elyx commited on
Commit
d6325c4
·
verified ·
1 Parent(s): 916ec3f

Upload Qwen2AudioForConditionalGeneration

Browse files
config.json CHANGED
@@ -5,7 +5,7 @@
5
  ],
6
  "audio_config": {
7
  "model_type": "qwen2_audio_encoder",
8
- "torch_dtype": "float32"
9
  },
10
  "audio_token_index": 151646,
11
  "ignore_index": -100,
@@ -44,7 +44,7 @@
44
  }
45
  },
46
  "format": "int-quantized",
47
- "global_compression_ratio": 2.02943919718519,
48
  "ignore": [],
49
  "kv_cache_scheme": null,
50
  "quant_method": "compressed-tensors",
@@ -58,11 +58,11 @@
58
  "model_type": "qwen2",
59
  "rms_norm_eps": 1e-05,
60
  "sliding_window": null,
61
- "torch_dtype": "float32",
62
  "use_mrope": false,
63
  "vocab_size": 156032
64
  },
65
- "torch_dtype": "float32",
66
  "transformers_version": "4.49.0",
67
  "vocab_size": 156032
68
  }
 
5
  ],
6
  "audio_config": {
7
  "model_type": "qwen2_audio_encoder",
8
+ "torch_dtype": "bfloat16"
9
  },
10
  "audio_token_index": 151646,
11
  "ignore_index": -100,
 
44
  }
45
  },
46
  "format": "int-quantized",
47
+ "global_compression_ratio": 1.510915647565434,
48
  "ignore": [],
49
  "kv_cache_scheme": null,
50
  "quant_method": "compressed-tensors",
 
58
  "model_type": "qwen2",
59
  "rms_norm_eps": 1e-05,
60
  "sliding_window": null,
61
+ "torch_dtype": "bfloat16",
62
  "use_mrope": false,
63
  "vocab_size": 156032
64
  },
65
+ "torch_dtype": "bfloat16",
66
  "transformers_version": "4.49.0",
67
  "vocab_size": 156032
68
  }
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:de8191ef44c9e32fe0b6d7b88688998843fcce1f69217513069ee4c91e92abb1
3
  size 4983429624
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9bd6c0780a60409335aa5db67a96b858e0c5cd0c5fcf9684a44d684cde40fb7
3
  size 4983429624
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a9863ce817c5ffc2e4b49e71514981bd9e26e01bda1281d323450c2a91902ba3
3
  size 4065193744
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7aaedfb5091b9e1047962df0869c02dfbff66d71841aa617afcfaee761c560f7
3
  size 4065193744
model.safetensors.index.json CHANGED
The diff for this file is too large to render. See raw diff
 
recipe.yaml CHANGED
@@ -1,6 +1,6 @@
1
  DEFAULT_stage:
2
  DEFAULT_modifiers:
3
- SmoothQuantModifier: {smoothing_strength: 0.8}
4
  GPTQModifier:
5
  scheme: W8A8
6
  targets: Linear
 
1
  DEFAULT_stage:
2
  DEFAULT_modifiers:
3
+ SmoothQuantModifier: {smoothing_strength: 0.5}
4
  GPTQModifier:
5
  scheme: W8A8
6
  targets: Linear