hermeschen-ezcon commited on
Commit
839bd19
·
verified ·
1 Parent(s): 72f1d3e

Upload folder using huggingface_hub

Browse files
README.md CHANGED
@@ -9,7 +9,7 @@ tags:
9
  ---
10
 
11
  # EZCon/RolmOCR-4bit-mlx
12
- This model was converted to MLX format from [`reducto/RolmOCR`]() using mlx-vlm version **0.1.26**.
13
  Refer to the [original model card](https://huggingface.co/reducto/RolmOCR) for more details on the model.
14
  ## Use with mlx
15
 
 
9
  ---
10
 
11
  # EZCon/RolmOCR-4bit-mlx
12
+ This model was converted to MLX format from [`reducto/RolmOCR`]() using mlx-vlm version **0.3.2**.
13
  Refer to the [original model card](https://huggingface.co/reducto/RolmOCR) for more details on the model.
14
  ## Use with mlx
15
 
config.json CHANGED
@@ -1,5 +1,4 @@
1
  {
2
- "_attn_implementation_autoset": false,
3
  "add_cross_attention": false,
4
  "architectures": [
5
  "Qwen2_5_VLForConditionalGeneration"
@@ -57,7 +56,981 @@
57
  "pruned_heads": {},
58
  "quantization": {
59
  "group_size": 64,
60
- "bits": 4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
61
  },
62
  "remove_invalid_values": false,
63
  "repetition_penalty": 1.0,
@@ -79,35 +1052,97 @@
79
  "suppress_tokens": null,
80
  "task_specific_params": null,
81
  "temperature": 1.0,
82
- "tf_legacy_loss": false,
83
- "tie_encoder_decoder": false,
84
- "tie_word_embeddings": false,
85
- "tokenizer_class": null,
86
- "top_k": 50,
87
- "top_p": 1.0,
88
- "torchscript": false,
89
- "transformers_version": "4.51.3",
90
- "typical_p": 1.0,
91
- "use_bfloat16": false,
92
- "use_cache": false,
93
- "use_sliding_window": false,
94
- "video_token_id": 151656,
95
- "vision_config": {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96
  "return_dict": true,
97
  "output_hidden_states": false,
98
- "output_attentions": false,
99
  "torchscript": false,
100
- "torch_dtype": "float32",
101
- "use_bfloat16": false,
102
- "tf_legacy_loss": false,
103
  "pruned_heads": {},
104
- "tie_word_embeddings": true,
105
  "chunk_size_feed_forward": 0,
106
  "is_encoder_decoder": false,
107
  "is_decoder": false,
108
  "cross_attention_hidden_size": null,
109
  "add_cross_attention": false,
110
  "tie_encoder_decoder": false,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
111
  "max_length": 20,
112
  "min_length": 0,
113
  "do_sample": false,
@@ -133,6 +1168,41 @@
133
  "exponential_decay_length_penalty": null,
134
  "suppress_tokens": null,
135
  "begin_suppress_tokens": null,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
136
  "architectures": null,
137
  "finetuning_task": null,
138
  "id2label": {
@@ -143,6 +1213,8 @@
143
  "LABEL_0": 0,
144
  "LABEL_1": 1
145
  },
 
 
146
  "tokenizer_class": null,
147
  "prefix": null,
148
  "bos_token_id": null,
@@ -150,13 +1222,37 @@
150
  "eos_token_id": null,
151
  "sep_token_id": null,
152
  "decoder_start_token_id": null,
153
- "task_specific_params": null,
154
- "problem_type": null,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
155
  "_name_or_path": "",
156
- "_attn_implementation_autoset": false,
157
  "in_chans": 3,
158
  "model_type": "qwen2_5_vl",
159
  "spatial_patch_size": 14,
 
 
160
  "depth": 32,
161
  "hidden_size": 1280,
162
  "hidden_act": "silu",
@@ -174,7 +1270,9 @@
174
  23,
175
  31
176
  ],
177
- "out_hidden_size": 3584
 
 
178
  },
179
  "vision_end_token_id": 151653,
180
  "vision_start_token_id": 151652,
 
1
  {
 
2
  "add_cross_attention": false,
3
  "architectures": [
4
  "Qwen2_5_VLForConditionalGeneration"
 
56
  "pruned_heads": {},
57
  "quantization": {
58
  "group_size": 64,
59
+ "bits": 4,
60
+ "vision_tower.patch_embed.proj": false,
61
+ "vision_tower.rotary_pos_emb": false,
62
+ "vision_tower.blocks.0.norm1": false,
63
+ "vision_tower.blocks.0.norm2": false,
64
+ "vision_tower.blocks.0.attn.qkv": false,
65
+ "vision_tower.blocks.0.attn.proj": false,
66
+ "vision_tower.blocks.0.mlp.gate_proj": false,
67
+ "vision_tower.blocks.0.mlp.up_proj": false,
68
+ "vision_tower.blocks.0.mlp.down_proj": false,
69
+ "vision_tower.blocks.1.norm1": false,
70
+ "vision_tower.blocks.1.norm2": false,
71
+ "vision_tower.blocks.1.attn.qkv": false,
72
+ "vision_tower.blocks.1.attn.proj": false,
73
+ "vision_tower.blocks.1.mlp.gate_proj": false,
74
+ "vision_tower.blocks.1.mlp.up_proj": false,
75
+ "vision_tower.blocks.1.mlp.down_proj": false,
76
+ "vision_tower.blocks.2.norm1": false,
77
+ "vision_tower.blocks.2.norm2": false,
78
+ "vision_tower.blocks.2.attn.qkv": false,
79
+ "vision_tower.blocks.2.attn.proj": false,
80
+ "vision_tower.blocks.2.mlp.gate_proj": false,
81
+ "vision_tower.blocks.2.mlp.up_proj": false,
82
+ "vision_tower.blocks.2.mlp.down_proj": false,
83
+ "vision_tower.blocks.3.norm1": false,
84
+ "vision_tower.blocks.3.norm2": false,
85
+ "vision_tower.blocks.3.attn.qkv": false,
86
+ "vision_tower.blocks.3.attn.proj": false,
87
+ "vision_tower.blocks.3.mlp.gate_proj": false,
88
+ "vision_tower.blocks.3.mlp.up_proj": false,
89
+ "vision_tower.blocks.3.mlp.down_proj": false,
90
+ "vision_tower.blocks.4.norm1": false,
91
+ "vision_tower.blocks.4.norm2": false,
92
+ "vision_tower.blocks.4.attn.qkv": false,
93
+ "vision_tower.blocks.4.attn.proj": false,
94
+ "vision_tower.blocks.4.mlp.gate_proj": false,
95
+ "vision_tower.blocks.4.mlp.up_proj": false,
96
+ "vision_tower.blocks.4.mlp.down_proj": false,
97
+ "vision_tower.blocks.5.norm1": false,
98
+ "vision_tower.blocks.5.norm2": false,
99
+ "vision_tower.blocks.5.attn.qkv": false,
100
+ "vision_tower.blocks.5.attn.proj": false,
101
+ "vision_tower.blocks.5.mlp.gate_proj": false,
102
+ "vision_tower.blocks.5.mlp.up_proj": false,
103
+ "vision_tower.blocks.5.mlp.down_proj": false,
104
+ "vision_tower.blocks.6.norm1": false,
105
+ "vision_tower.blocks.6.norm2": false,
106
+ "vision_tower.blocks.6.attn.qkv": false,
107
+ "vision_tower.blocks.6.attn.proj": false,
108
+ "vision_tower.blocks.6.mlp.gate_proj": false,
109
+ "vision_tower.blocks.6.mlp.up_proj": false,
110
+ "vision_tower.blocks.6.mlp.down_proj": false,
111
+ "vision_tower.blocks.7.norm1": false,
112
+ "vision_tower.blocks.7.norm2": false,
113
+ "vision_tower.blocks.7.attn.qkv": false,
114
+ "vision_tower.blocks.7.attn.proj": false,
115
+ "vision_tower.blocks.7.mlp.gate_proj": false,
116
+ "vision_tower.blocks.7.mlp.up_proj": false,
117
+ "vision_tower.blocks.7.mlp.down_proj": false,
118
+ "vision_tower.blocks.8.norm1": false,
119
+ "vision_tower.blocks.8.norm2": false,
120
+ "vision_tower.blocks.8.attn.qkv": false,
121
+ "vision_tower.blocks.8.attn.proj": false,
122
+ "vision_tower.blocks.8.mlp.gate_proj": false,
123
+ "vision_tower.blocks.8.mlp.up_proj": false,
124
+ "vision_tower.blocks.8.mlp.down_proj": false,
125
+ "vision_tower.blocks.9.norm1": false,
126
+ "vision_tower.blocks.9.norm2": false,
127
+ "vision_tower.blocks.9.attn.qkv": false,
128
+ "vision_tower.blocks.9.attn.proj": false,
129
+ "vision_tower.blocks.9.mlp.gate_proj": false,
130
+ "vision_tower.blocks.9.mlp.up_proj": false,
131
+ "vision_tower.blocks.9.mlp.down_proj": false,
132
+ "vision_tower.blocks.10.norm1": false,
133
+ "vision_tower.blocks.10.norm2": false,
134
+ "vision_tower.blocks.10.attn.qkv": false,
135
+ "vision_tower.blocks.10.attn.proj": false,
136
+ "vision_tower.blocks.10.mlp.gate_proj": false,
137
+ "vision_tower.blocks.10.mlp.up_proj": false,
138
+ "vision_tower.blocks.10.mlp.down_proj": false,
139
+ "vision_tower.blocks.11.norm1": false,
140
+ "vision_tower.blocks.11.norm2": false,
141
+ "vision_tower.blocks.11.attn.qkv": false,
142
+ "vision_tower.blocks.11.attn.proj": false,
143
+ "vision_tower.blocks.11.mlp.gate_proj": false,
144
+ "vision_tower.blocks.11.mlp.up_proj": false,
145
+ "vision_tower.blocks.11.mlp.down_proj": false,
146
+ "vision_tower.blocks.12.norm1": false,
147
+ "vision_tower.blocks.12.norm2": false,
148
+ "vision_tower.blocks.12.attn.qkv": false,
149
+ "vision_tower.blocks.12.attn.proj": false,
150
+ "vision_tower.blocks.12.mlp.gate_proj": false,
151
+ "vision_tower.blocks.12.mlp.up_proj": false,
152
+ "vision_tower.blocks.12.mlp.down_proj": false,
153
+ "vision_tower.blocks.13.norm1": false,
154
+ "vision_tower.blocks.13.norm2": false,
155
+ "vision_tower.blocks.13.attn.qkv": false,
156
+ "vision_tower.blocks.13.attn.proj": false,
157
+ "vision_tower.blocks.13.mlp.gate_proj": false,
158
+ "vision_tower.blocks.13.mlp.up_proj": false,
159
+ "vision_tower.blocks.13.mlp.down_proj": false,
160
+ "vision_tower.blocks.14.norm1": false,
161
+ "vision_tower.blocks.14.norm2": false,
162
+ "vision_tower.blocks.14.attn.qkv": false,
163
+ "vision_tower.blocks.14.attn.proj": false,
164
+ "vision_tower.blocks.14.mlp.gate_proj": false,
165
+ "vision_tower.blocks.14.mlp.up_proj": false,
166
+ "vision_tower.blocks.14.mlp.down_proj": false,
167
+ "vision_tower.blocks.15.norm1": false,
168
+ "vision_tower.blocks.15.norm2": false,
169
+ "vision_tower.blocks.15.attn.qkv": false,
170
+ "vision_tower.blocks.15.attn.proj": false,
171
+ "vision_tower.blocks.15.mlp.gate_proj": false,
172
+ "vision_tower.blocks.15.mlp.up_proj": false,
173
+ "vision_tower.blocks.15.mlp.down_proj": false,
174
+ "vision_tower.blocks.16.norm1": false,
175
+ "vision_tower.blocks.16.norm2": false,
176
+ "vision_tower.blocks.16.attn.qkv": false,
177
+ "vision_tower.blocks.16.attn.proj": false,
178
+ "vision_tower.blocks.16.mlp.gate_proj": false,
179
+ "vision_tower.blocks.16.mlp.up_proj": false,
180
+ "vision_tower.blocks.16.mlp.down_proj": false,
181
+ "vision_tower.blocks.17.norm1": false,
182
+ "vision_tower.blocks.17.norm2": false,
183
+ "vision_tower.blocks.17.attn.qkv": false,
184
+ "vision_tower.blocks.17.attn.proj": false,
185
+ "vision_tower.blocks.17.mlp.gate_proj": false,
186
+ "vision_tower.blocks.17.mlp.up_proj": false,
187
+ "vision_tower.blocks.17.mlp.down_proj": false,
188
+ "vision_tower.blocks.18.norm1": false,
189
+ "vision_tower.blocks.18.norm2": false,
190
+ "vision_tower.blocks.18.attn.qkv": false,
191
+ "vision_tower.blocks.18.attn.proj": false,
192
+ "vision_tower.blocks.18.mlp.gate_proj": false,
193
+ "vision_tower.blocks.18.mlp.up_proj": false,
194
+ "vision_tower.blocks.18.mlp.down_proj": false,
195
+ "vision_tower.blocks.19.norm1": false,
196
+ "vision_tower.blocks.19.norm2": false,
197
+ "vision_tower.blocks.19.attn.qkv": false,
198
+ "vision_tower.blocks.19.attn.proj": false,
199
+ "vision_tower.blocks.19.mlp.gate_proj": false,
200
+ "vision_tower.blocks.19.mlp.up_proj": false,
201
+ "vision_tower.blocks.19.mlp.down_proj": false,
202
+ "vision_tower.blocks.20.norm1": false,
203
+ "vision_tower.blocks.20.norm2": false,
204
+ "vision_tower.blocks.20.attn.qkv": false,
205
+ "vision_tower.blocks.20.attn.proj": false,
206
+ "vision_tower.blocks.20.mlp.gate_proj": false,
207
+ "vision_tower.blocks.20.mlp.up_proj": false,
208
+ "vision_tower.blocks.20.mlp.down_proj": false,
209
+ "vision_tower.blocks.21.norm1": false,
210
+ "vision_tower.blocks.21.norm2": false,
211
+ "vision_tower.blocks.21.attn.qkv": false,
212
+ "vision_tower.blocks.21.attn.proj": false,
213
+ "vision_tower.blocks.21.mlp.gate_proj": false,
214
+ "vision_tower.blocks.21.mlp.up_proj": false,
215
+ "vision_tower.blocks.21.mlp.down_proj": false,
216
+ "vision_tower.blocks.22.norm1": false,
217
+ "vision_tower.blocks.22.norm2": false,
218
+ "vision_tower.blocks.22.attn.qkv": false,
219
+ "vision_tower.blocks.22.attn.proj": false,
220
+ "vision_tower.blocks.22.mlp.gate_proj": false,
221
+ "vision_tower.blocks.22.mlp.up_proj": false,
222
+ "vision_tower.blocks.22.mlp.down_proj": false,
223
+ "vision_tower.blocks.23.norm1": false,
224
+ "vision_tower.blocks.23.norm2": false,
225
+ "vision_tower.blocks.23.attn.qkv": false,
226
+ "vision_tower.blocks.23.attn.proj": false,
227
+ "vision_tower.blocks.23.mlp.gate_proj": false,
228
+ "vision_tower.blocks.23.mlp.up_proj": false,
229
+ "vision_tower.blocks.23.mlp.down_proj": false,
230
+ "vision_tower.blocks.24.norm1": false,
231
+ "vision_tower.blocks.24.norm2": false,
232
+ "vision_tower.blocks.24.attn.qkv": false,
233
+ "vision_tower.blocks.24.attn.proj": false,
234
+ "vision_tower.blocks.24.mlp.gate_proj": false,
235
+ "vision_tower.blocks.24.mlp.up_proj": false,
236
+ "vision_tower.blocks.24.mlp.down_proj": false,
237
+ "vision_tower.blocks.25.norm1": false,
238
+ "vision_tower.blocks.25.norm2": false,
239
+ "vision_tower.blocks.25.attn.qkv": false,
240
+ "vision_tower.blocks.25.attn.proj": false,
241
+ "vision_tower.blocks.25.mlp.gate_proj": false,
242
+ "vision_tower.blocks.25.mlp.up_proj": false,
243
+ "vision_tower.blocks.25.mlp.down_proj": false,
244
+ "vision_tower.blocks.26.norm1": false,
245
+ "vision_tower.blocks.26.norm2": false,
246
+ "vision_tower.blocks.26.attn.qkv": false,
247
+ "vision_tower.blocks.26.attn.proj": false,
248
+ "vision_tower.blocks.26.mlp.gate_proj": false,
249
+ "vision_tower.blocks.26.mlp.up_proj": false,
250
+ "vision_tower.blocks.26.mlp.down_proj": false,
251
+ "vision_tower.blocks.27.norm1": false,
252
+ "vision_tower.blocks.27.norm2": false,
253
+ "vision_tower.blocks.27.attn.qkv": false,
254
+ "vision_tower.blocks.27.attn.proj": false,
255
+ "vision_tower.blocks.27.mlp.gate_proj": false,
256
+ "vision_tower.blocks.27.mlp.up_proj": false,
257
+ "vision_tower.blocks.27.mlp.down_proj": false,
258
+ "vision_tower.blocks.28.norm1": false,
259
+ "vision_tower.blocks.28.norm2": false,
260
+ "vision_tower.blocks.28.attn.qkv": false,
261
+ "vision_tower.blocks.28.attn.proj": false,
262
+ "vision_tower.blocks.28.mlp.gate_proj": false,
263
+ "vision_tower.blocks.28.mlp.up_proj": false,
264
+ "vision_tower.blocks.28.mlp.down_proj": false,
265
+ "vision_tower.blocks.29.norm1": false,
266
+ "vision_tower.blocks.29.norm2": false,
267
+ "vision_tower.blocks.29.attn.qkv": false,
268
+ "vision_tower.blocks.29.attn.proj": false,
269
+ "vision_tower.blocks.29.mlp.gate_proj": false,
270
+ "vision_tower.blocks.29.mlp.up_proj": false,
271
+ "vision_tower.blocks.29.mlp.down_proj": false,
272
+ "vision_tower.blocks.30.norm1": false,
273
+ "vision_tower.blocks.30.norm2": false,
274
+ "vision_tower.blocks.30.attn.qkv": false,
275
+ "vision_tower.blocks.30.attn.proj": false,
276
+ "vision_tower.blocks.30.mlp.gate_proj": false,
277
+ "vision_tower.blocks.30.mlp.up_proj": false,
278
+ "vision_tower.blocks.30.mlp.down_proj": false,
279
+ "vision_tower.blocks.31.norm1": false,
280
+ "vision_tower.blocks.31.norm2": false,
281
+ "vision_tower.blocks.31.attn.qkv": false,
282
+ "vision_tower.blocks.31.attn.proj": false,
283
+ "vision_tower.blocks.31.mlp.gate_proj": false,
284
+ "vision_tower.blocks.31.mlp.up_proj": false,
285
+ "vision_tower.blocks.31.mlp.down_proj": false,
286
+ "vision_tower.merger.ln_q": false,
287
+ "vision_tower.merger.mlp.0": false,
288
+ "vision_tower.merger.mlp.1": false,
289
+ "vision_tower.merger.mlp.2": false,
290
+ "language_model.model.embed_tokens": true,
291
+ "language_model.model.layers.0.self_attn.q_proj": true,
292
+ "language_model.model.layers.0.self_attn.k_proj": true,
293
+ "language_model.model.layers.0.self_attn.v_proj": true,
294
+ "language_model.model.layers.0.self_attn.o_proj": true,
295
+ "language_model.model.layers.0.mlp.gate_proj": true,
296
+ "language_model.model.layers.0.mlp.down_proj": true,
297
+ "language_model.model.layers.0.mlp.up_proj": true,
298
+ "language_model.model.layers.0.input_layernorm": false,
299
+ "language_model.model.layers.0.post_attention_layernorm": false,
300
+ "language_model.model.layers.1.self_attn.q_proj": true,
301
+ "language_model.model.layers.1.self_attn.k_proj": true,
302
+ "language_model.model.layers.1.self_attn.v_proj": true,
303
+ "language_model.model.layers.1.self_attn.o_proj": true,
304
+ "language_model.model.layers.1.mlp.gate_proj": true,
305
+ "language_model.model.layers.1.mlp.down_proj": true,
306
+ "language_model.model.layers.1.mlp.up_proj": true,
307
+ "language_model.model.layers.1.input_layernorm": false,
308
+ "language_model.model.layers.1.post_attention_layernorm": false,
309
+ "language_model.model.layers.2.self_attn.q_proj": true,
310
+ "language_model.model.layers.2.self_attn.k_proj": true,
311
+ "language_model.model.layers.2.self_attn.v_proj": true,
312
+ "language_model.model.layers.2.self_attn.o_proj": true,
313
+ "language_model.model.layers.2.mlp.gate_proj": true,
314
+ "language_model.model.layers.2.mlp.down_proj": true,
315
+ "language_model.model.layers.2.mlp.up_proj": true,
316
+ "language_model.model.layers.2.input_layernorm": false,
317
+ "language_model.model.layers.2.post_attention_layernorm": false,
318
+ "language_model.model.layers.3.self_attn.q_proj": true,
319
+ "language_model.model.layers.3.self_attn.k_proj": true,
320
+ "language_model.model.layers.3.self_attn.v_proj": true,
321
+ "language_model.model.layers.3.self_attn.o_proj": true,
322
+ "language_model.model.layers.3.mlp.gate_proj": true,
323
+ "language_model.model.layers.3.mlp.down_proj": true,
324
+ "language_model.model.layers.3.mlp.up_proj": true,
325
+ "language_model.model.layers.3.input_layernorm": false,
326
+ "language_model.model.layers.3.post_attention_layernorm": false,
327
+ "language_model.model.layers.4.self_attn.q_proj": true,
328
+ "language_model.model.layers.4.self_attn.k_proj": true,
329
+ "language_model.model.layers.4.self_attn.v_proj": true,
330
+ "language_model.model.layers.4.self_attn.o_proj": true,
331
+ "language_model.model.layers.4.mlp.gate_proj": true,
332
+ "language_model.model.layers.4.mlp.down_proj": true,
333
+ "language_model.model.layers.4.mlp.up_proj": true,
334
+ "language_model.model.layers.4.input_layernorm": false,
335
+ "language_model.model.layers.4.post_attention_layernorm": false,
336
+ "language_model.model.layers.5.self_attn.q_proj": true,
337
+ "language_model.model.layers.5.self_attn.k_proj": true,
338
+ "language_model.model.layers.5.self_attn.v_proj": true,
339
+ "language_model.model.layers.5.self_attn.o_proj": true,
340
+ "language_model.model.layers.5.mlp.gate_proj": true,
341
+ "language_model.model.layers.5.mlp.down_proj": true,
342
+ "language_model.model.layers.5.mlp.up_proj": true,
343
+ "language_model.model.layers.5.input_layernorm": false,
344
+ "language_model.model.layers.5.post_attention_layernorm": false,
345
+ "language_model.model.layers.6.self_attn.q_proj": true,
346
+ "language_model.model.layers.6.self_attn.k_proj": true,
347
+ "language_model.model.layers.6.self_attn.v_proj": true,
348
+ "language_model.model.layers.6.self_attn.o_proj": true,
349
+ "language_model.model.layers.6.mlp.gate_proj": true,
350
+ "language_model.model.layers.6.mlp.down_proj": true,
351
+ "language_model.model.layers.6.mlp.up_proj": true,
352
+ "language_model.model.layers.6.input_layernorm": false,
353
+ "language_model.model.layers.6.post_attention_layernorm": false,
354
+ "language_model.model.layers.7.self_attn.q_proj": true,
355
+ "language_model.model.layers.7.self_attn.k_proj": true,
356
+ "language_model.model.layers.7.self_attn.v_proj": true,
357
+ "language_model.model.layers.7.self_attn.o_proj": true,
358
+ "language_model.model.layers.7.mlp.gate_proj": true,
359
+ "language_model.model.layers.7.mlp.down_proj": true,
360
+ "language_model.model.layers.7.mlp.up_proj": true,
361
+ "language_model.model.layers.7.input_layernorm": false,
362
+ "language_model.model.layers.7.post_attention_layernorm": false,
363
+ "language_model.model.layers.8.self_attn.q_proj": true,
364
+ "language_model.model.layers.8.self_attn.k_proj": true,
365
+ "language_model.model.layers.8.self_attn.v_proj": true,
366
+ "language_model.model.layers.8.self_attn.o_proj": true,
367
+ "language_model.model.layers.8.mlp.gate_proj": true,
368
+ "language_model.model.layers.8.mlp.down_proj": true,
369
+ "language_model.model.layers.8.mlp.up_proj": true,
370
+ "language_model.model.layers.8.input_layernorm": false,
371
+ "language_model.model.layers.8.post_attention_layernorm": false,
372
+ "language_model.model.layers.9.self_attn.q_proj": true,
373
+ "language_model.model.layers.9.self_attn.k_proj": true,
374
+ "language_model.model.layers.9.self_attn.v_proj": true,
375
+ "language_model.model.layers.9.self_attn.o_proj": true,
376
+ "language_model.model.layers.9.mlp.gate_proj": true,
377
+ "language_model.model.layers.9.mlp.down_proj": true,
378
+ "language_model.model.layers.9.mlp.up_proj": true,
379
+ "language_model.model.layers.9.input_layernorm": false,
380
+ "language_model.model.layers.9.post_attention_layernorm": false,
381
+ "language_model.model.layers.10.self_attn.q_proj": true,
382
+ "language_model.model.layers.10.self_attn.k_proj": true,
383
+ "language_model.model.layers.10.self_attn.v_proj": true,
384
+ "language_model.model.layers.10.self_attn.o_proj": true,
385
+ "language_model.model.layers.10.mlp.gate_proj": true,
386
+ "language_model.model.layers.10.mlp.down_proj": true,
387
+ "language_model.model.layers.10.mlp.up_proj": true,
388
+ "language_model.model.layers.10.input_layernorm": false,
389
+ "language_model.model.layers.10.post_attention_layernorm": false,
390
+ "language_model.model.layers.11.self_attn.q_proj": true,
391
+ "language_model.model.layers.11.self_attn.k_proj": true,
392
+ "language_model.model.layers.11.self_attn.v_proj": true,
393
+ "language_model.model.layers.11.self_attn.o_proj": true,
394
+ "language_model.model.layers.11.mlp.gate_proj": true,
395
+ "language_model.model.layers.11.mlp.down_proj": true,
396
+ "language_model.model.layers.11.mlp.up_proj": true,
397
+ "language_model.model.layers.11.input_layernorm": false,
398
+ "language_model.model.layers.11.post_attention_layernorm": false,
399
+ "language_model.model.layers.12.self_attn.q_proj": true,
400
+ "language_model.model.layers.12.self_attn.k_proj": true,
401
+ "language_model.model.layers.12.self_attn.v_proj": true,
402
+ "language_model.model.layers.12.self_attn.o_proj": true,
403
+ "language_model.model.layers.12.mlp.gate_proj": true,
404
+ "language_model.model.layers.12.mlp.down_proj": true,
405
+ "language_model.model.layers.12.mlp.up_proj": true,
406
+ "language_model.model.layers.12.input_layernorm": false,
407
+ "language_model.model.layers.12.post_attention_layernorm": false,
408
+ "language_model.model.layers.13.self_attn.q_proj": true,
409
+ "language_model.model.layers.13.self_attn.k_proj": true,
410
+ "language_model.model.layers.13.self_attn.v_proj": true,
411
+ "language_model.model.layers.13.self_attn.o_proj": true,
412
+ "language_model.model.layers.13.mlp.gate_proj": true,
413
+ "language_model.model.layers.13.mlp.down_proj": true,
414
+ "language_model.model.layers.13.mlp.up_proj": true,
415
+ "language_model.model.layers.13.input_layernorm": false,
416
+ "language_model.model.layers.13.post_attention_layernorm": false,
417
+ "language_model.model.layers.14.self_attn.q_proj": true,
418
+ "language_model.model.layers.14.self_attn.k_proj": true,
419
+ "language_model.model.layers.14.self_attn.v_proj": true,
420
+ "language_model.model.layers.14.self_attn.o_proj": true,
421
+ "language_model.model.layers.14.mlp.gate_proj": true,
422
+ "language_model.model.layers.14.mlp.down_proj": true,
423
+ "language_model.model.layers.14.mlp.up_proj": true,
424
+ "language_model.model.layers.14.input_layernorm": false,
425
+ "language_model.model.layers.14.post_attention_layernorm": false,
426
+ "language_model.model.layers.15.self_attn.q_proj": true,
427
+ "language_model.model.layers.15.self_attn.k_proj": true,
428
+ "language_model.model.layers.15.self_attn.v_proj": true,
429
+ "language_model.model.layers.15.self_attn.o_proj": true,
430
+ "language_model.model.layers.15.mlp.gate_proj": true,
431
+ "language_model.model.layers.15.mlp.down_proj": true,
432
+ "language_model.model.layers.15.mlp.up_proj": true,
433
+ "language_model.model.layers.15.input_layernorm": false,
434
+ "language_model.model.layers.15.post_attention_layernorm": false,
435
+ "language_model.model.layers.16.self_attn.q_proj": true,
436
+ "language_model.model.layers.16.self_attn.k_proj": true,
437
+ "language_model.model.layers.16.self_attn.v_proj": true,
438
+ "language_model.model.layers.16.self_attn.o_proj": true,
439
+ "language_model.model.layers.16.mlp.gate_proj": true,
440
+ "language_model.model.layers.16.mlp.down_proj": true,
441
+ "language_model.model.layers.16.mlp.up_proj": true,
442
+ "language_model.model.layers.16.input_layernorm": false,
443
+ "language_model.model.layers.16.post_attention_layernorm": false,
444
+ "language_model.model.layers.17.self_attn.q_proj": true,
445
+ "language_model.model.layers.17.self_attn.k_proj": true,
446
+ "language_model.model.layers.17.self_attn.v_proj": true,
447
+ "language_model.model.layers.17.self_attn.o_proj": true,
448
+ "language_model.model.layers.17.mlp.gate_proj": true,
449
+ "language_model.model.layers.17.mlp.down_proj": true,
450
+ "language_model.model.layers.17.mlp.up_proj": true,
451
+ "language_model.model.layers.17.input_layernorm": false,
452
+ "language_model.model.layers.17.post_attention_layernorm": false,
453
+ "language_model.model.layers.18.self_attn.q_proj": true,
454
+ "language_model.model.layers.18.self_attn.k_proj": true,
455
+ "language_model.model.layers.18.self_attn.v_proj": true,
456
+ "language_model.model.layers.18.self_attn.o_proj": true,
457
+ "language_model.model.layers.18.mlp.gate_proj": true,
458
+ "language_model.model.layers.18.mlp.down_proj": true,
459
+ "language_model.model.layers.18.mlp.up_proj": true,
460
+ "language_model.model.layers.18.input_layernorm": false,
461
+ "language_model.model.layers.18.post_attention_layernorm": false,
462
+ "language_model.model.layers.19.self_attn.q_proj": true,
463
+ "language_model.model.layers.19.self_attn.k_proj": true,
464
+ "language_model.model.layers.19.self_attn.v_proj": true,
465
+ "language_model.model.layers.19.self_attn.o_proj": true,
466
+ "language_model.model.layers.19.mlp.gate_proj": true,
467
+ "language_model.model.layers.19.mlp.down_proj": true,
468
+ "language_model.model.layers.19.mlp.up_proj": true,
469
+ "language_model.model.layers.19.input_layernorm": false,
470
+ "language_model.model.layers.19.post_attention_layernorm": false,
471
+ "language_model.model.layers.20.self_attn.q_proj": true,
472
+ "language_model.model.layers.20.self_attn.k_proj": true,
473
+ "language_model.model.layers.20.self_attn.v_proj": true,
474
+ "language_model.model.layers.20.self_attn.o_proj": true,
475
+ "language_model.model.layers.20.mlp.gate_proj": true,
476
+ "language_model.model.layers.20.mlp.down_proj": true,
477
+ "language_model.model.layers.20.mlp.up_proj": true,
478
+ "language_model.model.layers.20.input_layernorm": false,
479
+ "language_model.model.layers.20.post_attention_layernorm": false,
480
+ "language_model.model.layers.21.self_attn.q_proj": true,
481
+ "language_model.model.layers.21.self_attn.k_proj": true,
482
+ "language_model.model.layers.21.self_attn.v_proj": true,
483
+ "language_model.model.layers.21.self_attn.o_proj": true,
484
+ "language_model.model.layers.21.mlp.gate_proj": true,
485
+ "language_model.model.layers.21.mlp.down_proj": true,
486
+ "language_model.model.layers.21.mlp.up_proj": true,
487
+ "language_model.model.layers.21.input_layernorm": false,
488
+ "language_model.model.layers.21.post_attention_layernorm": false,
489
+ "language_model.model.layers.22.self_attn.q_proj": true,
490
+ "language_model.model.layers.22.self_attn.k_proj": true,
491
+ "language_model.model.layers.22.self_attn.v_proj": true,
492
+ "language_model.model.layers.22.self_attn.o_proj": true,
493
+ "language_model.model.layers.22.mlp.gate_proj": true,
494
+ "language_model.model.layers.22.mlp.down_proj": true,
495
+ "language_model.model.layers.22.mlp.up_proj": true,
496
+ "language_model.model.layers.22.input_layernorm": false,
497
+ "language_model.model.layers.22.post_attention_layernorm": false,
498
+ "language_model.model.layers.23.self_attn.q_proj": true,
499
+ "language_model.model.layers.23.self_attn.k_proj": true,
500
+ "language_model.model.layers.23.self_attn.v_proj": true,
501
+ "language_model.model.layers.23.self_attn.o_proj": true,
502
+ "language_model.model.layers.23.mlp.gate_proj": true,
503
+ "language_model.model.layers.23.mlp.down_proj": true,
504
+ "language_model.model.layers.23.mlp.up_proj": true,
505
+ "language_model.model.layers.23.input_layernorm": false,
506
+ "language_model.model.layers.23.post_attention_layernorm": false,
507
+ "language_model.model.layers.24.self_attn.q_proj": true,
508
+ "language_model.model.layers.24.self_attn.k_proj": true,
509
+ "language_model.model.layers.24.self_attn.v_proj": true,
510
+ "language_model.model.layers.24.self_attn.o_proj": true,
511
+ "language_model.model.layers.24.mlp.gate_proj": true,
512
+ "language_model.model.layers.24.mlp.down_proj": true,
513
+ "language_model.model.layers.24.mlp.up_proj": true,
514
+ "language_model.model.layers.24.input_layernorm": false,
515
+ "language_model.model.layers.24.post_attention_layernorm": false,
516
+ "language_model.model.layers.25.self_attn.q_proj": true,
517
+ "language_model.model.layers.25.self_attn.k_proj": true,
518
+ "language_model.model.layers.25.self_attn.v_proj": true,
519
+ "language_model.model.layers.25.self_attn.o_proj": true,
520
+ "language_model.model.layers.25.mlp.gate_proj": true,
521
+ "language_model.model.layers.25.mlp.down_proj": true,
522
+ "language_model.model.layers.25.mlp.up_proj": true,
523
+ "language_model.model.layers.25.input_layernorm": false,
524
+ "language_model.model.layers.25.post_attention_layernorm": false,
525
+ "language_model.model.layers.26.self_attn.q_proj": true,
526
+ "language_model.model.layers.26.self_attn.k_proj": true,
527
+ "language_model.model.layers.26.self_attn.v_proj": true,
528
+ "language_model.model.layers.26.self_attn.o_proj": true,
529
+ "language_model.model.layers.26.mlp.gate_proj": true,
530
+ "language_model.model.layers.26.mlp.down_proj": true,
531
+ "language_model.model.layers.26.mlp.up_proj": true,
532
+ "language_model.model.layers.26.input_layernorm": false,
533
+ "language_model.model.layers.26.post_attention_layernorm": false,
534
+ "language_model.model.layers.27.self_attn.q_proj": true,
535
+ "language_model.model.layers.27.self_attn.k_proj": true,
536
+ "language_model.model.layers.27.self_attn.v_proj": true,
537
+ "language_model.model.layers.27.self_attn.o_proj": true,
538
+ "language_model.model.layers.27.mlp.gate_proj": true,
539
+ "language_model.model.layers.27.mlp.down_proj": true,
540
+ "language_model.model.layers.27.mlp.up_proj": true,
541
+ "language_model.model.layers.27.input_layernorm": false,
542
+ "language_model.model.layers.27.post_attention_layernorm": false,
543
+ "language_model.model.norm": false,
544
+ "language_model.lm_head": true
545
+ },
546
+ "quantization_config": {
547
+ "group_size": 64,
548
+ "bits": 4,
549
+ "vision_tower.patch_embed.proj": false,
550
+ "vision_tower.rotary_pos_emb": false,
551
+ "vision_tower.blocks.0.norm1": false,
552
+ "vision_tower.blocks.0.norm2": false,
553
+ "vision_tower.blocks.0.attn.qkv": false,
554
+ "vision_tower.blocks.0.attn.proj": false,
555
+ "vision_tower.blocks.0.mlp.gate_proj": false,
556
+ "vision_tower.blocks.0.mlp.up_proj": false,
557
+ "vision_tower.blocks.0.mlp.down_proj": false,
558
+ "vision_tower.blocks.1.norm1": false,
559
+ "vision_tower.blocks.1.norm2": false,
560
+ "vision_tower.blocks.1.attn.qkv": false,
561
+ "vision_tower.blocks.1.attn.proj": false,
562
+ "vision_tower.blocks.1.mlp.gate_proj": false,
563
+ "vision_tower.blocks.1.mlp.up_proj": false,
564
+ "vision_tower.blocks.1.mlp.down_proj": false,
565
+ "vision_tower.blocks.2.norm1": false,
566
+ "vision_tower.blocks.2.norm2": false,
567
+ "vision_tower.blocks.2.attn.qkv": false,
568
+ "vision_tower.blocks.2.attn.proj": false,
569
+ "vision_tower.blocks.2.mlp.gate_proj": false,
570
+ "vision_tower.blocks.2.mlp.up_proj": false,
571
+ "vision_tower.blocks.2.mlp.down_proj": false,
572
+ "vision_tower.blocks.3.norm1": false,
573
+ "vision_tower.blocks.3.norm2": false,
574
+ "vision_tower.blocks.3.attn.qkv": false,
575
+ "vision_tower.blocks.3.attn.proj": false,
576
+ "vision_tower.blocks.3.mlp.gate_proj": false,
577
+ "vision_tower.blocks.3.mlp.up_proj": false,
578
+ "vision_tower.blocks.3.mlp.down_proj": false,
579
+ "vision_tower.blocks.4.norm1": false,
580
+ "vision_tower.blocks.4.norm2": false,
581
+ "vision_tower.blocks.4.attn.qkv": false,
582
+ "vision_tower.blocks.4.attn.proj": false,
583
+ "vision_tower.blocks.4.mlp.gate_proj": false,
584
+ "vision_tower.blocks.4.mlp.up_proj": false,
585
+ "vision_tower.blocks.4.mlp.down_proj": false,
586
+ "vision_tower.blocks.5.norm1": false,
587
+ "vision_tower.blocks.5.norm2": false,
588
+ "vision_tower.blocks.5.attn.qkv": false,
589
+ "vision_tower.blocks.5.attn.proj": false,
590
+ "vision_tower.blocks.5.mlp.gate_proj": false,
591
+ "vision_tower.blocks.5.mlp.up_proj": false,
592
+ "vision_tower.blocks.5.mlp.down_proj": false,
593
+ "vision_tower.blocks.6.norm1": false,
594
+ "vision_tower.blocks.6.norm2": false,
595
+ "vision_tower.blocks.6.attn.qkv": false,
596
+ "vision_tower.blocks.6.attn.proj": false,
597
+ "vision_tower.blocks.6.mlp.gate_proj": false,
598
+ "vision_tower.blocks.6.mlp.up_proj": false,
599
+ "vision_tower.blocks.6.mlp.down_proj": false,
600
+ "vision_tower.blocks.7.norm1": false,
601
+ "vision_tower.blocks.7.norm2": false,
602
+ "vision_tower.blocks.7.attn.qkv": false,
603
+ "vision_tower.blocks.7.attn.proj": false,
604
+ "vision_tower.blocks.7.mlp.gate_proj": false,
605
+ "vision_tower.blocks.7.mlp.up_proj": false,
606
+ "vision_tower.blocks.7.mlp.down_proj": false,
607
+ "vision_tower.blocks.8.norm1": false,
608
+ "vision_tower.blocks.8.norm2": false,
609
+ "vision_tower.blocks.8.attn.qkv": false,
610
+ "vision_tower.blocks.8.attn.proj": false,
611
+ "vision_tower.blocks.8.mlp.gate_proj": false,
612
+ "vision_tower.blocks.8.mlp.up_proj": false,
613
+ "vision_tower.blocks.8.mlp.down_proj": false,
614
+ "vision_tower.blocks.9.norm1": false,
615
+ "vision_tower.blocks.9.norm2": false,
616
+ "vision_tower.blocks.9.attn.qkv": false,
617
+ "vision_tower.blocks.9.attn.proj": false,
618
+ "vision_tower.blocks.9.mlp.gate_proj": false,
619
+ "vision_tower.blocks.9.mlp.up_proj": false,
620
+ "vision_tower.blocks.9.mlp.down_proj": false,
621
+ "vision_tower.blocks.10.norm1": false,
622
+ "vision_tower.blocks.10.norm2": false,
623
+ "vision_tower.blocks.10.attn.qkv": false,
624
+ "vision_tower.blocks.10.attn.proj": false,
625
+ "vision_tower.blocks.10.mlp.gate_proj": false,
626
+ "vision_tower.blocks.10.mlp.up_proj": false,
627
+ "vision_tower.blocks.10.mlp.down_proj": false,
628
+ "vision_tower.blocks.11.norm1": false,
629
+ "vision_tower.blocks.11.norm2": false,
630
+ "vision_tower.blocks.11.attn.qkv": false,
631
+ "vision_tower.blocks.11.attn.proj": false,
632
+ "vision_tower.blocks.11.mlp.gate_proj": false,
633
+ "vision_tower.blocks.11.mlp.up_proj": false,
634
+ "vision_tower.blocks.11.mlp.down_proj": false,
635
+ "vision_tower.blocks.12.norm1": false,
636
+ "vision_tower.blocks.12.norm2": false,
637
+ "vision_tower.blocks.12.attn.qkv": false,
638
+ "vision_tower.blocks.12.attn.proj": false,
639
+ "vision_tower.blocks.12.mlp.gate_proj": false,
640
+ "vision_tower.blocks.12.mlp.up_proj": false,
641
+ "vision_tower.blocks.12.mlp.down_proj": false,
642
+ "vision_tower.blocks.13.norm1": false,
643
+ "vision_tower.blocks.13.norm2": false,
644
+ "vision_tower.blocks.13.attn.qkv": false,
645
+ "vision_tower.blocks.13.attn.proj": false,
646
+ "vision_tower.blocks.13.mlp.gate_proj": false,
647
+ "vision_tower.blocks.13.mlp.up_proj": false,
648
+ "vision_tower.blocks.13.mlp.down_proj": false,
649
+ "vision_tower.blocks.14.norm1": false,
650
+ "vision_tower.blocks.14.norm2": false,
651
+ "vision_tower.blocks.14.attn.qkv": false,
652
+ "vision_tower.blocks.14.attn.proj": false,
653
+ "vision_tower.blocks.14.mlp.gate_proj": false,
654
+ "vision_tower.blocks.14.mlp.up_proj": false,
655
+ "vision_tower.blocks.14.mlp.down_proj": false,
656
+ "vision_tower.blocks.15.norm1": false,
657
+ "vision_tower.blocks.15.norm2": false,
658
+ "vision_tower.blocks.15.attn.qkv": false,
659
+ "vision_tower.blocks.15.attn.proj": false,
660
+ "vision_tower.blocks.15.mlp.gate_proj": false,
661
+ "vision_tower.blocks.15.mlp.up_proj": false,
662
+ "vision_tower.blocks.15.mlp.down_proj": false,
663
+ "vision_tower.blocks.16.norm1": false,
664
+ "vision_tower.blocks.16.norm2": false,
665
+ "vision_tower.blocks.16.attn.qkv": false,
666
+ "vision_tower.blocks.16.attn.proj": false,
667
+ "vision_tower.blocks.16.mlp.gate_proj": false,
668
+ "vision_tower.blocks.16.mlp.up_proj": false,
669
+ "vision_tower.blocks.16.mlp.down_proj": false,
670
+ "vision_tower.blocks.17.norm1": false,
671
+ "vision_tower.blocks.17.norm2": false,
672
+ "vision_tower.blocks.17.attn.qkv": false,
673
+ "vision_tower.blocks.17.attn.proj": false,
674
+ "vision_tower.blocks.17.mlp.gate_proj": false,
675
+ "vision_tower.blocks.17.mlp.up_proj": false,
676
+ "vision_tower.blocks.17.mlp.down_proj": false,
677
+ "vision_tower.blocks.18.norm1": false,
678
+ "vision_tower.blocks.18.norm2": false,
679
+ "vision_tower.blocks.18.attn.qkv": false,
680
+ "vision_tower.blocks.18.attn.proj": false,
681
+ "vision_tower.blocks.18.mlp.gate_proj": false,
682
+ "vision_tower.blocks.18.mlp.up_proj": false,
683
+ "vision_tower.blocks.18.mlp.down_proj": false,
684
+ "vision_tower.blocks.19.norm1": false,
685
+ "vision_tower.blocks.19.norm2": false,
686
+ "vision_tower.blocks.19.attn.qkv": false,
687
+ "vision_tower.blocks.19.attn.proj": false,
688
+ "vision_tower.blocks.19.mlp.gate_proj": false,
689
+ "vision_tower.blocks.19.mlp.up_proj": false,
690
+ "vision_tower.blocks.19.mlp.down_proj": false,
691
+ "vision_tower.blocks.20.norm1": false,
692
+ "vision_tower.blocks.20.norm2": false,
693
+ "vision_tower.blocks.20.attn.qkv": false,
694
+ "vision_tower.blocks.20.attn.proj": false,
695
+ "vision_tower.blocks.20.mlp.gate_proj": false,
696
+ "vision_tower.blocks.20.mlp.up_proj": false,
697
+ "vision_tower.blocks.20.mlp.down_proj": false,
698
+ "vision_tower.blocks.21.norm1": false,
699
+ "vision_tower.blocks.21.norm2": false,
700
+ "vision_tower.blocks.21.attn.qkv": false,
701
+ "vision_tower.blocks.21.attn.proj": false,
702
+ "vision_tower.blocks.21.mlp.gate_proj": false,
703
+ "vision_tower.blocks.21.mlp.up_proj": false,
704
+ "vision_tower.blocks.21.mlp.down_proj": false,
705
+ "vision_tower.blocks.22.norm1": false,
706
+ "vision_tower.blocks.22.norm2": false,
707
+ "vision_tower.blocks.22.attn.qkv": false,
708
+ "vision_tower.blocks.22.attn.proj": false,
709
+ "vision_tower.blocks.22.mlp.gate_proj": false,
710
+ "vision_tower.blocks.22.mlp.up_proj": false,
711
+ "vision_tower.blocks.22.mlp.down_proj": false,
712
+ "vision_tower.blocks.23.norm1": false,
713
+ "vision_tower.blocks.23.norm2": false,
714
+ "vision_tower.blocks.23.attn.qkv": false,
715
+ "vision_tower.blocks.23.attn.proj": false,
716
+ "vision_tower.blocks.23.mlp.gate_proj": false,
717
+ "vision_tower.blocks.23.mlp.up_proj": false,
718
+ "vision_tower.blocks.23.mlp.down_proj": false,
719
+ "vision_tower.blocks.24.norm1": false,
720
+ "vision_tower.blocks.24.norm2": false,
721
+ "vision_tower.blocks.24.attn.qkv": false,
722
+ "vision_tower.blocks.24.attn.proj": false,
723
+ "vision_tower.blocks.24.mlp.gate_proj": false,
724
+ "vision_tower.blocks.24.mlp.up_proj": false,
725
+ "vision_tower.blocks.24.mlp.down_proj": false,
726
+ "vision_tower.blocks.25.norm1": false,
727
+ "vision_tower.blocks.25.norm2": false,
728
+ "vision_tower.blocks.25.attn.qkv": false,
729
+ "vision_tower.blocks.25.attn.proj": false,
730
+ "vision_tower.blocks.25.mlp.gate_proj": false,
731
+ "vision_tower.blocks.25.mlp.up_proj": false,
732
+ "vision_tower.blocks.25.mlp.down_proj": false,
733
+ "vision_tower.blocks.26.norm1": false,
734
+ "vision_tower.blocks.26.norm2": false,
735
+ "vision_tower.blocks.26.attn.qkv": false,
736
+ "vision_tower.blocks.26.attn.proj": false,
737
+ "vision_tower.blocks.26.mlp.gate_proj": false,
738
+ "vision_tower.blocks.26.mlp.up_proj": false,
739
+ "vision_tower.blocks.26.mlp.down_proj": false,
740
+ "vision_tower.blocks.27.norm1": false,
741
+ "vision_tower.blocks.27.norm2": false,
742
+ "vision_tower.blocks.27.attn.qkv": false,
743
+ "vision_tower.blocks.27.attn.proj": false,
744
+ "vision_tower.blocks.27.mlp.gate_proj": false,
745
+ "vision_tower.blocks.27.mlp.up_proj": false,
746
+ "vision_tower.blocks.27.mlp.down_proj": false,
747
+ "vision_tower.blocks.28.norm1": false,
748
+ "vision_tower.blocks.28.norm2": false,
749
+ "vision_tower.blocks.28.attn.qkv": false,
750
+ "vision_tower.blocks.28.attn.proj": false,
751
+ "vision_tower.blocks.28.mlp.gate_proj": false,
752
+ "vision_tower.blocks.28.mlp.up_proj": false,
753
+ "vision_tower.blocks.28.mlp.down_proj": false,
754
+ "vision_tower.blocks.29.norm1": false,
755
+ "vision_tower.blocks.29.norm2": false,
756
+ "vision_tower.blocks.29.attn.qkv": false,
757
+ "vision_tower.blocks.29.attn.proj": false,
758
+ "vision_tower.blocks.29.mlp.gate_proj": false,
759
+ "vision_tower.blocks.29.mlp.up_proj": false,
760
+ "vision_tower.blocks.29.mlp.down_proj": false,
761
+ "vision_tower.blocks.30.norm1": false,
762
+ "vision_tower.blocks.30.norm2": false,
763
+ "vision_tower.blocks.30.attn.qkv": false,
764
+ "vision_tower.blocks.30.attn.proj": false,
765
+ "vision_tower.blocks.30.mlp.gate_proj": false,
766
+ "vision_tower.blocks.30.mlp.up_proj": false,
767
+ "vision_tower.blocks.30.mlp.down_proj": false,
768
+ "vision_tower.blocks.31.norm1": false,
769
+ "vision_tower.blocks.31.norm2": false,
770
+ "vision_tower.blocks.31.attn.qkv": false,
771
+ "vision_tower.blocks.31.attn.proj": false,
772
+ "vision_tower.blocks.31.mlp.gate_proj": false,
773
+ "vision_tower.blocks.31.mlp.up_proj": false,
774
+ "vision_tower.blocks.31.mlp.down_proj": false,
775
+ "vision_tower.merger.ln_q": false,
776
+ "vision_tower.merger.mlp.0": false,
777
+ "vision_tower.merger.mlp.1": false,
778
+ "vision_tower.merger.mlp.2": false,
779
+ "language_model.model.embed_tokens": true,
780
+ "language_model.model.layers.0.self_attn.q_proj": true,
781
+ "language_model.model.layers.0.self_attn.k_proj": true,
782
+ "language_model.model.layers.0.self_attn.v_proj": true,
783
+ "language_model.model.layers.0.self_attn.o_proj": true,
784
+ "language_model.model.layers.0.mlp.gate_proj": true,
785
+ "language_model.model.layers.0.mlp.down_proj": true,
786
+ "language_model.model.layers.0.mlp.up_proj": true,
787
+ "language_model.model.layers.0.input_layernorm": false,
788
+ "language_model.model.layers.0.post_attention_layernorm": false,
789
+ "language_model.model.layers.1.self_attn.q_proj": true,
790
+ "language_model.model.layers.1.self_attn.k_proj": true,
791
+ "language_model.model.layers.1.self_attn.v_proj": true,
792
+ "language_model.model.layers.1.self_attn.o_proj": true,
793
+ "language_model.model.layers.1.mlp.gate_proj": true,
794
+ "language_model.model.layers.1.mlp.down_proj": true,
795
+ "language_model.model.layers.1.mlp.up_proj": true,
796
+ "language_model.model.layers.1.input_layernorm": false,
797
+ "language_model.model.layers.1.post_attention_layernorm": false,
798
+ "language_model.model.layers.2.self_attn.q_proj": true,
799
+ "language_model.model.layers.2.self_attn.k_proj": true,
800
+ "language_model.model.layers.2.self_attn.v_proj": true,
801
+ "language_model.model.layers.2.self_attn.o_proj": true,
802
+ "language_model.model.layers.2.mlp.gate_proj": true,
803
+ "language_model.model.layers.2.mlp.down_proj": true,
804
+ "language_model.model.layers.2.mlp.up_proj": true,
805
+ "language_model.model.layers.2.input_layernorm": false,
806
+ "language_model.model.layers.2.post_attention_layernorm": false,
807
+ "language_model.model.layers.3.self_attn.q_proj": true,
808
+ "language_model.model.layers.3.self_attn.k_proj": true,
809
+ "language_model.model.layers.3.self_attn.v_proj": true,
810
+ "language_model.model.layers.3.self_attn.o_proj": true,
811
+ "language_model.model.layers.3.mlp.gate_proj": true,
812
+ "language_model.model.layers.3.mlp.down_proj": true,
813
+ "language_model.model.layers.3.mlp.up_proj": true,
814
+ "language_model.model.layers.3.input_layernorm": false,
815
+ "language_model.model.layers.3.post_attention_layernorm": false,
816
+ "language_model.model.layers.4.self_attn.q_proj": true,
817
+ "language_model.model.layers.4.self_attn.k_proj": true,
818
+ "language_model.model.layers.4.self_attn.v_proj": true,
819
+ "language_model.model.layers.4.self_attn.o_proj": true,
820
+ "language_model.model.layers.4.mlp.gate_proj": true,
821
+ "language_model.model.layers.4.mlp.down_proj": true,
822
+ "language_model.model.layers.4.mlp.up_proj": true,
823
+ "language_model.model.layers.4.input_layernorm": false,
824
+ "language_model.model.layers.4.post_attention_layernorm": false,
825
+ "language_model.model.layers.5.self_attn.q_proj": true,
826
+ "language_model.model.layers.5.self_attn.k_proj": true,
827
+ "language_model.model.layers.5.self_attn.v_proj": true,
828
+ "language_model.model.layers.5.self_attn.o_proj": true,
829
+ "language_model.model.layers.5.mlp.gate_proj": true,
830
+ "language_model.model.layers.5.mlp.down_proj": true,
831
+ "language_model.model.layers.5.mlp.up_proj": true,
832
+ "language_model.model.layers.5.input_layernorm": false,
833
+ "language_model.model.layers.5.post_attention_layernorm": false,
834
+ "language_model.model.layers.6.self_attn.q_proj": true,
835
+ "language_model.model.layers.6.self_attn.k_proj": true,
836
+ "language_model.model.layers.6.self_attn.v_proj": true,
837
+ "language_model.model.layers.6.self_attn.o_proj": true,
838
+ "language_model.model.layers.6.mlp.gate_proj": true,
839
+ "language_model.model.layers.6.mlp.down_proj": true,
840
+ "language_model.model.layers.6.mlp.up_proj": true,
841
+ "language_model.model.layers.6.input_layernorm": false,
842
+ "language_model.model.layers.6.post_attention_layernorm": false,
843
+ "language_model.model.layers.7.self_attn.q_proj": true,
844
+ "language_model.model.layers.7.self_attn.k_proj": true,
845
+ "language_model.model.layers.7.self_attn.v_proj": true,
846
+ "language_model.model.layers.7.self_attn.o_proj": true,
847
+ "language_model.model.layers.7.mlp.gate_proj": true,
848
+ "language_model.model.layers.7.mlp.down_proj": true,
849
+ "language_model.model.layers.7.mlp.up_proj": true,
850
+ "language_model.model.layers.7.input_layernorm": false,
851
+ "language_model.model.layers.7.post_attention_layernorm": false,
852
+ "language_model.model.layers.8.self_attn.q_proj": true,
853
+ "language_model.model.layers.8.self_attn.k_proj": true,
854
+ "language_model.model.layers.8.self_attn.v_proj": true,
855
+ "language_model.model.layers.8.self_attn.o_proj": true,
856
+ "language_model.model.layers.8.mlp.gate_proj": true,
857
+ "language_model.model.layers.8.mlp.down_proj": true,
858
+ "language_model.model.layers.8.mlp.up_proj": true,
859
+ "language_model.model.layers.8.input_layernorm": false,
860
+ "language_model.model.layers.8.post_attention_layernorm": false,
861
+ "language_model.model.layers.9.self_attn.q_proj": true,
862
+ "language_model.model.layers.9.self_attn.k_proj": true,
863
+ "language_model.model.layers.9.self_attn.v_proj": true,
864
+ "language_model.model.layers.9.self_attn.o_proj": true,
865
+ "language_model.model.layers.9.mlp.gate_proj": true,
866
+ "language_model.model.layers.9.mlp.down_proj": true,
867
+ "language_model.model.layers.9.mlp.up_proj": true,
868
+ "language_model.model.layers.9.input_layernorm": false,
869
+ "language_model.model.layers.9.post_attention_layernorm": false,
870
+ "language_model.model.layers.10.self_attn.q_proj": true,
871
+ "language_model.model.layers.10.self_attn.k_proj": true,
872
+ "language_model.model.layers.10.self_attn.v_proj": true,
873
+ "language_model.model.layers.10.self_attn.o_proj": true,
874
+ "language_model.model.layers.10.mlp.gate_proj": true,
875
+ "language_model.model.layers.10.mlp.down_proj": true,
876
+ "language_model.model.layers.10.mlp.up_proj": true,
877
+ "language_model.model.layers.10.input_layernorm": false,
878
+ "language_model.model.layers.10.post_attention_layernorm": false,
879
+ "language_model.model.layers.11.self_attn.q_proj": true,
880
+ "language_model.model.layers.11.self_attn.k_proj": true,
881
+ "language_model.model.layers.11.self_attn.v_proj": true,
882
+ "language_model.model.layers.11.self_attn.o_proj": true,
883
+ "language_model.model.layers.11.mlp.gate_proj": true,
884
+ "language_model.model.layers.11.mlp.down_proj": true,
885
+ "language_model.model.layers.11.mlp.up_proj": true,
886
+ "language_model.model.layers.11.input_layernorm": false,
887
+ "language_model.model.layers.11.post_attention_layernorm": false,
888
+ "language_model.model.layers.12.self_attn.q_proj": true,
889
+ "language_model.model.layers.12.self_attn.k_proj": true,
890
+ "language_model.model.layers.12.self_attn.v_proj": true,
891
+ "language_model.model.layers.12.self_attn.o_proj": true,
892
+ "language_model.model.layers.12.mlp.gate_proj": true,
893
+ "language_model.model.layers.12.mlp.down_proj": true,
894
+ "language_model.model.layers.12.mlp.up_proj": true,
895
+ "language_model.model.layers.12.input_layernorm": false,
896
+ "language_model.model.layers.12.post_attention_layernorm": false,
897
+ "language_model.model.layers.13.self_attn.q_proj": true,
898
+ "language_model.model.layers.13.self_attn.k_proj": true,
899
+ "language_model.model.layers.13.self_attn.v_proj": true,
900
+ "language_model.model.layers.13.self_attn.o_proj": true,
901
+ "language_model.model.layers.13.mlp.gate_proj": true,
902
+ "language_model.model.layers.13.mlp.down_proj": true,
903
+ "language_model.model.layers.13.mlp.up_proj": true,
904
+ "language_model.model.layers.13.input_layernorm": false,
905
+ "language_model.model.layers.13.post_attention_layernorm": false,
906
+ "language_model.model.layers.14.self_attn.q_proj": true,
907
+ "language_model.model.layers.14.self_attn.k_proj": true,
908
+ "language_model.model.layers.14.self_attn.v_proj": true,
909
+ "language_model.model.layers.14.self_attn.o_proj": true,
910
+ "language_model.model.layers.14.mlp.gate_proj": true,
911
+ "language_model.model.layers.14.mlp.down_proj": true,
912
+ "language_model.model.layers.14.mlp.up_proj": true,
913
+ "language_model.model.layers.14.input_layernorm": false,
914
+ "language_model.model.layers.14.post_attention_layernorm": false,
915
+ "language_model.model.layers.15.self_attn.q_proj": true,
916
+ "language_model.model.layers.15.self_attn.k_proj": true,
917
+ "language_model.model.layers.15.self_attn.v_proj": true,
918
+ "language_model.model.layers.15.self_attn.o_proj": true,
919
+ "language_model.model.layers.15.mlp.gate_proj": true,
920
+ "language_model.model.layers.15.mlp.down_proj": true,
921
+ "language_model.model.layers.15.mlp.up_proj": true,
922
+ "language_model.model.layers.15.input_layernorm": false,
923
+ "language_model.model.layers.15.post_attention_layernorm": false,
924
+ "language_model.model.layers.16.self_attn.q_proj": true,
925
+ "language_model.model.layers.16.self_attn.k_proj": true,
926
+ "language_model.model.layers.16.self_attn.v_proj": true,
927
+ "language_model.model.layers.16.self_attn.o_proj": true,
928
+ "language_model.model.layers.16.mlp.gate_proj": true,
929
+ "language_model.model.layers.16.mlp.down_proj": true,
930
+ "language_model.model.layers.16.mlp.up_proj": true,
931
+ "language_model.model.layers.16.input_layernorm": false,
932
+ "language_model.model.layers.16.post_attention_layernorm": false,
933
+ "language_model.model.layers.17.self_attn.q_proj": true,
934
+ "language_model.model.layers.17.self_attn.k_proj": true,
935
+ "language_model.model.layers.17.self_attn.v_proj": true,
936
+ "language_model.model.layers.17.self_attn.o_proj": true,
937
+ "language_model.model.layers.17.mlp.gate_proj": true,
938
+ "language_model.model.layers.17.mlp.down_proj": true,
939
+ "language_model.model.layers.17.mlp.up_proj": true,
940
+ "language_model.model.layers.17.input_layernorm": false,
941
+ "language_model.model.layers.17.post_attention_layernorm": false,
942
+ "language_model.model.layers.18.self_attn.q_proj": true,
943
+ "language_model.model.layers.18.self_attn.k_proj": true,
944
+ "language_model.model.layers.18.self_attn.v_proj": true,
945
+ "language_model.model.layers.18.self_attn.o_proj": true,
946
+ "language_model.model.layers.18.mlp.gate_proj": true,
947
+ "language_model.model.layers.18.mlp.down_proj": true,
948
+ "language_model.model.layers.18.mlp.up_proj": true,
949
+ "language_model.model.layers.18.input_layernorm": false,
950
+ "language_model.model.layers.18.post_attention_layernorm": false,
951
+ "language_model.model.layers.19.self_attn.q_proj": true,
952
+ "language_model.model.layers.19.self_attn.k_proj": true,
953
+ "language_model.model.layers.19.self_attn.v_proj": true,
954
+ "language_model.model.layers.19.self_attn.o_proj": true,
955
+ "language_model.model.layers.19.mlp.gate_proj": true,
956
+ "language_model.model.layers.19.mlp.down_proj": true,
957
+ "language_model.model.layers.19.mlp.up_proj": true,
958
+ "language_model.model.layers.19.input_layernorm": false,
959
+ "language_model.model.layers.19.post_attention_layernorm": false,
960
+ "language_model.model.layers.20.self_attn.q_proj": true,
961
+ "language_model.model.layers.20.self_attn.k_proj": true,
962
+ "language_model.model.layers.20.self_attn.v_proj": true,
963
+ "language_model.model.layers.20.self_attn.o_proj": true,
964
+ "language_model.model.layers.20.mlp.gate_proj": true,
965
+ "language_model.model.layers.20.mlp.down_proj": true,
966
+ "language_model.model.layers.20.mlp.up_proj": true,
967
+ "language_model.model.layers.20.input_layernorm": false,
968
+ "language_model.model.layers.20.post_attention_layernorm": false,
969
+ "language_model.model.layers.21.self_attn.q_proj": true,
970
+ "language_model.model.layers.21.self_attn.k_proj": true,
971
+ "language_model.model.layers.21.self_attn.v_proj": true,
972
+ "language_model.model.layers.21.self_attn.o_proj": true,
973
+ "language_model.model.layers.21.mlp.gate_proj": true,
974
+ "language_model.model.layers.21.mlp.down_proj": true,
975
+ "language_model.model.layers.21.mlp.up_proj": true,
976
+ "language_model.model.layers.21.input_layernorm": false,
977
+ "language_model.model.layers.21.post_attention_layernorm": false,
978
+ "language_model.model.layers.22.self_attn.q_proj": true,
979
+ "language_model.model.layers.22.self_attn.k_proj": true,
980
+ "language_model.model.layers.22.self_attn.v_proj": true,
981
+ "language_model.model.layers.22.self_attn.o_proj": true,
982
+ "language_model.model.layers.22.mlp.gate_proj": true,
983
+ "language_model.model.layers.22.mlp.down_proj": true,
984
+ "language_model.model.layers.22.mlp.up_proj": true,
985
+ "language_model.model.layers.22.input_layernorm": false,
986
+ "language_model.model.layers.22.post_attention_layernorm": false,
987
+ "language_model.model.layers.23.self_attn.q_proj": true,
988
+ "language_model.model.layers.23.self_attn.k_proj": true,
989
+ "language_model.model.layers.23.self_attn.v_proj": true,
990
+ "language_model.model.layers.23.self_attn.o_proj": true,
991
+ "language_model.model.layers.23.mlp.gate_proj": true,
992
+ "language_model.model.layers.23.mlp.down_proj": true,
993
+ "language_model.model.layers.23.mlp.up_proj": true,
994
+ "language_model.model.layers.23.input_layernorm": false,
995
+ "language_model.model.layers.23.post_attention_layernorm": false,
996
+ "language_model.model.layers.24.self_attn.q_proj": true,
997
+ "language_model.model.layers.24.self_attn.k_proj": true,
998
+ "language_model.model.layers.24.self_attn.v_proj": true,
999
+ "language_model.model.layers.24.self_attn.o_proj": true,
1000
+ "language_model.model.layers.24.mlp.gate_proj": true,
1001
+ "language_model.model.layers.24.mlp.down_proj": true,
1002
+ "language_model.model.layers.24.mlp.up_proj": true,
1003
+ "language_model.model.layers.24.input_layernorm": false,
1004
+ "language_model.model.layers.24.post_attention_layernorm": false,
1005
+ "language_model.model.layers.25.self_attn.q_proj": true,
1006
+ "language_model.model.layers.25.self_attn.k_proj": true,
1007
+ "language_model.model.layers.25.self_attn.v_proj": true,
1008
+ "language_model.model.layers.25.self_attn.o_proj": true,
1009
+ "language_model.model.layers.25.mlp.gate_proj": true,
1010
+ "language_model.model.layers.25.mlp.down_proj": true,
1011
+ "language_model.model.layers.25.mlp.up_proj": true,
1012
+ "language_model.model.layers.25.input_layernorm": false,
1013
+ "language_model.model.layers.25.post_attention_layernorm": false,
1014
+ "language_model.model.layers.26.self_attn.q_proj": true,
1015
+ "language_model.model.layers.26.self_attn.k_proj": true,
1016
+ "language_model.model.layers.26.self_attn.v_proj": true,
1017
+ "language_model.model.layers.26.self_attn.o_proj": true,
1018
+ "language_model.model.layers.26.mlp.gate_proj": true,
1019
+ "language_model.model.layers.26.mlp.down_proj": true,
1020
+ "language_model.model.layers.26.mlp.up_proj": true,
1021
+ "language_model.model.layers.26.input_layernorm": false,
1022
+ "language_model.model.layers.26.post_attention_layernorm": false,
1023
+ "language_model.model.layers.27.self_attn.q_proj": true,
1024
+ "language_model.model.layers.27.self_attn.k_proj": true,
1025
+ "language_model.model.layers.27.self_attn.v_proj": true,
1026
+ "language_model.model.layers.27.self_attn.o_proj": true,
1027
+ "language_model.model.layers.27.mlp.gate_proj": true,
1028
+ "language_model.model.layers.27.mlp.down_proj": true,
1029
+ "language_model.model.layers.27.mlp.up_proj": true,
1030
+ "language_model.model.layers.27.input_layernorm": false,
1031
+ "language_model.model.layers.27.post_attention_layernorm": false,
1032
+ "language_model.model.norm": false,
1033
+ "language_model.lm_head": true
1034
  },
1035
  "remove_invalid_values": false,
1036
  "repetition_penalty": 1.0,
 
1052
  "suppress_tokens": null,
1053
  "task_specific_params": null,
1054
  "temperature": 1.0,
1055
+ "text_config": {
1056
+ "vocab_size": 152064,
1057
+ "max_position_embeddings": 128000,
1058
+ "hidden_size": 3584,
1059
+ "intermediate_size": 18944,
1060
+ "num_hidden_layers": 28,
1061
+ "num_attention_heads": 28,
1062
+ "use_sliding_window": false,
1063
+ "sliding_window": null,
1064
+ "max_window_layers": 28,
1065
+ "num_key_value_heads": 4,
1066
+ "hidden_act": "silu",
1067
+ "initializer_range": 0.02,
1068
+ "rms_norm_eps": 1e-06,
1069
+ "use_cache": false,
1070
+ "rope_theta": 1000000.0,
1071
+ "attention_dropout": 0.0,
1072
+ "rope_scaling": {
1073
+ "mrope_section": [
1074
+ 16,
1075
+ 24,
1076
+ 24
1077
+ ],
1078
+ "rope_type": "default",
1079
+ "type": "default"
1080
+ },
1081
+ "layer_types": [
1082
+ "full_attention",
1083
+ "full_attention",
1084
+ "full_attention",
1085
+ "full_attention",
1086
+ "full_attention",
1087
+ "full_attention",
1088
+ "full_attention",
1089
+ "full_attention",
1090
+ "full_attention",
1091
+ "full_attention",
1092
+ "full_attention",
1093
+ "full_attention",
1094
+ "full_attention",
1095
+ "full_attention",
1096
+ "full_attention",
1097
+ "full_attention",
1098
+ "full_attention",
1099
+ "full_attention",
1100
+ "full_attention",
1101
+ "full_attention",
1102
+ "full_attention",
1103
+ "full_attention",
1104
+ "full_attention",
1105
+ "full_attention",
1106
+ "full_attention",
1107
+ "full_attention",
1108
+ "full_attention",
1109
+ "full_attention"
1110
+ ],
1111
+ "image_token_id": null,
1112
+ "video_token_id": null,
1113
  "return_dict": true,
1114
  "output_hidden_states": false,
 
1115
  "torchscript": false,
1116
+ "torch_dtype": "bfloat16",
 
 
1117
  "pruned_heads": {},
1118
+ "tie_word_embeddings": false,
1119
  "chunk_size_feed_forward": 0,
1120
  "is_encoder_decoder": false,
1121
  "is_decoder": false,
1122
  "cross_attention_hidden_size": null,
1123
  "add_cross_attention": false,
1124
  "tie_encoder_decoder": false,
1125
+ "architectures": [
1126
+ "Qwen2_5_VLForConditionalGeneration"
1127
+ ],
1128
+ "finetuning_task": null,
1129
+ "id2label": {
1130
+ "0": "LABEL_0",
1131
+ "1": "LABEL_1"
1132
+ },
1133
+ "label2id": {
1134
+ "LABEL_0": 0,
1135
+ "LABEL_1": 1
1136
+ },
1137
+ "task_specific_params": null,
1138
+ "problem_type": null,
1139
+ "tokenizer_class": null,
1140
+ "prefix": null,
1141
+ "bos_token_id": 151643,
1142
+ "pad_token_id": null,
1143
+ "eos_token_id": 151645,
1144
+ "sep_token_id": null,
1145
+ "decoder_start_token_id": null,
1146
  "max_length": 20,
1147
  "min_length": 0,
1148
  "do_sample": false,
 
1168
  "exponential_decay_length_penalty": null,
1169
  "suppress_tokens": null,
1170
  "begin_suppress_tokens": null,
1171
+ "_name_or_path": "",
1172
+ "model_type": "qwen2_5_vl_text",
1173
+ "vision_end_token_id": 151653,
1174
+ "vision_start_token_id": 151652,
1175
+ "vision_token_id": 151654,
1176
+ "tf_legacy_loss": false,
1177
+ "use_bfloat16": false,
1178
+ "output_attentions": false
1179
+ },
1180
+ "tf_legacy_loss": false,
1181
+ "tie_encoder_decoder": false,
1182
+ "tie_word_embeddings": false,
1183
+ "tokenizer_class": null,
1184
+ "top_k": 50,
1185
+ "top_p": 1.0,
1186
+ "torchscript": false,
1187
+ "transformers_version": "4.55.0.dev0",
1188
+ "typical_p": 1.0,
1189
+ "use_bfloat16": false,
1190
+ "use_cache": false,
1191
+ "use_sliding_window": false,
1192
+ "video_token_id": 151656,
1193
+ "vision_config": {
1194
+ "return_dict": true,
1195
+ "output_hidden_states": false,
1196
+ "torchscript": false,
1197
+ "torch_dtype": "float32",
1198
+ "pruned_heads": {},
1199
+ "tie_word_embeddings": true,
1200
+ "chunk_size_feed_forward": 0,
1201
+ "is_encoder_decoder": false,
1202
+ "is_decoder": false,
1203
+ "cross_attention_hidden_size": null,
1204
+ "add_cross_attention": false,
1205
+ "tie_encoder_decoder": false,
1206
  "architectures": null,
1207
  "finetuning_task": null,
1208
  "id2label": {
 
1213
  "LABEL_0": 0,
1214
  "LABEL_1": 1
1215
  },
1216
+ "task_specific_params": null,
1217
+ "problem_type": null,
1218
  "tokenizer_class": null,
1219
  "prefix": null,
1220
  "bos_token_id": null,
 
1222
  "eos_token_id": null,
1223
  "sep_token_id": null,
1224
  "decoder_start_token_id": null,
1225
+ "max_length": 20,
1226
+ "min_length": 0,
1227
+ "do_sample": false,
1228
+ "early_stopping": false,
1229
+ "num_beams": 1,
1230
+ "num_beam_groups": 1,
1231
+ "diversity_penalty": 0.0,
1232
+ "temperature": 1.0,
1233
+ "top_k": 50,
1234
+ "top_p": 1.0,
1235
+ "typical_p": 1.0,
1236
+ "repetition_penalty": 1.0,
1237
+ "length_penalty": 1.0,
1238
+ "no_repeat_ngram_size": 0,
1239
+ "encoder_no_repeat_ngram_size": 0,
1240
+ "bad_words_ids": null,
1241
+ "num_return_sequences": 1,
1242
+ "output_scores": false,
1243
+ "return_dict_in_generate": false,
1244
+ "forced_bos_token_id": null,
1245
+ "forced_eos_token_id": null,
1246
+ "remove_invalid_values": false,
1247
+ "exponential_decay_length_penalty": null,
1248
+ "suppress_tokens": null,
1249
+ "begin_suppress_tokens": null,
1250
  "_name_or_path": "",
 
1251
  "in_chans": 3,
1252
  "model_type": "qwen2_5_vl",
1253
  "spatial_patch_size": 14,
1254
+ "tf_legacy_loss": false,
1255
+ "use_bfloat16": false,
1256
  "depth": 32,
1257
  "hidden_size": 1280,
1258
  "hidden_act": "silu",
 
1270
  23,
1271
  31
1272
  ],
1273
+ "out_hidden_size": 3584,
1274
+ "initializer_range": 0.02,
1275
+ "output_attentions": false
1276
  },
1277
  "vision_end_token_id": 151653,
1278
  "vision_start_token_id": 151652,
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ced1de9fabf1333aab42f4d6508412f8dbb72b82bb07eb9c8c8360f0617c78a6
3
+ size 5330941254
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cee803fec95b6c57ba997220197b6eb07c325ef8147557bbdf9c5c6cdc8891d8
3
+ size 306561371
preprocessor_config.json CHANGED
@@ -1,4 +1,10 @@
1
  {
 
 
 
 
 
 
2
  "do_convert_rgb": true,
3
  "do_normalize": true,
4
  "do_rescale": true,
@@ -8,12 +14,13 @@
8
  0.4578275,
9
  0.40821073
10
  ],
11
- "image_processor_type": "Qwen2VLImageProcessor",
12
  "image_std": [
13
  0.26862954,
14
  0.26130258,
15
  0.27577711
16
  ],
 
17
  "max_pixels": 12845056,
18
  "merge_size": 2,
19
  "min_pixels": 3136,
@@ -21,6 +28,7 @@
21
  "processor_class": "Qwen2_5_VLProcessor",
22
  "resample": 3,
23
  "rescale_factor": 0.00392156862745098,
 
24
  "size": {
25
  "longest_edge": 12845056,
26
  "shortest_edge": 3136
 
1
  {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": true,
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
  "do_convert_rgb": true,
9
  "do_normalize": true,
10
  "do_rescale": true,
 
14
  0.4578275,
15
  0.40821073
16
  ],
17
+ "image_processor_type": "Qwen2VLImageProcessorFast",
18
  "image_std": [
19
  0.26862954,
20
  0.26130258,
21
  0.27577711
22
  ],
23
+ "input_data_format": null,
24
  "max_pixels": 12845056,
25
  "merge_size": 2,
26
  "min_pixels": 3136,
 
28
  "processor_class": "Qwen2_5_VLProcessor",
29
  "resample": 3,
30
  "rescale_factor": 0.00392156862745098,
31
+ "return_tensors": null,
32
  "size": {
33
  "longest_edge": 12845056,
34
  "shortest_edge": 3136
tokenizer_config.json CHANGED
@@ -195,7 +195,6 @@
195
  "<|video_pad|>"
196
  ],
197
  "bos_token": null,
198
- "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0]['role'] == 'system' %}\n {{- messages[0]['content'] }}\n {%- else %}\n {{- 'You are a helpful assistant.' }}\n {%- endif %}\n {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0]['role'] == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n {%- else %}\n {{- '<|im_start|>system\\nYou are a helpful assistant.<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {{- '<|im_start|>' + message.role }}\n {%- if message.content %}\n {{- '\\n' + message.content }}\n {%- endif %}\n {%- for tool_call in message.tool_calls %}\n {%- if tool_call.function is defined %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '\\n<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {{- tool_call.arguments | tojson }}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
199
  "clean_up_tokenization_spaces": false,
200
  "eos_token": "<|im_end|>",
201
  "errors": "replace",
 
195
  "<|video_pad|>"
196
  ],
197
  "bos_token": null,
 
198
  "clean_up_tokenization_spaces": false,
199
  "eos_token": "<|im_end|>",
200
  "errors": "replace",
video_preprocessor_config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": true,
5
+ "device": null,
6
+ "do_center_crop": null,
7
+ "do_convert_rgb": true,
8
+ "do_normalize": true,
9
+ "do_pad": null,
10
+ "do_rescale": true,
11
+ "do_resize": true,
12
+ "do_sample_frames": false,
13
+ "fps": null,
14
+ "image_mean": [
15
+ 0.48145466,
16
+ 0.4578275,
17
+ 0.40821073
18
+ ],
19
+ "image_std": [
20
+ 0.26862954,
21
+ 0.26130258,
22
+ 0.27577711
23
+ ],
24
+ "input_data_format": null,
25
+ "max_frames": 768,
26
+ "max_pixels": 12845056,
27
+ "merge_size": 2,
28
+ "min_frames": 4,
29
+ "min_pixels": 3136,
30
+ "num_frames": null,
31
+ "patch_size": 14,
32
+ "processor_class": "Qwen2_5_VLProcessor",
33
+ "resample": 3,
34
+ "rescale_factor": 0.00392156862745098,
35
+ "size": {
36
+ "longest_edge": 12845056,
37
+ "shortest_edge": 3136
38
+ },
39
+ "size_divisor": null,
40
+ "temporal_patch_size": 2,
41
+ "video_metadata": null,
42
+ "video_processor_type": "Qwen2VLVideoProcessor"
43
+ }