88363611,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 88363613,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.down_proj.output_max": {
    "atom_offset": 88363646,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 88363648,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 88363650,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_start.output_min": {
    "atom_offset": 88363652,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 88363654,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.574126357212663e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.input_layernorm.weight": {
    "atom_offset": 88429191,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.input_layernorm.weight": {
    "atom_offset": 88429216,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 88429265,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.pre_feedforward_layernorm.weight": {
    "atom_offset": 88429267,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.k_norm.weight": {
    "atom_offset": 88429292,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.k_proj.input_min": {
    "atom_offset": 88429295,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.k_proj.linear.weight": {
    "atom_offset": 88429297,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003629429033026099,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.gate_proj.input_max": {
    "atom_offset": 88438514,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.relative_k_proj.weight": {
    "atom_offset": 88438516,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0019146776758134365,
    "track": "int8"
  },
  "model.language_model.layers.28.post_feedforward_layernorm.weight": {
    "atom_offset": 88454901,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.v_proj.linear.weight": {
    "atom_offset": 88454950,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0023068406153470278,
    "track": "int8"
  },
  "model.language_model.layers.7.self_attn.q_proj.weight": {
    "atom_offset": 88464167,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003429502947255969,
    "track": "int8"
  },
  "model.language_model.layers.19.self_attn.o_proj.weight": {
    "atom_offset": 88513320,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.003967765718698502,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_end.input_min": {
    "atom_offset": 88611625,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.q_proj.input_max": {
    "atom_offset": 88611627,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.up_proj.input_max": {
    "atom_offset": 88611629,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.v_proj.output_max": {
    "atom_offset": 88611631,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.per_dim_scale": {
    "atom_offset": 88611633,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.mlp.gate_proj.weight": {
    "atom_offset": 88611638,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004183070734143257,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.k_proj.input_max": {
    "atom_offset": 88759095,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.post_per_layer_input_norm.weight": {
    "atom_offset": 88759097,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.input_layernorm.weight": {
    "atom_offset": 88759146,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.down_proj.input_max": {
    "atom_offset": 88759195,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.k_norm.weight": {
    "atom_offset": 88759197,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.up_proj.output_max": {
    "atom_offset": 88759200,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 88759202,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.q_norm.weight": {
    "atom_offset": 88759204,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.q_proj.output_max": {
    "atom_offset": 88759207,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.gate_proj.input_max": {
    "atom_offset": 88759209,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.down_proj.input_min": {
    "atom_offset": 88759211,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.v_proj.output_min": {
    "atom_offset": 88759213,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.conv_norm.weight": {
    "atom_offset": 88759215,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.mlp.up_proj.weight": {
    "atom_offset": 88759248,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.language_model.layers.26.input_layernorm.weight": {
    "atom_offset": 89054161,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.v_proj.input_min": {
    "atom_offset": 89054210,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.self_attn.q_norm.weight": {
    "atom_offset": 89054212,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.norm_out.weight": {
    "atom_offset": 89054221,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.19389763474464417,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.post_attention_layernorm.weight": {
    "atom_offset": 89054238,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.self_attn.v_proj.weight": {
    "atom_offset": 89054263,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.003229576861485839,
    "track": "int8"
  },
  "model.language_model.layers.17.self_attn.o_proj.weight": {
    "atom_offset": 89060408,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0026144192088395357,
    "track": "int8"
  },
  "model.language_model.layers.19.post_per_layer_input_norm.weight": {
    "atom_offset": 89109561,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.down_proj.output_max": {
    "atom_offset": 89109610,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.q_proj.output_max": {
    "atom_offset": 89109612,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.self_attn.k_proj.weight": {
    "atom_offset": 89109614,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0028912401758134365,
    "track": "int8"
  },
  "model.language_model.layers.12.post_per_layer_input_norm.weight": {
    "atom_offset": 89115759,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.self_attn.q_proj.weight": {
    "atom_offset": 89115808,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0033218504395335913,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.o_proj.linear.weight": {
    "atom_offset": 89164961,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0028143455274403095,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.q_proj.output_max": {
    "atom_offset": 89174178,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 89174180,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.per_layer_input_gate.weight": {
    "atom_offset": 89174213,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.layer_scalar": {
    "atom_offset": 89186502,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.k_proj.output_min": {
    "atom_offset": 89186504,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.down_proj.output_max": {
    "atom_offset": 89186506,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.q_proj.output_max": {
    "atom_offset": 89186508,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 89186510,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 89186512,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.mlp.down_proj.weight": {
    "atom_offset": 89186545,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004213828593492508,
    "track": "int8"
  },
  "model.language_model.layers.30.mlp.up_proj.weight": {
    "atom_offset": 89334002,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0030450294725596905,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 89628915,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.post_attention_layernorm.weight": {
    "atom_offset": 89628948,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.norm_out.weight": {
    "atom_offset": 89628973,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.10383858531713486,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.k_proj.input_max": {
    "atom_offset": 89628990,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.mlp.down_proj.weight": {
    "atom_offset": 89628992,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004213828593492508,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.q_proj.output_max": {
    "atom_offset": 89923905,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.post_feedforward_layernorm.weight": {
    "atom_offset": 89923907,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.k_proj.input_max": {
    "atom_offset": 89923956,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_end.output_max": {
    "atom_offset": 89923958,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.o_proj.linear.weight": {
    "atom_offset": 89923960,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0024606299120932817,
    "track": "int8"
  },
  "model.language_model.layers.10.per_layer_input_gate.weight": {
    "atom_offset": 89933177,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.up_proj.linear.weight": {
    "atom_offset": 89945466,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0025682824198156595,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_end.output_min": {
    "atom_offset": 89982331,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.o_proj.output_max": {
    "atom_offset": 89982333,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.o_proj.linear.weight": {
    "atom_offset": 89982335,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0021684300154447556,
    "track": "int8"
  },
  "model.language_model.layers.27.mlp.gate_proj.weight": {
    "atom_offset": 89991552,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004244586452841759,
    "track": "int8"
  },
  "model.language_model.layers.14.self_attn.k_proj.weight": {
    "atom_offset": 90286465,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 0.0018454724922776222,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.q_proj.output_min": {
    "atom_offset": 90298754,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.per_layer_input_gate.weight": {
    "atom_offset": 90298756,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.gate_proj.input_max": {
    "atom_offset": 90311045,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.v_proj.output_min": {
    "atom_offset": 90311047,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.self_attn.q_proj.weight": {
    "atom_offset": 90311049,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0038447342813014984,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.up_proj.linear.weight": {
    "atom_offset": 90360202,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002352977404370904,
    "track": "int8"
  },
  "model.language_model.layers.9.per_layer_projection.weight": {
    "atom_offset": 90397067,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.pre_feedforward_layernorm.weight": {
    "atom_offset": 90409356,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.post.input_max": {
    "atom_offset": 90409405,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.down_proj.output_min": {
    "atom_offset": 90409407,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.v_proj.input_min": {
    "atom_offset": 90409409,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.per_layer_projection.weight": {
    "atom_offset": 90409411,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.down_proj.output_max": {
    "atom_offset": 90421700,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.mlp.down_proj.weight": {
    "atom_offset": 90421702,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.003875492140650749,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.q_proj.input_min": {
    "atom_offset": 90716615,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.down_proj.linear.weight": {
    "atom_offset": 90716617,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0017993355868384242,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.q_proj.input_max": {
    "atom_offset": 90753482,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 90753484,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.input_layernorm.weight": {
    "atom_offset": 90753486,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 90753535,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 90753537,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.per_dim_scale": {
    "atom_offset": 90753539,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 90753544,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_end.output_max": {
    "atom_offset": 90753546,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 90753548,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_end.input_min": {
    "atom_offset": 90753550,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.post.input_min": {
    "atom_offset": 90753552,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 90753554,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.v_proj.input_max": {
    "atom_offset": 90753556,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 90753558,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.pre_feedforward_layernorm.weight": {
    "atom_offset": 90753560,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.gate_proj.output_min": {
    "atom_offset": 90753609,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 90753611,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.k_proj.input_max": {
    "atom_offset": 90753613,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.q_proj.input_max": {
    "atom_offset": 90753615,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.gate_proj.input_max": {
    "atom_offset": 90753617,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_start.input_max": {
    "atom_offset": 90753619,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.post_feedforward_layernorm.weight": {
    "atom_offset": 90753621,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.per_layer_input_gate.weight": {
    "atom_offset": 90753670,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.k_proj.input_max": {
    "atom_offset": 90765959,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.self_attn.o_proj.weight": {
    "atom_offset": 90765961,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.00479822838678956,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.k_norm.weight": {
    "atom_offset": 90815114,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.o_proj.output_max": {
    "atom_offset": 90815117,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.self_attn.q_proj.weight": {
    "atom_offset": 90815119,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0035679133143275976,
    "track": "int8"
  },
  "model.language_model.layers.1.per_layer_projection.weight": {
    "atom_offset": 90864272,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.k_proj.input_min": {
    "atom_offset": 90876561,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.o_proj.output_min": {
    "atom_offset": 90876563,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.q_proj.input_max": {
    "atom_offset": 90876565,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.self_attn.v_proj.weight": {
    "atom_offset": 90876567,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0034448818769305944,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.v_proj.output_min": {
    "atom_offset": 90882712,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.q_norm.weight": {
    "atom_offset": 90882714,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 90882717,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.self_attn.k_norm.weight": {
    "atom_offset": 90882750,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 90882767,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 8.842888637445867e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.o_proj.output_min": {
    "atom_offset": 90948304,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 90948306,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.per_layer_projection.weight": {
    "atom_offset": 90948308,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.o_proj.output_min": {
    "atom_offset": 90960597,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.post_feedforward_layernorm.weight": {
    "atom_offset": 90960599,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_end.linear.weight": {
    "atom_offset": 90960648,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.534941054880619e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.down_proj.input_min": {
    "atom_offset": 90977033,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.self_attn.q_norm.weight": {
    "atom_offset": 90977035,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.post_attention_layernorm.weight": {
    "atom_offset": 90977044,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.k_proj.input_max": {
    "atom_offset": 90977093,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.k_proj.output_min": {
    "atom_offset": 90977095,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.v_proj.linear.weight": {
    "atom_offset": 90977097,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010842150077223778,
    "track": "int8"
  },
  "model.language_model.layers.3.self_attn.k_norm.weight": {
    "atom_offset": 90993482,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.o_proj.output_min": {
    "atom_offset": 90993491,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.down_proj.input_min": {
    "atom_offset": 90993493,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.down_proj.input_min": {
    "atom_offset": 90993495,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 90993497,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.down_proj.linear.weight": {
    "atom_offset": 90993499,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0017916461220011115,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 91030364,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.embed_vision.embedding_projection.weight": {
    "atom_offset": 91030397,
    "byte_length": 1179712,
    "num_atoms": 18433,
    "scale": 0.005228838417679071,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.v_proj.output_min": {
    "atom_offset": 91048830,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.k_proj.input_max": {
    "atom_offset": 91048832,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 91048834,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_end.output_min": {
    "atom_offset": 91048867,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.up_proj.output_min": {
    "atom_offset": 91048869,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.post_attention_layernorm.weight": {
    "atom_offset": 91048871,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.layer_scalar": {
    "atom_offset": 91048896,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.self_attn.k_proj.weight": {
    "atom_offset": 91048898,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.402220656629652e-4,
    "track": "int8"
  },
  "model.language_model.layers.24.input_layernorm.weight": {
    "atom_offset": 91055043,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.self_attn.v_proj.weight": {
    "atom_offset": 91055092,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.38299699453637e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 91061237,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 91061270,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.per_dim_scale": {
    "atom_offset": 91061303,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.post_attention_layernorm.weight": {
    "atom_offset": 91061308,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.self_attn.q_proj.weight": {
    "atom_offset": 91061357,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.004060039296746254,
    "track": "int8"
  },
  "model.language_model.layers.5.per_layer_input_gate.weight": {
    "atom_offset": 91159662,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.up_proj.output_max": {
    "atom_offset": 91171951,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.layer_scalar": {
    "atom_offset": 91171953,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 91171955,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.self_attn.q_proj.weight": {
    "atom_offset": 91171957,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0038447342813014984,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.post_attention_layernorm.weight": {
    "atom_offset": 91221110,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.q_norm.weight": {
    "atom_offset": 91221135,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.v_proj.input_max": {
    "atom_offset": 91221138,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.post_feedforward_layernorm.weight": {
    "atom_offset": 91221140,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.k_proj.output_max": {
    "atom_offset": 91221189,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_end.input_min": {
    "atom_offset": 91221191,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.o_proj.input_max": {
    "atom_offset": 91221193,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.input_layernorm.weight": {
    "atom_offset": 91221195,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.per_layer_projection.weight": {
    "atom_offset": 91221220,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.self_attn.v_proj.weight": {
    "atom_offset": 91233509,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.805917851626873e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.input_layernorm.weight": {
    "atom_offset": 91239654,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.post.input_min": {
    "atom_offset": 91239679,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.k_norm.weight": {
    "atom_offset": 91239681,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.mlp.gate_proj.weight": {
    "atom_offset": 91239684,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0054133860394358635,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_start.output_max": {
    "atom_offset": 91534597,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.v_proj.output_min": {
    "atom_offset": 91534599,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_end.output_max": {
    "atom_offset": 91534601,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.v_proj.input_min": {
    "atom_offset": 91534603,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.v_proj.linear.weight": {
    "atom_offset": 91534605,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0038908710703253746,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.q_proj.input_min": {
    "atom_offset": 91543822,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.up_proj.output_min": {
    "atom_offset": 91543824,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.v_proj.linear.weight": {
    "atom_offset": 91543826,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0014917568769305944,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.k_proj.output_min": {
    "atom_offset": 91553043,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.post_attention_layernorm.weight": {
    "atom_offset": 91553045,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.mlp.down_proj.weight": {
    "atom_offset": 91553070,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004552165511995554,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.k_proj.input_min": {
    "atom_offset": 91700527,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.post_feedforward_layernorm.weight": {
    "atom_offset": 91700529,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.q_proj.output_max": {
    "atom_offset": 91700578,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_start.input_min": {
    "atom_offset": 91700580,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.self_attn.v_proj.weight": {
    "atom_offset": 91700582,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.7674705274403095e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.down_proj.linear.weight": {
    "atom_offset": 91706727,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.001999261789023876,
    "track": "int8"
  },
  "model.language_model.layers.18.post_per_layer_input_norm.weight": {
    "atom_offset": 91743592,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 91743641,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.v_proj.input_max": {
    "atom_offset": 91743643,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.per_dim_scale": {
    "atom_offset": 91743645,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.up_proj.linear.weight": {
    "atom_offset": 91743650,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0024144931230694056,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_start.input_min": {
    "atom_offset": 91780515,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.pre_feedforward_layernorm.weight": {
    "atom_offset": 91780517,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 91780566,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.v_proj.input_min": {
    "atom_offset": 91780599,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.self_attn.v_proj.weight": {
    "atom_offset": 91780601,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 0.0029681348241865635,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.v_proj.input_min": {
    "atom_offset": 91792890,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_start.input_max": {
    "atom_offset": 91792892,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.down_proj.output_min": {
    "atom_offset": 91792894,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.down_proj.input_min": {
    "atom_offset": 91792896,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_start.output_min": {
    "atom_offset": 91792898,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.self_attn.q_norm.weight": {
    "atom_offset": 91792900,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_start.input_max": {
    "atom_offset": 91792909,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.k_proj.output_min": {
    "atom_offset": 91792911,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.self_attn.q_norm.weight": {
    "atom_offset": 91792913,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.post_feedforward_layernorm.weight": {
    "atom_offset": 91792930,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.mlp.up_proj.weight": {
    "atom_offset": 91792979,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.00370632391422987,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.v_proj.linear.weight": {
    "atom_offset": 92087892,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0030757873319089413,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.q_proj.output_max": {
    "atom_offset": 92097109,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.conv_norm.weight": {
    "atom_offset": 92097111,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.input_layernorm.weight": {
    "atom_offset": 92097144,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.gate_proj.output_min": {
    "atom_offset": 92097169,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.self_attn.q_norm.weight": {
    "atom_offset": 92097171,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.post_feedforward_layernorm.weight": {
    "atom_offset": 92097180,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 92097229,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.pre_feedforward_layernorm.weight": {
    "atom_offset": 92097231,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.mlp.down_proj.weight": {
    "atom_offset": 92097280,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.003383366158232093,
    "track": "int8"
  },
  "model.language_model.layers.1.per_layer_input_gate.weight": {
    "atom_offset": 92244737,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 92257026,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.031496062874794006,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.v_proj.input_max": {
    "atom_offset": 92257107,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 92257109,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.958600181154907e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.down_proj.linear.weight": {
    "atom_offset": 92322646,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0014840674120932817,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.k_norm.weight": {
    "atom_offset": 92359511,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.post_attention_layernorm.weight": {
    "atom_offset": 92359514,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.self_attn.o_proj.weight": {
    "atom_offset": 92359563,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.005290354136377573,
    "track": "int8"
  },
  "model.language_model.layers.30.post_attention_layernorm.weight": {
    "atom_offset": 92408716,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.relative_k_proj.weight": {
    "atom_offset": 92408765,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0019454355351626873,
    "track": "int8"
  },
  "model.language_model.layers.2.mlp.down_proj.weight": {
    "atom_offset": 92425150,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004090797156095505,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.v_proj.input_max": {
    "atom_offset": 92572607,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.o_proj.input_min": {
    "atom_offset": 92572609,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.per_layer_input_gate.weight": {
    "atom_offset": 92572611,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_start.output_min": {
    "atom_offset": 92584900,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.mlp.gate_proj.weight": {
    "atom_offset": 92584902,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.006397638004273176,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 92732359,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_end.input_min": {
    "atom_offset": 92732392,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.subsample_conv_projection.layer0.norm.weight": {
    "atom_offset": 92732394,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.q_proj.output_min": {
    "atom_offset": 92732399,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.gate_proj.output_min": {
    "atom_offset": 92732401,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.gate_proj.output_max": {
    "atom_offset": 92732403,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.self_attn.q_proj.weight": {
    "atom_offset": 92732405,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004213828593492508,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 92781558,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.612573681399226e-4,
    "track": "int8"
  },
  "model.language_model.embed_tokens.weight": {
    "atom_offset": 92847095,
    "byte_length": 805306432,
    "num_atoms": 12582913,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.o_proj.output_min": {
    "atom_offset": 105430008,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.o_proj.output_max": {
    "atom_offset": 105430010,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.output_proj.weight": {
    "atom_offset": 105430012,
    "byte_length": 1572928,
    "num_atoms": 24577,
    "scale": 0.001999261789023876,
    "track": "int8"
  },
  "model.language_model.layers.21.mlp.gate_proj.weight": {
    "atom_offset": 105454589,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005044291261583567,
    "track": "int8"
  },
  "model.language_model.layers.3.post_feedforward_layernorm.weight": {
    "atom_offset": 105749502,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.q_proj.linear.weight": {
    "atom_offset": 105749551,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0024144931230694056,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 105758768,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 105758770,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.k_proj.linear.weight": {
    "atom_offset": 105758772,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.612573681399226e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.q_proj.output_min": {
    "atom_offset": 105775157,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_start.linear.weight": {
    "atom_offset": 105775159,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.001253383350558579,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.k_proj.input_min": {
    "atom_offset": 105807928,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.k_proj.output_max": {
    "atom_offset": 105807930,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 105807932,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.norm_out.weight": {
    "atom_offset": 105807934,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.21358267962932587,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.q_proj.input_min": {
    "atom_offset": 105807951,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 105807953,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.805179873481393e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.post_feedforward_layernorm.weight": {
    "atom_offset": 105873490,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.k_proj.linear.weight": {
    "atom_offset": 105873515,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.074311142787337e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.k_proj.output_max": {
    "atom_offset": 105889900,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.v_proj.output_max": {
    "atom_offset": 105889902,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.q_proj.input_min": {
    "atom_offset": 105889904,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.per_layer_projection.weight": {
    "atom_offset": 105889906,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.post_per_layer_input_norm.weight": {
    "atom_offset": 105902195,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.q_proj.linear.weight": {
    "atom_offset": 105902244,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.00119955709669739,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_start.linear.weight": {
    "atom_offset": 105918629,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.0010919044725596905,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.v_proj.output_min": {
    "atom_offset": 105951398,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.post_attention_layernorm.weight": {
    "atom_offset": 105951400,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.post.output_min": {
    "atom_offset": 105951449,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.mlp.up_proj.weight": {
    "atom_offset": 105951451,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0036448079627007246,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.gate_proj.output_min": {
    "atom_offset": 106098908,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.up_proj.output_min": {
    "atom_offset": 106098910,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.v_proj.output_max": {
    "atom_offset": 106098912,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.post.input_max": {
    "atom_offset": 106098914,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.self_attn.k_norm.weight": {
    "atom_offset": 106098916,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.post.input_max": {
    "atom_offset": 106098925,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.gate_proj.input_min": {
    "atom_offset": 106098927,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.self_attn.q_proj.weight": {
    "atom_offset": 106098929,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0031219241209328175,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 106148082,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.151205791160464e-4,
    "track": "int8"
  },
  "model.language_model.layers.16.self_attn.q_norm.weight": {
    "atom_offset": 106213619,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.self_attn.o_proj.weight": {
    "atom_offset": 106213628,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.005228838417679071,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 106262781,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.497231708839536e-4,
    "track": "int8"
  },
  "model.language_model.layers.7.layer_scalar": {
    "atom_offset": 106328318,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_1.linear.weight": {
    "atom_offset": 106328320,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.613312241621315e-4,
    "track": "int8"
  },
  "model.language_model.layers.8.mlp.gate_proj.weight": {
    "atom_offset": 106393857,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.006182332523167133,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.relative_k_proj.weight": {
    "atom_offset": 106541314,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0018377830274403095,
    "track": "int8"
  },
  "model.language_model.layers.8.per_layer_projection.weight": {
    "atom_offset": 106557699,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 106569988,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.q_proj.output_min": {
    "atom_offset": 106569990,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.self_attn.v_proj.weight": {
    "atom_offset": 106569992,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.3445496703498065e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 106576137,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.612573681399226e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.q_proj.linear.weight": {
    "atom_offset": 106641674,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0017147514736279845,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.up_proj.output_max": {
    "atom_offset": 106658059,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.gate_proj.output_max": {
    "atom_offset": 106658061,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 106658063,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.post_per_layer_input_norm.weight": {
    "atom_offset": 106658065,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.self_attn.o_proj.weight": {
    "atom_offset": 106658114,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.language_model.layers.26.mlp.gate_proj.weight": {
    "atom_offset": 106707267,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0049212598241865635,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 107002180,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.layer_scalar": {
    "atom_offset": 107002213,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 107002215,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.o_proj.input_max": {
    "atom_offset": 107002217,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.k_proj.linear.weight": {
    "atom_offset": 107002219,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.958230610005558e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.2.norm_out.weight": {
    "atom_offset": 107018604,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.10777559131383896,
    "track": "int8"
  },
  "model.language_model.layers.1.self_attn.q_proj.weight": {
    "atom_offset": 107018621,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.00430610217154026,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.down_proj.input_min": {
    "atom_offset": 107067774,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.v_proj.output_min": {
    "atom_offset": 107067776,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.mlp.gate_proj.weight": {
    "atom_offset": 107067778,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004859744105488062,
    "track": "int8"
  },
  "model.language_model.layers.7.post_per_layer_input_norm.weight": {
    "atom_offset": 107362691,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.q_norm.weight": {
    "atom_offset": 107362740,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.self_attn.v_proj.weight": {
    "atom_offset": 107362743,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.594457568600774e-4,
    "track": "int8"
  },
  "model.language_model.layers.31.self_attn.k_proj.weight": {
    "atom_offset": 107375032,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.748246865347028e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.q_proj.output_min": {
    "atom_offset": 107381177,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 107381179,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 107381181,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.536417593248188e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 107446718,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 107446751,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 107446753,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.self_attn.q_norm.weight": {
    "atom_offset": 107446786,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.relative_k_proj.weight": {
    "atom_offset": 107446795,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014609990175813437,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.k_proj.input_max": {
    "atom_offset": 107463180,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_start.input_min": {
    "atom_offset": 107463182,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.down_proj.input_max": {
    "atom_offset": 107463184,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.k_proj.input_max": {
    "atom_offset": 107463186,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.post_per_layer_input_norm.weight": {
    "atom_offset": 107463188,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.mlp.down_proj.weight": {
    "atom_offset": 107463237,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005105806980282068,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.gate_proj.linear.weight": {
    "atom_offset": 107758150,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.003352608298882842,
    "track": "int8"
  },
  "model.language_model.layers.21.layer_scalar": {
    "atom_offset": 107795015,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.self_attn.v_proj.weight": {
    "atom_offset": 107795017,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0031526819802820683,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 107801162,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.down_proj.output_min": {
    "atom_offset": 107801164,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 107801166,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.per_layer_input_gate.weight": {
    "atom_offset": 107801168,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.o_proj.output_max": {
    "atom_offset": 107813457,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.gate_proj.linear.weight": {
    "atom_offset": 107813459,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0026144192088395357,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 107850324,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.343811401166022e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 107915861,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.self_attn.q_norm.weight": {
    "atom_offset": 107915863,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 107915872,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.k_norm.weight": {
    "atom_offset": 107915874,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.per_dim_scale": {
    "atom_offset": 107915877,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.per_layer_input_gate.weight": {
    "atom_offset": 107915882,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 107928171,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.self_attn.q_proj.weight": {
    "atom_offset": 107928173,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0034910186659544706,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.post.output_max": {
    "atom_offset": 107977326,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 107977328,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.88244351092726e-4,
    "track": "int8"
  },
  "model.language_model.layers.31.mlp.up_proj.weight": {
    "atom_offset": 108042865,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.002983513753861189,
    "track": "int8"
  },
  "model.language_model.per_layer_projection_norm.weight": {
    "atom_offset": 108337778,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.mlp.gate_proj.weight": {
    "atom_offset": 108337787,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0038447342813014984,
    "track": "int8"
  },
  "model.language_model.layers.31.layer_scalar": {
    "atom_offset": 108632700,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 108632702,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.norm_pre_attn.weight": {
    "atom_offset": 108632704,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2696850299835205,
    "track": "int8"
  },
  "model.language_model.layers.3.per_layer_projection.weight": {
    "atom_offset": 108632721,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 108645010,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.v_proj.input_max": {
    "atom_offset": 108645043,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.k_proj.output_max": {
    "atom_offset": 108645045,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.q_proj.input_max": {
    "atom_offset": 108645047,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_start.linear.weight": {
    "atom_offset": 108645049,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.0016609252197667956,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.o_proj.output_max": {
    "atom_offset": 108677818,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.q_proj.input_max": {
    "atom_offset": 108677820,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.self_attn.k_norm.weight": {
    "atom_offset": 108677822,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 108677839,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.self_attn.k_norm.weight": {
    "atom_offset": 108677841,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 108677858,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.mlp.up_proj.weight": {
    "atom_offset": 108677860,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.003198819002136588,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.up_proj.output_max": {
    "atom_offset": 108972773,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.post_feedforward_layernorm.weight": {
    "atom_offset": 108972775,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_end.linear.weight": {
    "atom_offset": 108972824,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.76599398907274e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.post.output_min": {
    "atom_offset": 108989209,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 108989211,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.pre_feedforward_layernorm.weight": {
    "atom_offset": 108989213,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.post_per_layer_input_norm.weight": {
    "atom_offset": 108989262,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.mlp.gate_proj.weight": {
    "atom_offset": 108989311,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.005597933195531368,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 109136768,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_end.linear.weight": {
    "atom_offset": 109136770,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.99630894511938e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 109153155,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.layer_scalar": {
    "atom_offset": 109153157,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.q_proj.input_max": {
    "atom_offset": 109153159,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.k_proj.input_min": {
    "atom_offset": 109153161,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.down_proj.output_max": {
    "atom_offset": 109153163,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.mlp.up_proj.weight": {
    "atom_offset": 109153165,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004644439090043306,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.gate_proj.input_min": {
    "atom_offset": 109448078,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.self_attn.k_norm.weight": {
    "atom_offset": 109448080,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.self_attn.q_norm.weight": {
    "atom_offset": 109448089,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.v_proj.output_max": {
    "atom_offset": 109448106,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.q_proj.input_min": {
    "atom_offset": 109448108,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.k_proj.output_max": {
    "atom_offset": 109448110,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.gate_proj.linear.weight": {
    "atom_offset": 109448112,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0023375984746962786,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.k_proj.input_max": {
    "atom_offset": 109484977,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.k_proj.input_min": {
    "atom_offset": 109484979,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.self_attn.k_proj.weight": {
    "atom_offset": 109484981,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0031834400724619627,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.gate_proj.output_min": {
    "atom_offset": 109491126,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 109491128,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.self_attn.k_norm.weight": {
    "atom_offset": 109491130,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.post.linear.weight": {
    "atom_offset": 109491139,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.688730351626873e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.post.input_min": {
    "atom_offset": 109507524,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.up_proj.output_max": {
    "atom_offset": 109507526,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.v_proj.linear.weight": {
    "atom_offset": 109507528,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002183808945119381,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.post_feedforward_layernorm.weight": {
    "atom_offset": 109516745,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 109516770,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.gate_proj.output_min": {
    "atom_offset": 109516803,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.post_attention_layernorm.weight": {
    "atom_offset": 109516805,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 109516854,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 109516856,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.subsample_conv_projection.layer1.conv.weight": {
    "atom_offset": 109516858,
    "byte_length": 36928,
    "num_atoms": 577,
    "scale": 0.002245324896648526,
    "track": "int8"
  },
  "model.language_model.layers.34.per_layer_projection.weight": {
    "atom_offset": 109517435,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.mlp.up_proj.weight": {
    "atom_offset": 109529724,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.003198819002136588,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.gate_proj.input_min": {
    "atom_offset": 109677181,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.gate_proj.linear.weight": {
    "atom_offset": 109677183,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0024144931230694056,
    "track": "int8"
  },
  "model.language_model.layers.33.input_layernorm.weight": {
    "atom_offset": 109714048,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.input_layernorm.weight": {
    "atom_offset": 109714097,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.per_layer_projection.weight": {
    "atom_offset": 109714146,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 109726435,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.01833169348537922,
    "track": "int8"
  },
  "model.language_model.layers.25.self_attn.v_proj.weight": {
    "atom_offset": 109726516,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.6136812306940556e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.gate_proj.input_max": {
    "atom_offset": 109732661,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.conv_norm.weight": {
    "atom_offset": 109732663,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.per_layer_projection.weight": {
    "atom_offset": 109732696,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.per_layer_input_gate.weight": {
    "atom_offset": 109744985,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.q_norm.weight": {
    "atom_offset": 109757274,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.post_per_layer_input_norm.weight": {
    "atom_offset": 109757277,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.o_proj.output_max": {
    "atom_offset": 109757326,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.v_proj.linear.weight": {
    "atom_offset": 109757328,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.689837318845093e-4,
    "track": "int8"
  },
  "model.language_model.layers.22.layer_scalar": {
    "atom_offset": 109773713,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.gate_proj.linear.weight": {
    "atom_offset": 109773715,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0025682824198156595,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.q_proj.input_min": {
    "atom_offset": 109810580,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.post.output_min": {
    "atom_offset": 109810582,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.q_proj.output_max": {
    "atom_offset": 109810584,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.gate_proj.input_min": {
    "atom_offset": 109810586,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.k_proj.input_max": {
    "atom_offset": 109810588,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.k_proj.output_max": {
    "atom_offset": 109810590,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.k_proj.linear.weight": {
    "atom_offset": 109810592,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010150098241865635,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.q_proj.linear.weight": {
    "atom_offset": 109826977,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.536048022098839e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.10.norm_out.weight": {
    "atom_offset": 109843362,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.218503937125206,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.q_proj.linear.weight": {
    "atom_offset": 109843379,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.728285225108266e-4,
    "track": "int8"
  },
  "model.language_model.layers.11.self_attn.k_proj.weight": {
    "atom_offset": 109859764,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.002583661349490285,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.v_proj.output_max": {
    "atom_offset": 109865909,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.q_proj.output_max": {
    "atom_offset": 109865911,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.v_proj.linear.weight": {
    "atom_offset": 109865913,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002322219545021653,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.v_proj.output_max": {
    "atom_offset": 109875130,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.k_proj.input_max": {
    "atom_offset": 109875132,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.layer_scalar": {
    "atom_offset": 109875134,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.down_proj.input_max": {
    "atom_offset": 109875136,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 109875138,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.self_attn.o_proj.weight": {
    "atom_offset": 109875171,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0042753443121910095,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.conv_norm.weight": {
    "atom_offset": 109924324,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.down_proj.output_min": {
    "atom_offset": 109924357,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.mlp.up_proj.weight": {
    "atom_offset": 109924359,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.00565944891422987,
    "track": "int8"
  },
  "model.language_model.layers.23.self_attn.q_norm.weight": {
    "atom_offset": 110219272,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.q_proj.output_max": {
    "atom_offset": 110219281,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.input_layernorm.weight": {
    "atom_offset": 110219283,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.o_proj.input_min": {
    "atom_offset": 110219332,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 110219334,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 110219367,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.self_attn.k_proj.weight": {
    "atom_offset": 110219369,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0026913140900433064,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.down_proj.input_max": {
    "atom_offset": 110225514,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.gate_proj.output_max": {
    "atom_offset": 110225516,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_end.input_min": {
    "atom_offset": 110225518,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.self_attn.k_proj.weight": {
    "atom_offset": 110225520,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.6329048927873373e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.v_proj.output_min": {
    "atom_offset": 110231665,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.post_feedforward_layernorm.weight": {
    "atom_offset": 110231667,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.up_proj.input_max": {
    "atom_offset": 110231716,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.post_feedforward_layernorm.weight": {
    "atom_offset": 110231718,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.mlp.up_proj.weight": {
    "atom_offset": 110231767,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005259596277028322,
    "track": "int8"
  },
  "model.language_model.layers.9.post_per_layer_input_norm.weight": {
    "atom_offset": 110526680,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.mlp.down_proj.weight": {
    "atom_offset": 110526729,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.005259596277028322,
    "track": "int8"
  },
  "model.language_model.layers.14.layer_scalar": {
    "atom_offset": 110674186,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 110674188,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.self_attn.k_proj.weight": {
    "atom_offset": 110674221,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.479115305002779e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.v_proj.linear.weight": {
    "atom_offset": 110680366,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003060408402234316,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.up_proj.input_max": {
    "atom_offset": 110689583,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.k_proj.output_max": {
    "atom_offset": 110689585,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.down_proj.output_min": {
    "atom_offset": 110689587,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.k_proj.output_min": {
    "atom_offset": 110689589,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 110689591,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.k_proj.linear.weight": {
    "atom_offset": 110689624,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003614050103351474,
    "track": "int8"
  },
  "model.language_model.layers.33.post_per_layer_input_norm.weight": {
    "atom_offset": 110698841,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.q_proj.output_min": {
    "atom_offset": 110698890,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.per_layer_projection.weight": {
    "atom_offset": 110698892,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.self_attn.v_proj.weight": {
    "atom_offset": 110711181,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.786694189533591e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.q_proj.input_min": {
    "atom_offset": 110717326,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.post.input_max": {
    "atom_offset": 110717328,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.gate_proj.output_min": {
    "atom_offset": 110717330,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.v_proj.linear.weight": {
    "atom_offset": 110717332,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0012380044208839536,
    "track": "int8"
  },
  "model.language_model.layers.19.self_attn.q_proj.weight": {
    "atom_offset": 110733717,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.005444143898785114,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.k_proj.linear.weight": {
    "atom_offset": 110832022,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003275713650509715,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 110841239,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 110841241,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.574126357212663e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.pre_feedforward_layernorm.weight": {
    "atom_offset": 110906778,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 110906803,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_end.input_max": {
    "atom_offset": 110906805,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.up_proj.linear.weight": {
    "atom_offset": 110906807,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.00239911419339478,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.down_proj.linear.weight": {
    "atom_offset": 110943672,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0012380044208839536,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.q_proj.input_min": {
    "atom_offset": 110980537,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.self_attn.q_proj.weight": {
    "atom_offset": 110980539,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003967765718698502,
    "track": "int8"
  },
  "model.language_model.layers.3.self_attn.o_proj.weight": {
    "atom_offset": 111029692,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.005198080558329821,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.o_proj.linear.weight": {
    "atom_offset": 111078845,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003229576861485839,
    "track": "int8"
  },
  "model.language_model.layers.9.mlp.down_proj.weight": {
    "atom_offset": 111088062,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0035217765253037214,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.q_norm.weight": {
    "atom_offset": 111235519,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_start.input_min": {
    "atom_offset": 111235522,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.norm_out.weight": {
    "atom_offset": 111235524,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.15452755987644196,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.k_norm.weight": {
    "atom_offset": 111235541,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 111235544,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.o_proj.linear.weight": {
    "atom_offset": 111235546,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0037217028439044952,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_end.output_min": {
    "atom_offset": 111244763,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.v_proj.output_min": {
    "atom_offset": 111244765,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.up_proj.input_min": {
    "atom_offset": 111244767,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.up_proj.output_max": {
    "atom_offset": 111244769,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 111244771,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 111244773,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.layer_scalar": {
    "atom_offset": 111244775,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.q_proj.linear.weight": {
    "atom_offset": 111244777,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0042753443121910095,
    "track": "int8"
  },
  "model.vision_tower.patch_embedder.input_proj.weight": {
    "atom_offset": 111253994,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002660555997863412,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 111263211,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.q_proj.linear.weight": {
    "atom_offset": 111263213,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.00366018689237535,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.k_proj.input_min": {
    "atom_offset": 111272430,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.self_attn.v_proj.weight": {
    "atom_offset": 111272432,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0030142716132104397,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.k_proj.linear.weight": {
    "atom_offset": 111278577,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003275713650509715,
    "track": "int8"
  },
  "model.audio_tower.layers.1.norm_post_attn.weight": {
    "atom_offset": 111287794,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.08415354043245316,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.o_proj.linear.weight": {
    "atom_offset": 111287811,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002752829808741808,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.o_proj.output_min": {
    "atom_offset": 111297028,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 111297030,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.conv_norm.weight": {
    "atom_offset": 111297032,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.k_proj.output_min": {
    "atom_offset": 111297065,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_start.linear.weight": {
    "atom_offset": 111297067,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.001122662448324263,
    "track": "int8"
  },
  "model.language_model.layers.22.mlp.gate_proj.weight": {
    "atom_offset": 111329836,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.post_attention_layernorm.weight": {
    "atom_offset": 111624749,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.q_proj.output_min": {
    "atom_offset": 111624774,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.self_attn.o_proj.weight": {
    "atom_offset": 111624776,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.0035525343846529722,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 111723081,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.up_proj.input_max": {
    "atom_offset": 111723083,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.self_attn.o_proj.weight": {
    "atom_offset": 111723085,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0035986711736768484,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.post_feedforward_layernorm.weight": {
    "atom_offset": 111772238,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 111772263,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 111772265,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.conv_norm.weight": {
    "atom_offset": 111772298,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_1.input_min": {
    "atom_offset": 111772331,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.self_attn.q_norm.weight": {
    "atom_offset": 111772333,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.q_proj.linear.weight": {
    "atom_offset": 111772342,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0031834400724619627,
    "track": "int8"
  },
  "model.language_model.layers.32.self_attn.q_proj.weight": {
    "atom_offset": 111781559,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003383366158232093,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.k_proj.input_min": {
    "atom_offset": 111830712,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 111830714,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.o_proj.linear.weight": {
    "atom_offset": 111830716,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0027682087384164333,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 111839933,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.mlp.up_proj.weight": {
    "atom_offset": 111839935,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0043983762152493,
    "track": "int8"
  },
  "model.language_model.layers.7.per_layer_projection.weight": {
    "atom_offset": 112134848,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.k_proj.output_max": {
    "atom_offset": 112147137,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_end.input_max": {
    "atom_offset": 112147139,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.q_proj.output_min": {
    "atom_offset": 112147141,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.self_attn.o_proj.weight": {
    "atom_offset": 112147143,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004367617890238762,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.relative_k_proj.weight": {
    "atom_offset": 112196296,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0023375984746962786,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.up_proj.output_min": {
    "atom_offset": 112212681,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.per_layer_input_gate.weight": {
    "atom_offset": 112212683,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.q_proj.input_min": {
    "atom_offset": 112224972,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.down_proj.input_max": {
    "atom_offset": 112224974,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.mlp.down_proj.weight": {
    "atom_offset": 112224976,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004213828593492508,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.post.output_max": {
    "atom_offset": 112519889,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.v_proj.output_min": {
    "atom_offset": 112519891,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.self_attn.o_proj.weight": {
    "atom_offset": 112519893,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0033218504395335913,
    "track": "int8"
  },
  "model.language_model.layers.12.mlp.up_proj.weight": {
    "atom_offset": 112569046,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.003998523578047752,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.k_proj.input_max": {
    "atom_offset": 112716503,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.post_attention_layernorm.weight": {
    "atom_offset": 112716505,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.down_proj.input_max": {
    "atom_offset": 112716530,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.self_attn.v_proj.weight": {
    "atom_offset": 112716532,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.7097995411604643e-4,
    "track": "int8"
  },
  "model.language_model.layers.4.pre_feedforward_layernorm.weight": {
    "atom_offset": 112728821,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.down_proj.linear.weight": {
    "atom_offset": 112728870,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0015378936659544706,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 112765735,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_start.linear.weight": {
    "atom_offset": 112765737,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.001076525542885065,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.v_proj.input_min": {
    "atom_offset": 112798506,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.q_proj.input_max": {
    "atom_offset": 112798508,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.down_proj.input_max": {
    "atom_offset": 112798510,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.self_attn.k_proj.weight": {
    "atom_offset": 112798512,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.805917851626873e-4,
    "track": "int8"
  },
  "model.language_model.layers.8.per_layer_input_gate.weight": {
    "atom_offset": 112804657,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_start.input_min": {
    "atom_offset": 112816946,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.q_proj.linear.weight": {
    "atom_offset": 112816948,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0027682087384164333,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.k_proj.input_max": {
    "atom_offset": 112826165,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 112826167,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.689468329772353e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.v_proj.linear.weight": {
    "atom_offset": 112891704,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 5.99778548348695e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 112908089,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.v_proj.output_min": {
    "atom_offset": 112908091,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 112908093,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_end.input_min": {
    "atom_offset": 112908126,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.gate_proj.input_max": {
    "atom_offset": 112908128,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 112908130,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.input_layernorm.weight": {
    "atom_offset": 112908163,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.mlp.up_proj.weight": {
    "atom_offset": 112908212,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006459153722971678,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.post.linear.weight": {
    "atom_offset": 113203125,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.001053457148373127,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.post_feedforward_layernorm.weight": {
    "atom_offset": 113219510,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_end.input_min": {
    "atom_offset": 113219535,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.mlp.down_proj.weight": {
    "atom_offset": 113219537,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.language_model.layers.10.self_attn.q_proj.weight": {
    "atom_offset": 113514450,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0041215550154447556,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 113563603,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 113563636,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.down_proj.linear.weight": {
    "atom_offset": 113563638,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0015225147362798452,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.gate_proj.output_min": {
    "atom_offset": 113600503,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.k_proj.linear.weight": {
    "atom_offset": 113600505,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0030911662615835667,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.k_proj.input_min": {
    "atom_offset": 113609722,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.k_proj.linear.weight": {
    "atom_offset": 113609724,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.227362461388111e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.pre_feedforward_layernorm.weight": {
    "atom_offset": 113626109,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.pre_feedforward_layernorm.weight": {
    "atom_offset": 113626134,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.mlp.gate_proj.weight": {
    "atom_offset": 113626183,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006028543226420879,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.v_proj.output_min": {
    "atom_offset": 113921096,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.post.output_max": {
    "atom_offset": 113921098,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.self_attn.q_norm.weight": {
    "atom_offset": 113921100,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.down_proj.input_min": {
    "atom_offset": 113921109,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.post_feedforward_layernorm.weight": {
    "atom_offset": 113921111,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.pre_feedforward_layernorm.weight": {
    "atom_offset": 113921160,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.down_proj.input_max": {
    "atom_offset": 113921209,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.layer_scalar": {
    "atom_offset": 113921211,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.gate_proj.output_max": {
    "atom_offset": 113921213,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.self_attn.k_proj.weight": {
    "atom_offset": 113921215,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.4983389670960605e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.post.input_min": {
    "atom_offset": 113927360,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.v_proj.output_min": {
    "atom_offset": 113927362,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.per_dim_scale": {
    "atom_offset": 113927364,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_start.output_min": {
    "atom_offset": 113927369,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.self_attn.k_proj.weight": {
    "atom_offset": 113927371,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.6136812306940556e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.relative_k_proj.weight": {
    "atom_offset": 113933516,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0043983762152493,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 113949901,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.per_dim_scale": {
    "atom_offset": 113949903,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.mlp.up_proj.weight": {
    "atom_offset": 113949908,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.005259596277028322,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.up_proj.output_min": {
    "atom_offset": 114097365,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 114097367,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 114097400,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 114097433,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 114097435,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.per_dim_scale": {
    "atom_offset": 114097437,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.v_proj.output_max": {
    "atom_offset": 114097442,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.post_attention_layernorm.weight": {
    "atom_offset": 114097444,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.self_attn.k_proj.weight": {
    "atom_offset": 114097469,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0027682087384164333,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 114103614,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.self_attn.o_proj.weight": {
    "atom_offset": 114103616,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004552165511995554,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 114152769,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 114152802,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 114152804,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.q_proj.input_min": {
    "atom_offset": 114152806,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 114152808,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 5.267285741865635e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.v_proj.input_min": {
    "atom_offset": 114218345,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.norm_pre_attn.weight": {
    "atom_offset": 114218347,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2303149551153183,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.up_proj.input_min": {
    "atom_offset": 114218364,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.q_proj.input_max": {
    "atom_offset": 114218366,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.norm_pre_attn.weight": {
    "atom_offset": 114218368,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.30314961075782776,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.v_proj.linear.weight": {
    "atom_offset": 114218385,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014225516933947802,
    "track": "int8"
  },
  "model.language_model.layers.30.self_attn.q_proj.weight": {
    "atom_offset": 114234770,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0033987450879067183,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.q_proj.linear.weight": {
    "atom_offset": 114283923,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.881335961632431e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.post.input_min": {
    "atom_offset": 114300308,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.q_proj.output_max": {
    "atom_offset": 114300310,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.v_proj.input_max": {
    "atom_offset": 114300312,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.post.output_max": {
    "atom_offset": 114300314,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.v_proj.input_min": {
    "atom_offset": 114300316,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.post.output_min": {
    "atom_offset": 114300318,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 114300320,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.gate_proj.linear.weight": {
    "atom_offset": 114300353,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0027066930197179317,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.post_attention_layernorm.weight": {
    "atom_offset": 114337218,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.gate_proj.output_min": {
    "atom_offset": 114337243,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.o_proj.input_max": {
    "atom_offset": 114337245,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.k_proj.linear.weight": {
    "atom_offset": 114337247,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.766732549294829e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.per_dim_scale": {
    "atom_offset": 114353632,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.post_feedforward_layernorm.weight": {
    "atom_offset": 114353637,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.k_proj.input_max": {
    "atom_offset": 114353662,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.self_attn.k_norm.weight": {
    "atom_offset": 114353664,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.k_proj.output_min": {
    "atom_offset": 114353673,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.pre_feedforward_layernorm.weight": {
    "atom_offset": 114353675,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.v_proj.input_min": {
    "atom_offset": 114353724,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.up_proj.output_max": {
    "atom_offset": 114353726,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.q_proj.input_min": {
    "atom_offset": 114353728,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.post.linear.weight": {
    "atom_offset": 114353730,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0011918676318600774,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 114370115,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 114370117,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.mlp.gate_proj.weight": {
    "atom_offset": 114370150,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0030911662615835667,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 114517607,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.input_layernorm.weight": {
    "atom_offset": 114517609,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.post_attention_layernorm.weight": {
    "atom_offset": 114517658,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.v_proj.input_min": {
    "atom_offset": 114517707,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.mlp.down_proj.weight": {
    "atom_offset": 114517709,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0055671753361821175,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.up_proj.input_min": {
    "atom_offset": 114812622,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_start.output_max": {
    "atom_offset": 114812624,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.per_layer_projection.weight": {
    "atom_offset": 114812626,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.q_proj.output_min": {
    "atom_offset": 114824915,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.input_layernorm.weight": {
    "atom_offset": 114824917,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 114824966,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.post.input_min": {
    "atom_offset": 114824999,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.self_attn.k_norm.weight": {
    "atom_offset": 114825001,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 114825010,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.q_proj.input_min": {
    "atom_offset": 114825012,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.input_layernorm.weight": {
    "atom_offset": 114825014,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_end.input_max": {
    "atom_offset": 114825039,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.per_dim_scale": {
    "atom_offset": 114825041,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.self_attn.k_proj.weight": {
    "atom_offset": 114825046,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.003306471509858966,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_end.output_max": {
    "atom_offset": 114831191,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.up_proj.input_min": {
    "atom_offset": 114831193,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.down_proj.input_max": {
    "atom_offset": 114831195,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.v_proj.output_min": {
    "atom_offset": 114831197,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.gate_proj.linear.weight": {
    "atom_offset": 114831199,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0029373769648373127,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.q_proj.output_min": {
    "atom_offset": 114868064,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_start.output_max": {
    "atom_offset": 114868066,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.input_layernorm.weight": {
    "atom_offset": 114868068,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.k_proj.linear.weight": {
    "atom_offset": 114868093,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.002798966597765684,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.o_proj.output_max": {
    "atom_offset": 114877310,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.o_proj.input_max": {
    "atom_offset": 114877312,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.pre_feedforward_layernorm.weight": {
    "atom_offset": 114877314,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.k_proj.output_max": {
    "atom_offset": 114877363,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.self_attn.v_proj.weight": {
    "atom_offset": 114877365,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0032141979318112135,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.k_proj.output_min": {
    "atom_offset": 114883510,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.post.output_min": {
    "atom_offset": 114883512,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 114883514,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.q_proj.linear.weight": {
    "atom_offset": 114883516,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.458784384652972e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.v_proj.input_max": {
    "atom_offset": 114899901,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.per_layer_projection.weight": {
    "atom_offset": 114899903,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.k_proj.output_min": {
    "atom_offset": 114912192,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.o_proj.input_max": {
    "atom_offset": 114912194,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.o_proj.output_min": {
    "atom_offset": 114912196,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.linear_end.linear.weight": {
    "atom_offset": 114912198,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.496493730694056e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.post_feedforward_layernorm.weight": {
    "atom_offset": 114928583,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 114928608,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_end.output_max": {
    "atom_offset": 114928610,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.v_proj.output_max": {
    "atom_offset": 114928612,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.k_proj.output_max": {
    "atom_offset": 114928614,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.k_proj.output_max": {
    "atom_offset": 114928616,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.k_proj.output_max": {
    "atom_offset": 114928618,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.post_attention_layernorm.weight": {
    "atom_offset": 114928620,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.v_proj.input_max": {
    "atom_offset": 114928645,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.q_proj.output_min": {
    "atom_offset": 114928647,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.post_feedforward_layernorm.weight": {
    "atom_offset": 114928649,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.norm_post_attn.weight": {
    "atom_offset": 114928698,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.09596456587314606,
    "track": "int8"
  },
  "model.language_model.layers.16.pre_feedforward_layernorm.weight": {
    "atom_offset": 114928715,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.input_layernorm.weight": {
    "atom_offset": 114928764,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.mlp.gate_proj.weight": {
    "atom_offset": 114928813,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0042753443121910095,
    "track": "int8"
  },
  "model.language_model.layers.33.self_attn.v_proj.weight": {
    "atom_offset": 115076270,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.671352216973901e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.up_proj.linear.weight": {
    "atom_offset": 115082415,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002245324896648526,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 115119280,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.v_proj.output_max": {
    "atom_offset": 115119282,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.input_layernorm.weight": {
    "atom_offset": 115119284,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.q_proj.output_max": {
    "atom_offset": 115119333,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_end.linear.weight": {
    "atom_offset": 115119335,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0012149360263720155,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.k_proj.input_max": {
    "atom_offset": 115135720,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.gate_proj.output_max": {
    "atom_offset": 115135722,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_end.input_min": {
    "atom_offset": 115135724,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.k_proj.input_min": {
    "atom_offset": 115135726,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.input_layernorm.weight": {
    "atom_offset": 115135728,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.per_layer_input_gate.weight": {
    "atom_offset": 115135777,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.post_feedforward_layernorm.weight": {
    "atom_offset": 115148066,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.q_proj.output_max": {
    "atom_offset": 115148115,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.norm_post_attn.weight": {
    "atom_offset": 115148117,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.1761811077594757,
    "track": "int8"
  },
  "model.language_model.layers.2.layer_scalar": {
    "atom_offset": 115148134,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.pre_feedforward_layernorm.weight": {
    "atom_offset": 115148136,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.k_proj.output_min": {
    "atom_offset": 115148161,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.self_attn.k_norm.weight": {
    "atom_offset": 115148163,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 115148180,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.gate_proj.linear.weight": {
    "atom_offset": 115148213,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002983513753861189,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.q_proj.input_min": {
    "atom_offset": 115185078,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.16.self_attn.q_proj.weight": {
    "atom_offset": 115185080,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003352608298882842,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.k_proj.input_min": {
    "atom_offset": 115234233,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.post_feedforward_layernorm.weight": {
    "atom_offset": 115234235,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.q_proj.output_min": {
    "atom_offset": 115234284,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.q_proj.input_min": {
    "atom_offset": 115234286,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.post_feedforward_layernorm.weight": {
    "atom_offset": 115234288,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 115234313,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.up_proj.output_min": {
    "atom_offset": 115234315,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.q_proj.linear.weight": {
    "atom_offset": 115234317,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003429502947255969,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_start.linear.weight": {
    "atom_offset": 115243534,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.001253383350558579,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.gate_proj.input_max": {
    "atom_offset": 115276303,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 115276305,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.mlp.down_proj.weight": {
    "atom_offset": 115276307,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005905511789023876,
    "track": "int8"
  },
  "model.language_model.layers.18.mlp.down_proj.weight": {
    "atom_offset": 115571220,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.00366018689237535,
    "track": "int8"
  },
  "model.language_model.layers.29.pre_feedforward_layernorm.weight": {
    "atom_offset": 115866133,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.k_proj.input_min": {
    "atom_offset": 115866182,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.self_attn.v_proj.weight": {
    "atom_offset": 115866184,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0026144192088395357,
    "track": "int8"
  },
  "model.language_model.layers.14.post_feedforward_layernorm.weight": {
    "atom_offset": 115872329,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.relative_k_proj.weight": {
    "atom_offset": 115872378,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0017378198681399226,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.post_attention_layernorm.weight": {
    "atom_offset": 115888763,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.v_proj.input_max": {
    "atom_offset": 115888788,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_end.output_min": {
    "atom_offset": 115888790,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.norm_pre_attn.weight": {
    "atom_offset": 115888792,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.31496062874794006,
    "track": "int8"
  },
  "model.language_model.layers.19.post_attention_layernorm.weight": {
    "atom_offset": 115888809,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.input_layernorm.weight": {
    "atom_offset": 115888858,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.self_attn.q_norm.weight": {
    "atom_offset": 115888883,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.self_attn.k_norm.weight": {
    "atom_offset": 115888892,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.subsample_conv_projection.layer0.conv.weight": {
    "atom_offset": 115888901,
    "byte_length": 1216,
    "num_atoms": 19,
    "scale": 0.010457676835358143,
    "track": "int8"
  },
  "model.language_model.layers.11.input_layernorm.weight": {
    "atom_offset": 115888920,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.up_proj.output_max": {
    "atom_offset": 115888969,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.down_proj.output_min": {
    "atom_offset": 115888971,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.pre_feedforward_layernorm.weight": {
    "atom_offset": 115888973,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.o_proj.output_min": {
    "atom_offset": 115889022,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.v_proj.linear.weight": {
    "atom_offset": 115889024,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.190022104419768e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 115905409,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 8.419968071393669e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 115970946,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.down_proj.linear.weight": {
    "atom_offset": 115970948,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0014994463417679071,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.gate_proj.output_max": {
    "atom_offset": 116007813,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.pre_feedforward_layernorm.weight": {
    "atom_offset": 116007815,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 116007840,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_end.input_min": {
    "atom_offset": 116007842,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.per_layer_input_gate.weight": {
    "atom_offset": 116007844,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.ffw_layer_2.linear.weight": {
    "atom_offset": 116020133,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.036232807673514e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.10.norm_pre_attn.weight": {
    "atom_offset": 116085670,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2263779491186142,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.k_proj.input_max": {
    "atom_offset": 116085687,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.q_proj.input_max": {
    "atom_offset": 116085689,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.post_attention_layernorm.weight": {
    "atom_offset": 116085691,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.self_attn.k_proj.weight": {
    "atom_offset": 116085740,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.7097995411604643e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.k_norm.weight": {
    "atom_offset": 116098029,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.self_attn.q_norm.weight": {
    "atom_offset": 116098032,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.self_attn.q_proj.weight": {
    "atom_offset": 116098041,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004675196949392557,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.o_proj.input_min": {
    "atom_offset": 116147194,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.norm.weight": {
    "atom_offset": 116147196,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.up_proj.input_max": {
    "atom_offset": 116147245,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.post_feedforward_layernorm.weight": {
    "atom_offset": 116147247,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.o_proj.input_min": {
    "atom_offset": 116147296,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.v_proj.output_max": {
    "atom_offset": 116147298,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.post.input_max": {
    "atom_offset": 116147300,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.self_attn.k_norm.weight": {
    "atom_offset": 116147302,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.input_layernorm.weight": {
    "atom_offset": 116147311,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.layer_scalar": {
    "atom_offset": 116147336,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.k_proj.input_min": {
    "atom_offset": 116147338,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.k_proj.output_max": {
    "atom_offset": 116147340,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.pre_feedforward_layernorm.weight": {
    "atom_offset": 116147342,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 116147391,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 116147393,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 3.5563792334869504e-4,
    "track": "int8"
  },
  "model.language_model.layers.13.self_attn.k_proj.weight": {
    "atom_offset": 116212930,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0017762671923264861,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_end.input_max": {
    "atom_offset": 116219075,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.self_attn.o_proj.weight": {
    "atom_offset": 116219077,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003429502947255969,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 116268230,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.pre_feedforward_layernorm.weight": {
    "atom_offset": 116268263,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.k_proj.output_min": {
    "atom_offset": 116268312,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.down_proj.output_max": {
    "atom_offset": 116268314,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_end.linear.weight": {
    "atom_offset": 116268316,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014148622285574675,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.q_proj.input_max": {
    "atom_offset": 116284701,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.q_proj.input_max": {
    "atom_offset": 116284703,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_start.input_max": {
    "atom_offset": 116284705,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.gate_proj.output_min": {
    "atom_offset": 116284707,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.7.mlp.up_proj.weight": {
    "atom_offset": 116284709,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004828986246138811,
    "track": "int8"
  },
  "model.language_model.layers.33.post_attention_layernorm.weight": {
    "atom_offset": 116432166,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.k_proj.input_min": {
    "atom_offset": 116432215,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.v_proj.output_min": {
    "atom_offset": 116432217,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.post_attention_layernorm.weight": {
    "atom_offset": 116432219,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.post_attention_layernorm.weight": {
    "atom_offset": 116432268,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.post_per_layer_input_norm.weight": {
    "atom_offset": 116432317,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.mlp.down_proj.weight": {
    "atom_offset": 116432366,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004552165511995554,
    "track": "int8"
  },
  "model.language_model.layers.15.layer_scalar": {
    "atom_offset": 116727279,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_start.input_max": {
    "atom_offset": 116727281,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.down_proj.linear.weight": {
    "atom_offset": 116727283,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.00173013040330261,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 116764148,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.03863188996911049,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.q_proj.output_min": {
    "atom_offset": 116764229,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 116764231,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.30.post_feedforward_layernorm.weight": {
    "atom_offset": 116764233,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.self_attn.o_proj.weight": {
    "atom_offset": 116764282,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004060039296746254,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 116813435,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.self_attn.o_proj.weight": {
    "atom_offset": 116813437,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0024606299120932817,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.o_proj.output_min": {
    "atom_offset": 116862590,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.self_attn.k_norm.weight": {
    "atom_offset": 116862592,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 116862601,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.post_per_layer_input_norm.weight": {
    "atom_offset": 116862634,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.self_attn.v_proj.weight": {
    "atom_offset": 116862683,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.748246865347028e-4,
    "track": "int8"
  },
  "model.language_model.layers.20.input_layernorm.weight": {
    "atom_offset": 116868828,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.layer_scalar": {
    "atom_offset": 116868877,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.o_proj.input_min": {
    "atom_offset": 116868879,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.post_per_layer_input_norm.weight": {
    "atom_offset": 116868881,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.26.pre_feedforward_layernorm.weight": {
    "atom_offset": 116868930,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.post_per_layer_input_norm.weight": {
    "atom_offset": 116868979,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 116869028,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 8.689099340699613e-4,
    "track": "int8"
  },
  "model.language_model.layers.24.mlp.gate_proj.weight": {
    "atom_offset": 116934565,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005105806980282068,
    "track": "int8"
  },
  "model.language_model.layers.23.mlp.down_proj.weight": {
    "atom_offset": 117229478,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.0032910925801843405,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.q_proj.output_max": {
    "atom_offset": 117524391,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.post_attention_layernorm.weight": {
    "atom_offset": 117524393,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 117524442,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.up_proj.output_min": {
    "atom_offset": 117524444,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_1.input_max": {
    "atom_offset": 117524446,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.post.input_max": {
    "atom_offset": 117524448,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.layer_scalar": {
    "atom_offset": 117524450,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.up_proj.output_max": {
    "atom_offset": 117524452,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.post.linear.weight": {
    "atom_offset": 117524454,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010919044725596905,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_end.output_max": {
    "atom_offset": 117540839,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.q_proj.output_max": {
    "atom_offset": 117540841,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.self_attn.q_proj.weight": {
    "atom_offset": 117540843,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0043983762152493,
    "track": "int8"
  },
  "model.language_model.layers.15.mlp.gate_proj.weight": {
    "atom_offset": 117589996,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005351869855076075,
    "track": "int8"
  },
  "model.language_model.layers.26.post_attention_layernorm.weight": {
    "atom_offset": 117884909,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 117884958,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 117884991,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_end.linear.weight": {
    "atom_offset": 117884993,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0011841781670227647,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.o_proj.input_max": {
    "atom_offset": 117901378,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.post_per_layer_input_norm.weight": {
    "atom_offset": 117901380,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.v_proj.input_min": {
    "atom_offset": 117901429,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.15.per_layer_projection.weight": {
    "atom_offset": 117901431,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.o_proj.input_max": {
    "atom_offset": 117913720,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.q_norm.weight": {
    "atom_offset": 117913722,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.per_layer_input_gate.weight": {
    "atom_offset": 117913725,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.v_proj.input_max": {
    "atom_offset": 117926014,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.per_layer_input_gate.weight": {
    "atom_offset": 117926016,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.k_proj.output_min": {
    "atom_offset": 117938305,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.q_proj.input_min": {
    "atom_offset": 117938307,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.self_attn.q_proj.weight": {
    "atom_offset": 117938309,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.00366018689237535,
    "track": "int8"
  },
  "model.language_model.layers.6.mlp.up_proj.weight": {
    "atom_offset": 117987462,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.003967765718698502,
    "track": "int8"
  },
  "model.language_model.layers.30.per_layer_projection.weight": {
    "atom_offset": 118134919,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.post_per_layer_input_norm.weight": {
    "atom_offset": 118147208,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.pre_feedforward_layernorm.weight": {
    "atom_offset": 118147257,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.8.self_attn.o_proj.weight": {
    "atom_offset": 118147282,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003998523578047752,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 118196435,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.per_layer_projection.weight": {
    "atom_offset": 118196437,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.pre_feedforward_layernorm.weight": {
    "atom_offset": 118208726,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.k_norm.weight": {
    "atom_offset": 118208775,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 118208778,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.k_proj.linear.weight": {
    "atom_offset": 118208811,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.004336860030889511,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.mlp.gate_proj.input_min": {
    "atom_offset": 118218028,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.v_proj.linear.weight": {
    "atom_offset": 118218030,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0027682087384164333,
    "track": "int8"
  },
  "model.language_model.layers.3.per_layer_input_gate.weight": {
    "atom_offset": 118227247,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.11.mlp.down_proj.weight": {
    "atom_offset": 118239536,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0036448079627007246,
    "track": "int8"
  },
  "model.audio_tower.layers.6.lconv1d.conv_norm.weight": {
    "atom_offset": 118386993,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 118387026,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.mlp.down_proj.weight": {
    "atom_offset": 118387028,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006182332523167133,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.relative_k_proj.weight": {
    "atom_offset": 118681941,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010611466132104397,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.q_proj.output_max": {
    "atom_offset": 118698326,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.post_feedforward_layernorm.weight": {
    "atom_offset": 118698328,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.per_layer_projection.weight": {
    "atom_offset": 118698377,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.self_attn.k_proj.weight": {
    "atom_offset": 118710666,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 0.0021222932264208794,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_start.input_min": {
    "atom_offset": 118722955,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.v_proj.output_max": {
    "atom_offset": 118722957,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.k_norm.weight": {
    "atom_offset": 118722959,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 118722962,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.up_proj.output_max": {
    "atom_offset": 118722964,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.up_proj.input_max": {
    "atom_offset": 118722966,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.o_proj.input_min": {
    "atom_offset": 118722968,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.k_proj.input_min": {
    "atom_offset": 118722970,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.q_proj.output_min": {
    "atom_offset": 118722972,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.k_proj.output_min": {
    "atom_offset": 118722974,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.self_attn.k_norm.weight": {
    "atom_offset": 118722976,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.post.linear.weight": {
    "atom_offset": 118722985,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010303887538611889,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.q_proj.output_max": {
    "atom_offset": 118739370,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.o_proj.output_max": {
    "atom_offset": 118739372,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.k_proj.output_max": {
    "atom_offset": 118739374,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.v_proj.output_max": {
    "atom_offset": 118739376,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.k_proj.linear.weight": {
    "atom_offset": 118739378,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0027374508790671825,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.v_proj.output_max": {
    "atom_offset": 118748595,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.self_attn.q_proj.weight": {
    "atom_offset": 118748597,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.004060039296746254,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.post.input_min": {
    "atom_offset": 118846902,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.up_proj.linear.weight": {
    "atom_offset": 118846904,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0031680609099566936,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_start.output_min": {
    "atom_offset": 118883769,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.per_layer_projection.weight": {
    "atom_offset": 118883771,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.o_proj.output_min": {
    "atom_offset": 118896060,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.27.per_layer_input_gate.weight": {
    "atom_offset": 118896062,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 118908351,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.k_proj.output_max": {
    "atom_offset": 118908384,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.pre_feedforward_layernorm.weight": {
    "atom_offset": 118908386,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.o_proj.input_max": {
    "atom_offset": 118908411,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.gate_proj.input_max": {
    "atom_offset": 118908413,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_end.input_min": {
    "atom_offset": 118908415,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.self_attn.q_proj.weight": {
    "atom_offset": 118908417,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.0032603347208350897,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.down_proj.linear.weight": {
    "atom_offset": 119006722,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0017224409384652972,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 119043587,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.026697834953665733,
    "track": "int8"
  },
  "model.language_model.layers.15.self_attn.v_proj.weight": {
    "atom_offset": 119043668,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.3445496703498065e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_end.output_min": {
    "atom_offset": 119049813,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.down_proj.input_min": {
    "atom_offset": 119049815,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.self_attn.k_norm.weight": {
    "atom_offset": 119049817,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.pre_feedforward_layernorm.weight": {
    "atom_offset": 119049826,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.norm_post_attn.weight": {
    "atom_offset": 119049851,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.06594488024711609,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.post.linear.weight": {
    "atom_offset": 119049868,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.001807025051675737,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_start.linear.weight": {
    "atom_offset": 119066253,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.0012303149560466409,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 119099022,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.down_proj.input_min": {
    "atom_offset": 119099024,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.v_proj.input_min": {
    "atom_offset": 119099026,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.self_attn.post.linear.weight": {
    "atom_offset": 119099028,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 8.266178774647415e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_end.input_max": {
    "atom_offset": 119115413,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.post_attention_layernorm.weight": {
    "atom_offset": 119115415,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.per_layer_projection.weight": {
    "atom_offset": 119115464,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 119127753,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.gate_proj.output_min": {
    "atom_offset": 119127755,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_start.input_min": {
    "atom_offset": 119127757,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.up_proj.linear.weight": {
    "atom_offset": 119127759,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002106914296746254,
    "track": "int8"
  },
  "model.language_model.layers.7.self_attn.v_proj.weight": {
    "atom_offset": 119164624,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0034448818769305944,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_end.linear.weight": {
    "atom_offset": 119170769,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0012918306747451425,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.k_proj.output_max": {
    "atom_offset": 119187154,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.post.linear.weight": {
    "atom_offset": 119187156,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014379306230694056,
    "track": "int8"
  },
  "model.language_model.layers.28.self_attn.k_proj.weight": {
    "atom_offset": 119203541,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.729023203253746e-4,
    "track": "int8"
  },
  "model.language_model.layers.6.input_layernorm.weight": {
    "atom_offset": 119209686,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.gate_proj.input_min": {
    "atom_offset": 119209735,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 119209737,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.k_proj.output_min": {
    "atom_offset": 119209739,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.mlp.up_proj.weight": {
    "atom_offset": 119209741,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.003998523578047752,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.down_proj.output_min": {
    "atom_offset": 119357198,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.v_proj.input_max": {
    "atom_offset": 119357200,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.pre_feedforward_layernorm.weight": {
    "atom_offset": 119357202,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.mlp.gate_proj.weight": {
    "atom_offset": 119357251,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0032910925801843405,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 119504708,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.gate_proj.linear.weight": {
    "atom_offset": 119504741,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0022607038263231516,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.depthwise_conv1d.weight": {
    "atom_offset": 119541606,
    "byte_length": 5184,
    "num_atoms": 81,
    "scale": 0.034202754497528076,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.conv_norm.weight": {
    "atom_offset": 119541687,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 119541720,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.q_proj.input_max": {
    "atom_offset": 119541722,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_2.output_min": {
    "atom_offset": 119541724,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.per_layer_projection.weight": {
    "atom_offset": 119541726,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 119554015,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.post_layer_norm.weight": {
    "atom_offset": 119554017,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 119554050,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.843257626518607e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 119619587,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.self_attn.v_proj.weight": {
    "atom_offset": 119619620,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 0.002752829808741808,
    "track": "int8"
  },
  "model.language_model.layers.24.self_attn.k_proj.weight": {
    "atom_offset": 119631909,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 4.6905758790671825e-4,
    "track": "int8"
  },
  "model.language_model.layers.0.self_attn.k_proj.weight": {
    "atom_offset": 119644198,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0021684300154447556,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 119650343,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.31.self_attn.o_proj.weight": {
    "atom_offset": 119650345,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.00479822838678956,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.q_proj.linear.weight": {
    "atom_offset": 119699498,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0055671753361821175,
    "track": "int8"
  },
  "model.language_model.layers.33.layer_scalar": {
    "atom_offset": 119708715,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.self_attn.k_norm.weight": {
    "atom_offset": 119708717,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.self_attn.k_proj.weight": {
    "atom_offset": 119708734,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.00215305108577013,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 119714879,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.gate_proj.input_min": {
    "atom_offset": 119714881,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.self_attn.q_proj.output_min": {
    "atom_offset": 119714883,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.pre_feedforward_layernorm.weight": {
    "atom_offset": 119714885,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.norm_post_attn.weight": {
    "atom_offset": 119714934,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.24901574850082397,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.k_proj.output_min": {
    "atom_offset": 119714951,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 119714953,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.17.mlp.gate_proj.weight": {
    "atom_offset": 119714955,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.006397638004273176,
    "track": "int8"
  },
  "model.language_model.layers.27.post_attention_layernorm.weight": {
    "atom_offset": 120009868,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.self_attn.o_proj.weight": {
    "atom_offset": 120009917,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.009534941054880619,
    "track": "int8"
  },
  "model.language_model.layers.10.post_feedforward_layernorm.weight": {
    "atom_offset": 120108222,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_start.output_max": {
    "atom_offset": 120108271,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.v_proj.input_min": {
    "atom_offset": 120108273,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.k_proj.output_max": {
    "atom_offset": 120108275,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.pre_feedforward_layernorm.weight": {
    "atom_offset": 120108277,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.1.mlp.gate_proj.weight": {
    "atom_offset": 120108302,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0029988926835358143,
    "track": "int8"
  },
  "model.language_model.layers.31.post_per_layer_input_norm.weight": {
    "atom_offset": 120255759,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.mlp.gate_proj.weight": {
    "atom_offset": 120255808,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004521407652646303,
    "track": "int8"
  },
  "model.audio_tower.layers.1.feed_forward1.post_layer_norm.weight": {
    "atom_offset": 120403265,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.self_attn.o_proj.weight": {
    "atom_offset": 120403298,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0034448818769305944,
    "track": "int8"
  },
  "model.language_model.layers.13.self_attn.q_proj.weight": {
    "atom_offset": 120452451,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.0034756397362798452,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_start.input_max": {
    "atom_offset": 120501604,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.post.output_min": {
    "atom_offset": 120501606,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.gate_proj.input_max": {
    "atom_offset": 120501608,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.self_attn.k_norm.weight": {
    "atom_offset": 120501610,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 120501619,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.14.self_attn.o_proj.weight": {
    "atom_offset": 120501621,
    "byte_length": 6291520,
    "num_atoms": 98305,
    "scale": 0.00565944891422987,
    "track": "int8"
  },
  "model.language_model.layers.23.self_attn.k_proj.weight": {
    "atom_offset": 120599926,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.7674705274403095e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 120606071,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.lconv1d.linear_start.input_max": {
    "atom_offset": 120606073,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.q_proj.output_max": {
    "atom_offset": 120606075,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.gate_proj.input_max": {
    "atom_offset": 120606077,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.gate_proj.linear.weight": {
    "atom_offset": 120606079,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0025375245604664087,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.v_proj.input_max": {
    "atom_offset": 120642944,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.down_proj.output_max": {
    "atom_offset": 120642946,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.k_proj.input_max": {
    "atom_offset": 120642948,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.pre_layer_norm.weight": {
    "atom_offset": 120642950,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_1.output_max": {
    "atom_offset": 120642983,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.gate_proj.linear.weight": {
    "atom_offset": 120642985,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002429872052744031,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.k_proj.input_min": {
    "atom_offset": 120679850,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.q_proj.linear.weight": {
    "atom_offset": 120679852,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 6.99741649441421e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.k_proj.output_max": {
    "atom_offset": 120696237,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward1.ffw_layer_2.output_min": {
    "atom_offset": 120696239,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 120696241,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.pre_feedforward_layernorm.weight": {
    "atom_offset": 120696243,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.gate_proj.output_max": {
    "atom_offset": 120696268,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.post.linear.weight": {
    "atom_offset": 120696270,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0010150098241865635,
    "track": "int8"
  },
  "model.language_model.layers.6.per_layer_projection.weight": {
    "atom_offset": 120712655,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.33.self_attn.q_proj.weight": {
    "atom_offset": 120724944,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003875492140650749,
    "track": "int8"
  },
  "model.language_model.layers.7.self_attn.q_norm.weight": {
    "atom_offset": 120774097,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.34.post_feedforward_layernorm.weight": {
    "atom_offset": 120774106,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.v_proj.input_min": {
    "atom_offset": 120774155,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.k_proj.input_min": {
    "atom_offset": 120774157,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.relative_k_proj.weight": {
    "atom_offset": 120774159,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0014456200879067183,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.q_proj.input_min": {
    "atom_offset": 120790544,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.norm_post_attn.weight": {
    "atom_offset": 120790546,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.11171259731054306,
    "track": "int8"
  },
  "model.language_model.layers.14.per_layer_input_gate.weight": {
    "atom_offset": 120790563,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.o_proj.input_min": {
    "atom_offset": 120802852,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.self_attn.k_norm.weight": {
    "atom_offset": 120802854,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_1.linear.weight": {
    "atom_offset": 120802863,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 9.496493730694056e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 120868400,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 6.459153373725712e-4,
    "track": "int8"
  },
  "model.language_model.layers.5.pre_feedforward_layernorm.weight": {
    "atom_offset": 120933937,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.down_proj.output_max": {
    "atom_offset": 120933986,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.v_proj.input_min": {
    "atom_offset": 120933988,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.q_proj.output_min": {
    "atom_offset": 120933990,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.o_proj.linear.weight": {
    "atom_offset": 120933992,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003675566054880619,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.v_proj.linear.weight": {
    "atom_offset": 120943209,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0031680609099566936,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.pre_feedforward_layernorm.weight": {
    "atom_offset": 120952426,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.10.self_attn.k_proj.weight": {
    "atom_offset": 120952451,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0028604823164641857,
    "track": "int8"
  },
  "model.language_model.layers.29.per_layer_projection.weight": {
    "atom_offset": 120958596,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.down_proj.input_max": {
    "atom_offset": 120970885,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.self_attn.v_proj.input_min": {
    "atom_offset": 120970887,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_start.input_max": {
    "atom_offset": 120970889,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.lconv1d.linear_end.output_max": {
    "atom_offset": 120970891,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.pre_layer_norm.weight": {
    "atom_offset": 120970893,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.v_proj.output_max": {
    "atom_offset": 120970926,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 120970928,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.layer_scalar": {
    "atom_offset": 120970930,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.down_proj.output_min": {
    "atom_offset": 120970932,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.gate_proj.output_max": {
    "atom_offset": 120970934,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.post.linear.weight": {
    "atom_offset": 120970936,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 0.0013917938340455294,
    "track": "int8"
  },
  "model.language_model.layers.4.post_attention_layernorm.weight": {
    "atom_offset": 120987321,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.q_proj.input_min": {
    "atom_offset": 120987370,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.mlp.up_proj.weight": {
    "atom_offset": 120987372,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.004029281437397003,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.gate_proj.input_min": {
    "atom_offset": 121134829,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_end.output_min": {
    "atom_offset": 121134831,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.up_proj.input_min": {
    "atom_offset": 121134833,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.self_attn.k_norm.weight": {
    "atom_offset": 121134835,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.v_proj.output_max": {
    "atom_offset": 121134844,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.32.self_attn.k_norm.weight": {
    "atom_offset": 121134846,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.mlp.down_proj.input_max": {
    "atom_offset": 121134855,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.29.mlp.gate_proj.weight": {
    "atom_offset": 121134857,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005536417476832867,
    "track": "int8"
  },
  "model.audio_tower.layers.0.norm_pre_attn.weight": {
    "atom_offset": 121429770,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.156496062874794,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_start.linear.weight": {
    "atom_offset": 121429787,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.0013456570450216532,
    "track": "int8"
  },
  "model.language_model.layers.17.self_attn.q_norm.weight": {
    "atom_offset": 121462556,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.post.input_min": {
    "atom_offset": 121462565,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 121462567,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.self_attn.q_norm.weight": {
    "atom_offset": 121462569,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.mlp.down_proj.input_max": {
    "atom_offset": 121462572,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 121462574,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.self_attn.q_proj.weight": {
    "atom_offset": 121462576,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003783218562602997,
    "track": "int8"
  },
  "model.audio_tower.layers.2.lconv1d.linear_end.output_max": {
    "atom_offset": 121511729,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.k_proj.input_min": {
    "atom_offset": 121511731,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.mlp.gate_proj.weight": {
    "atom_offset": 121511733,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.005751722492277622,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.k_proj.output_min": {
    "atom_offset": 121806646,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.self_attn.k_proj.input_max": {
    "atom_offset": 121806648,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.q_proj.output_min": {
    "atom_offset": 121806650,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.mlp.up_proj.weight": {
    "atom_offset": 121806652,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0034756397362798452,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.q_proj.linear.weight": {
    "atom_offset": 121954109,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.003921628929674625,
    "track": "int8"
  },
  "model.language_model.layers.15.self_attn.k_proj.weight": {
    "atom_offset": 121963326,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.3445496703498065e-4,
    "track": "int8"
  },
  "model.language_model.layers.12.self_attn.q_norm.weight": {
    "atom_offset": 121969471,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.6.self_attn.k_proj.weight": {
    "atom_offset": 121969480,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 0.0023068406153470278,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.gate_proj.input_min": {
    "atom_offset": 121975625,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 121975627,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.1.self_attn.k_proj.output_min": {
    "atom_offset": 121975629,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.self_attn.o_proj.output_max": {
    "atom_offset": 121975631,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.self_attn.q_proj.weight": {
    "atom_offset": 121975633,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003506397595629096,
    "track": "int8"
  },
  "model.audio_tower.layers.3.norm_out.weight": {
    "atom_offset": 122024786,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.1761811077594757,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward1.ffw_layer_1.input_max": {
    "atom_offset": 122024803,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_2.output_max": {
    "atom_offset": 122024805,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 122024807,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.down_proj.output_max": {
    "atom_offset": 122024809,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.up_proj.input_min": {
    "atom_offset": 122024811,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.v_proj.output_min": {
    "atom_offset": 122024813,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.q_proj.output_max": {
    "atom_offset": 122024815,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.norm_out.weight": {
    "atom_offset": 122024817,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.18307086825370789,
    "track": "int8"
  },
  "model.audio_tower.layers.8.lconv1d.linear_end.linear.weight": {
    "atom_offset": 122024834,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 9.804072324186563e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.up_proj.input_min": {
    "atom_offset": 122041219,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.self_attn.q_proj.output_min": {
    "atom_offset": 122041221,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_start.output_min": {
    "atom_offset": 122041223,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.pre_feedforward_layernorm.weight": {
    "atom_offset": 122041225,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.k_proj.output_min": {
    "atom_offset": 122041274,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.down_proj.linear.weight": {
    "atom_offset": 122041276,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0016147884307429194,
    "track": "int8"
  },
  "model.language_model.layers.18.self_attn.v_proj.weight": {
    "atom_offset": 122078141,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.594457568600774e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.1.lconv1d.linear_start.output_min": {
    "atom_offset": 122084286,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.gate_proj.linear.weight": {
    "atom_offset": 122084288,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002429872052744031,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.q_proj.linear.weight": {
    "atom_offset": 122121153,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0021684300154447556,
    "track": "int8"
  },
  "model.language_model.layers.34.post_attention_layernorm.weight": {
    "atom_offset": 122130370,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.mlp.gate_proj.input_min": {
    "atom_offset": 122130419,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.o_proj.input_min": {
    "atom_offset": 122130421,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.self_attn.q_norm.weight": {
    "atom_offset": 122130423,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.9.self_attn.o_proj.input_min": {
    "atom_offset": 122130432,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.input_layernorm.weight": {
    "atom_offset": 122130434,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.18.input_layernorm.weight": {
    "atom_offset": 122130459,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.self_attn.k_norm.weight": {
    "atom_offset": 122130508,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.22.self_attn.k_norm.weight": {
    "atom_offset": 122130511,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.post_per_layer_input_norm.weight": {
    "atom_offset": 122130520,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.feed_forward2.ffw_layer_1.input_min": {
    "atom_offset": 122130569,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.v_proj.output_max": {
    "atom_offset": 122130571,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.v_proj.linear.weight": {
    "atom_offset": 122130573,
    "byte_length": 1048640,
    "num_atoms": 16385,
    "scale": 7.574126357212663e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_2.input_max": {
    "atom_offset": 122146958,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.3.mlp.up_proj.input_max": {
    "atom_offset": 122146960,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.mlp.gate_proj.weight": {
    "atom_offset": 122146962,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.006520669441670179,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.v_proj.input_max": {
    "atom_offset": 122294419,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_start.input_max": {
    "atom_offset": 122294421,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.3.lconv1d.linear_start.output_max": {
    "atom_offset": 122294423,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 122294425,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.post.output_min": {
    "atom_offset": 122294427,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.per_layer_input_gate.weight": {
    "atom_offset": 122294429,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.v_proj.input_min": {
    "atom_offset": 122306718,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.gate_proj.linear.weight": {
    "atom_offset": 122306720,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002506766701117158,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.down_proj.output_max": {
    "atom_offset": 122343585,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.self_attn.o_proj.output_min": {
    "atom_offset": 122343587,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.self_attn.k_proj.weight": {
    "atom_offset": 122343589,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.3445496703498065e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.5.feed_forward1.ffw_layer_2.input_min": {
    "atom_offset": 122349734,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.mlp.up_proj.weight": {
    "atom_offset": 122349736,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.003829355351626873,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.mlp.up_proj.output_max": {
    "atom_offset": 122644649,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.23.self_attn.q_proj.weight": {
    "atom_offset": 122644651,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004459891933947802,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_2.input_max": {
    "atom_offset": 122693804,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward1.ffw_layer_1.output_min": {
    "atom_offset": 122693806,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.per_layer_projection.weight": {
    "atom_offset": 122693808,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.19.mlp.down_proj.weight": {
    "atom_offset": 122706097,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004213828593492508,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_end.output_min": {
    "atom_offset": 123001010,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.feed_forward1.pre_layer_norm.weight": {
    "atom_offset": 123001012,
    "byte_length": 2112,
    "num_atoms": 33,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.down_proj.linear.weight": {
    "atom_offset": 123001045,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.002322219545021653,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.input_layernorm.weight": {
    "atom_offset": 123037910,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.self_attn.k_proj.output_min": {
    "atom_offset": 123037935,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.k_proj.output_min": {
    "atom_offset": 123037937,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.norm_pre_attn.weight": {
    "atom_offset": 123037939,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 0.2539370059967041,
    "track": "int8"
  },
  "model.audio_tower.layers.3.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 123037956,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_start.output_max": {
    "atom_offset": 123037958,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.subsample_conv_projection.layer1.norm.weight": {
    "atom_offset": 123037960,
    "byte_length": 128,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.6.feed_forward2.ffw_layer_1.output_min": {
    "atom_offset": 123037962,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.k_proj.linear.weight": {
    "atom_offset": 123037964,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.0037985974922776222,
    "track": "int8"
  },
  "model.language_model.layers.10.mlp.gate_proj.weight": {
    "atom_offset": 123047181,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0050750491209328175,
    "track": "int8"
  },
  "model.language_model.layers.17.self_attn.v_proj.weight": {
    "atom_offset": 123194638,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.479115305002779e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.down_proj.input_max": {
    "atom_offset": 123200783,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.feed_forward2.ffw_layer_2.input_min": {
    "atom_offset": 123200785,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.self_attn.q_proj.input_max": {
    "atom_offset": 123200787,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.0.self_attn.o_proj.weight": {
    "atom_offset": 123200789,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.003306471509858966,
    "track": "int8"
  },
  "model.language_model.layers.30.pre_feedforward_layernorm.weight": {
    "atom_offset": 123249942,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.2.self_attn.k_proj.input_max": {
    "atom_offset": 123249991,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.lconv1d.linear_start.output_min": {
    "atom_offset": 123249993,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.2.post_attention_layernorm.weight": {
    "atom_offset": 123249995,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.mlp.up_proj.input_min": {
    "atom_offset": 123250044,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.self_attn.q_proj.input_max": {
    "atom_offset": 123250046,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.20.self_attn.k_proj.weight": {
    "atom_offset": 123250048,
    "byte_length": 393280,
    "num_atoms": 6145,
    "scale": 4.7674705274403095e-4,
    "track": "int8"
  },
  "model.audio_tower.layers.2.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 123256193,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.q_proj.input_min": {
    "atom_offset": 123256195,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.self_attn.q_proj.input_max": {
    "atom_offset": 123256197,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.mlp.up_proj.input_min": {
    "atom_offset": 123256199,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.3.mlp.down_proj.weight": {
    "atom_offset": 123256201,
    "byte_length": 9437248,
    "num_atoms": 147457,
    "scale": 0.0055671753361821175,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.gate_proj.input_max": {
    "atom_offset": 123403658,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.self_attn.v_proj.input_max": {
    "atom_offset": 123403660,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.4.mlp.down_proj.input_min": {
    "atom_offset": 123403662,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.0.lconv1d.linear_end.output_min": {
    "atom_offset": 123403664,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.11.lconv1d.linear_start.linear.weight": {
    "atom_offset": 123403666,
    "byte_length": 2097216,
    "num_atoms": 32769,
    "scale": 0.002229945966973901,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.2.self_attn.k_proj.output_min": {
    "atom_offset": 123436435,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.7.feed_forward2.ffw_layer_2.linear.weight": {
    "atom_offset": 123436437,
    "byte_length": 4194368,
    "num_atoms": 65537,
    "scale": 7.99704750534147e-4,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.self_attn.v_proj.input_min": {
    "atom_offset": 123501974,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.9.self_attn.post.output_max": {
    "atom_offset": 123501976,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.gate_proj.linear.weight": {
    "atom_offset": 123501978,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0022914616856724024,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.12.self_attn.o_proj.output_max": {
    "atom_offset": 123538843,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.0.post_attention_layernorm.weight": {
    "atom_offset": 123538845,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.v_proj.input_min": {
    "atom_offset": 123538870,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.lconv1d.linear_end.input_min": {
    "atom_offset": 123538872,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.q_norm.weight": {
    "atom_offset": 123538874,
    "byte_length": 192,
    "num_atoms": 3,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.4.input_layernorm.weight": {
    "atom_offset": 123538877,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.self_attn.k_proj.linear.weight": {
    "atom_offset": 123538926,
    "byte_length": 589888,
    "num_atoms": 9217,
    "scale": 0.00366018689237535,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.14.post_feedforward_layernorm.weight": {
    "atom_offset": 123548143,
    "byte_length": 1600,
    "num_atoms": 25,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.self_attn.per_dim_scale": {
    "atom_offset": 123548168,
    "byte_length": 320,
    "num_atoms": 5,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.8.self_attn.k_proj.input_min": {
    "atom_offset": 123548173,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.25.pre_feedforward_layernorm.weight": {
    "atom_offset": 123548175,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.21.self_attn.q_proj.weight": {
    "atom_offset": 123548224,
    "byte_length": 3145792,
    "num_atoms": 49153,
    "scale": 0.004552165511995554,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.mlp.up_proj.input_min": {
    "atom_offset": 123597377,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.1.self_attn.o_proj.output_max": {
    "atom_offset": 123597379,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.8.mlp.up_proj.output_min": {
    "atom_offset": 123597381,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.13.mlp.gate_proj.output_max": {
    "atom_offset": 123597383,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.7.self_attn.k_proj.output_min": {
    "atom_offset": 123597385,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.o_proj.output_max": {
    "atom_offset": 123597387,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.5.mlp.gate_proj.input_min": {
    "atom_offset": 123597389,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.per_layer_projection.weight": {
    "atom_offset": 123597391,
    "byte_length": 786496,
    "num_atoms": 12289,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.5.lconv1d.linear_end.output_min": {
    "atom_offset": 123609680,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.4.feed_forward2.ffw_layer_1.output_max": {
    "atom_offset": 123609682,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.13.post_attention_layernorm.weight": {
    "atom_offset": 123609684,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.10.mlp.down_proj.input_min": {
    "atom_offset": 123609733,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.15.mlp.up_proj.input_min": {
    "atom_offset": 123609735,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.12.input_layernorm.weight": {
    "atom_offset": 123609737,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.5.input_layernorm.weight": {
    "atom_offset": 123609786,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.6.self_attn.k_proj.output_max": {
    "atom_offset": 123609835,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.9.self_attn.k_norm.weight": {
    "atom_offset": 123609837,
    "byte_length": 1088,
    "num_atoms": 17,
    "scale": 1.0,
    "track": "int8"
  },
  "model.vision_tower.encoder.layers.11.mlp.up_proj.linear.weight": {
    "atom_offset": 123609854,
    "byte_length": 2359360,
    "num_atoms": 36865,
    "scale": 0.0021222932264208794,
    "track": "int8"
  },
  "model.language_model.layers.28.self_attn.q_norm.weight": {
    "atom_offset": 123646719,
    "byte_length": 576,
    "num_atoms": 9,
    "scale": 1.0,
    "track": "int8"
  },
  "model.audio_tower.layers.10.feed_forward2.ffw_layer_2.output_max": {
    "atom_offset": 123646728,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.28.mlp.down_proj.weight": {
    "atom_offset": 123646730,
    "byte_length": 18874432,
    "num_atoms": 294913,
    "scale": 0.004060039296746254,
    "track": "int8"
  },
  "model.audio_tower.layers.4.self_attn.q_proj.output_min": {
    "atom_offset": 123941643,
    "byte_length": 66,
    "num_atoms": 2,
    "scale": 1.0,
    "track": "int8"
  },
  "model.language_model.layers.24.post_per_layer_input_norm.weight": {
    "atom_offset": 123941645,
    "byte_length": 3136,
    "num_atoms": 49,
    "scale": 1.0,
    "track": "int8"
  }
}