Hugging Face's logo

pytorch
/
gemma-3-27b-it-FP8

Any-to-Any
Transformers
PyTorch
English
gemma3
torchao
text-generation-inference
Model card Files
xet
Community
jerryzh168 commited on
Commit
d818521
·
verified ·
1 Parent(s): 51a8a11

Upload Gemma3ForConditionalGeneration

Browse files
Files changed (9) hide show
  1. config.json +72 -7
  2. generation_config.json +1 -1
  3. pytorch_model-00001-of-00006.bin +2 -2
  4. pytorch_model-00002-of-00006.bin +2 -2
  5. pytorch_model-00003-of-00006.bin +2 -2
  6. pytorch_model-00004-of-00006.bin +2 -2
  7. pytorch_model-00005-of-00006.bin +2 -2
  8. pytorch_model-00006-of-00006.bin +2 -2
  9. pytorch_model.bin.index.json +1 -0
config.json CHANGED
@@ -3,6 +3,7 @@
3
  "Gemma3ForConditionalGeneration"
4
  ],
5
  "boi_token_index": 255999,
 
6
  "eoi_token_index": 256000,
7
  "eos_token_id": [
8
  1,
@@ -13,6 +14,7 @@
13
  "mm_tokens_per_image": 256,
14
  "model_type": "gemma3",
15
  "quantization_config": {
 
16
  "modules_to_not_convert": null,
17
  "quant_method": "torchao",
18
  "quant_type": {
@@ -59,19 +61,85 @@
59
  "_version": 2
60
  }
61
  },
62
- "quant_type_kwargs": {}
 
63
  },
64
  "text_config": {
 
65
  "attention_bias": false,
66
  "attention_dropout": 0.0,
67
  "attn_logit_softcapping": null,
68
- "cache_implementation": "hybrid",
69
  "final_logit_softcapping": null,
70
  "head_dim": 128,
71
  "hidden_activation": "gelu_pytorch_tanh",
72
  "hidden_size": 5376,
73
  "initializer_range": 0.02,
74
  "intermediate_size": 21504,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
  "max_position_embeddings": 131072,
76
  "model_type": "gemma3_text",
77
  "num_attention_heads": 32,
@@ -86,15 +154,13 @@
86
  },
87
  "rope_theta": 1000000.0,
88
  "sliding_window": 1024,
89
- "sliding_window_pattern": 6,
90
- "torch_dtype": "bfloat16",
91
  "use_cache": true,
92
  "vocab_size": 262208
93
  },
94
- "torch_dtype": "bfloat16",
95
- "transformers_version": "4.52.0.dev0",
96
  "vision_config": {
97
  "attention_dropout": 0.0,
 
98
  "hidden_act": "gelu_pytorch_tanh",
99
  "hidden_size": 1152,
100
  "image_size": 896,
@@ -105,7 +171,6 @@
105
  "num_channels": 3,
106
  "num_hidden_layers": 27,
107
  "patch_size": 14,
108
- "torch_dtype": "bfloat16",
109
  "vision_use_head": false
110
  }
111
  }
 
3
  "Gemma3ForConditionalGeneration"
4
  ],
5
  "boi_token_index": 255999,
6
+ "dtype": "bfloat16",
7
  "eoi_token_index": 256000,
8
  "eos_token_id": [
9
  1,
 
14
  "mm_tokens_per_image": 256,
15
  "model_type": "gemma3",
16
  "quantization_config": {
17
+ "include_input_output_embeddings": false,
18
  "modules_to_not_convert": null,
19
  "quant_method": "torchao",
20
  "quant_type": {
 
61
  "_version": 2
62
  }
63
  },
64
+ "quant_type_kwargs": {},
65
+ "untie_embedding_weights": false
66
  },
67
  "text_config": {
68
+ "_sliding_window_pattern": 6,
69
  "attention_bias": false,
70
  "attention_dropout": 0.0,
71
  "attn_logit_softcapping": null,
72
+ "dtype": "bfloat16",
73
  "final_logit_softcapping": null,
74
  "head_dim": 128,
75
  "hidden_activation": "gelu_pytorch_tanh",
76
  "hidden_size": 5376,
77
  "initializer_range": 0.02,
78
  "intermediate_size": 21504,
79
+ "layer_types": [
80
+ "sliding_attention",
81
+ "sliding_attention",
82
+ "sliding_attention",
83
+ "sliding_attention",
84
+ "sliding_attention",
85
+ "full_attention",
86
+ "sliding_attention",
87
+ "sliding_attention",
88
+ "sliding_attention",
89
+ "sliding_attention",
90
+ "sliding_attention",
91
+ "full_attention",
92
+ "sliding_attention",
93
+ "sliding_attention",
94
+ "sliding_attention",
95
+ "sliding_attention",
96
+ "sliding_attention",
97
+ "full_attention",
98
+ "sliding_attention",
99
+ "sliding_attention",
100
+ "sliding_attention",
101
+ "sliding_attention",
102
+ "sliding_attention",
103
+ "full_attention",
104
+ "sliding_attention",
105
+ "sliding_attention",
106
+ "sliding_attention",
107
+ "sliding_attention",
108
+ "sliding_attention",
109
+ "full_attention",
110
+ "sliding_attention",
111
+ "sliding_attention",
112
+ "sliding_attention",
113
+ "sliding_attention",
114
+ "sliding_attention",
115
+ "full_attention",
116
+ "sliding_attention",
117
+ "sliding_attention",
118
+ "sliding_attention",
119
+ "sliding_attention",
120
+ "sliding_attention",
121
+ "full_attention",
122
+ "sliding_attention",
123
+ "sliding_attention",
124
+ "sliding_attention",
125
+ "sliding_attention",
126
+ "sliding_attention",
127
+ "full_attention",
128
+ "sliding_attention",
129
+ "sliding_attention",
130
+ "sliding_attention",
131
+ "sliding_attention",
132
+ "sliding_attention",
133
+ "full_attention",
134
+ "sliding_attention",
135
+ "sliding_attention",
136
+ "sliding_attention",
137
+ "sliding_attention",
138
+ "sliding_attention",
139
+ "full_attention",
140
+ "sliding_attention",
141
+ "sliding_attention"
142
+ ],
143
  "max_position_embeddings": 131072,
144
  "model_type": "gemma3_text",
145
  "num_attention_heads": 32,
 
154
  },
155
  "rope_theta": 1000000.0,
156
  "sliding_window": 1024,
 
 
157
  "use_cache": true,
158
  "vocab_size": 262208
159
  },
160
+ "transformers_version": "4.56.1",
 
161
  "vision_config": {
162
  "attention_dropout": 0.0,
163
+ "dtype": "bfloat16",
164
  "hidden_act": "gelu_pytorch_tanh",
165
  "hidden_size": 1152,
166
  "image_size": 896,
 
171
  "num_channels": 3,
172
  "num_hidden_layers": 27,
173
  "patch_size": 14,
 
174
  "vision_use_head": false
175
  }
176
  }
generation_config.json CHANGED
@@ -9,5 +9,5 @@
9
  "pad_token_id": 0,
10
  "top_k": 64,
11
  "top_p": 0.95,
12
- "transformers_version": "4.52.0.dev0"
13
  }
 
9
  "pad_token_id": 0,
10
  "top_k": 64,
11
  "top_p": 0.95,
12
+ "transformers_version": "4.56.1"
13
  }
pytorch_model-00001-of-00006.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d0f16a9f6f17baff4bc33b32bbec519e19bdc0f5e1892e59f567d2fb4acf3c52
3
- size 4974464373
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ca95abb975180893fb279fbcf7558dbad1f45f614735a49710ad85b1ea5a2de
3
+ size 4974463221
pytorch_model-00002-of-00006.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3c01065d30bb26597a53cda3552989b0d0462ea8f89d50a681c19e7748e6e4a7
3
- size 4958116713
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0afb4d643b54b73011525ca190b04531be99c19d40d9082dd856e1cf0c00aa60
3
+ size 4958116201
pytorch_model-00003-of-00006.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a9a7bbb1a667d9c31df8d9df940a7be1afbf667005944b6f515b5fbb0ee39f01
3
- size 4958116777
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6751d04c9981ddbaa026aa5e2003a93ebc63a4d9eef0ccc3f4d18f02ec0ef3d9
3
+ size 4958116265
pytorch_model-00004-of-00006.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:35d2d810eff91d347d31b8fcc7fab0ced38f6899aba3aab70b7d38126ff34aec
3
- size 4958116777
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:83f273ec90fd4a94f5263a144afbaf4332c917d9379d05571d57ddf63a28b9dd
3
+ size 4958116265
pytorch_model-00005-of-00006.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5ca0ac3b5c5b5a5071b0ae4212703ac925d5bc2f83e56af1c54eabddd1d85304
3
- size 4958116777
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ddea3cf322fb38ee76c4c8b4b47614a464c037e1b4d767b2028ac4bcf4bdff1
3
+ size 4958116265
pytorch_model-00006-of-00006.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:82b3c0b234328b8095fc30db089dc8f056b6b383a609719da73bc9fa5b87bfbd
3
- size 4065655722
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9230b00b73bf55b4faf1e70ec77ca993c0ceb0aa109e9b40a56b0512bf399469
3
+ size 4065655338
pytorch_model.bin.index.json CHANGED
@@ -1,5 +1,6 @@
1
  {
2
  "metadata": {
 
3
  "total_size": 28871832224
4
  },
5
  "weight_map": {
 
1
  {
2
  "metadata": {
3
+ "total_parameters": 27432406640,
4
  "total_size": 28871832224
5
  },
6
  "weight_map": {