Sendpns commited on
Commit
d51039e
·
verified ·
1 Parent(s): b973e3b

Upload 3 files

Browse files
Files changed (4) hide show
  1. .gitattributes +1 -0
  2. config.json +646 -0
  3. model.safetensors +3 -0
  4. tokenizer.json +3 -0
.gitattributes CHANGED
@@ -36,3 +36,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
36
  medium_tier.json filter=lfs diff=lfs merge=lfs -text
37
  model.gguf filter=lfs diff=lfs merge=lfs -text
38
  opus46_final.jsonl filter=lfs diff=lfs merge=lfs -text
 
 
36
  medium_tier.json filter=lfs diff=lfs merge=lfs -text
37
  model.gguf filter=lfs diff=lfs merge=lfs -text
38
  opus46_final.jsonl filter=lfs diff=lfs merge=lfs -text
39
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
config.json ADDED
@@ -0,0 +1,646 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Glm4vMoeForConditionalGeneration"
4
+ ],
5
+ "model_type": "glm4v_moe",
6
+ "text_config": {
7
+ "attention_bias": true,
8
+ "attention_dropout": 0.0,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": [
11
+ 151329,
12
+ 151336,
13
+ 151338
14
+ ],
15
+ "first_k_dense_replace": 1,
16
+ "head_dim": 128,
17
+ "hidden_act": "silu",
18
+ "hidden_size": 4096,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 10944,
21
+ "max_position_embeddings": 131072,
22
+ "model_type": "glm4v_moe_text",
23
+ "moe_intermediate_size": 1408,
24
+ "n_group": 1,
25
+ "n_routed_experts": 128,
26
+ "n_shared_experts": 1,
27
+ "norm_topk_prob": true,
28
+ "num_attention_heads": 96,
29
+ "num_experts_per_tok": 8,
30
+ "num_hidden_layers": 46,
31
+ "num_key_value_heads": 8,
32
+ "num_nextn_predict_layers": 0,
33
+ "pad_token_id": 151329,
34
+ "partial_rotary_factor": 0.5,
35
+ "qk_layernorm": false,
36
+ "rms_norm_eps": 1e-05,
37
+ "rope_parameters": {
38
+ "mrope_section": [
39
+ 8,
40
+ 12,
41
+ 12
42
+ ],
43
+ "partial_rotary_factor": 0.5,
44
+ "rope_theta": 500000,
45
+ "rope_type": "default"
46
+ },
47
+ "routed_scaling_factor": 1.0,
48
+ "topk_group": 1,
49
+ "use_cache": true,
50
+ "use_qk_norm": false,
51
+ "vocab_size": 151552
52
+ },
53
+ "tie_word_embeddings": false,
54
+ "transformers_version": "5.0.0rc0",
55
+ "image_start_token_id": 151339,
56
+ "image_end_token_id": 151340,
57
+ "video_start_token_id": 151341,
58
+ "video_end_token_id": 151342,
59
+ "image_token_id": 151363,
60
+ "video_token_id": 151364,
61
+ "vision_config": {
62
+ "attention_bias": false,
63
+ "attention_dropout": 0.0,
64
+ "depth": 24,
65
+ "hidden_act": "silu",
66
+ "hidden_dropout_prob": 0.0,
67
+ "hidden_size": 1536,
68
+ "image_size": 336,
69
+ "in_channels": 3,
70
+ "initializer_range": 0.02,
71
+ "intermediate_size": 10944,
72
+ "model_type": "glm4v_moe_vision",
73
+ "num_heads": 12,
74
+ "out_hidden_size": 4096,
75
+ "patch_size": 14,
76
+ "rms_norm_eps": 1e-05,
77
+ "spatial_merge_size": 2,
78
+ "temporal_patch_size": 2
79
+ },
80
+ "quantization_config": {
81
+ "config_groups": {
82
+ "group_0": {
83
+ "input_activations": {
84
+ "actorder": null,
85
+ "block_structure": null,
86
+ "dynamic": true,
87
+ "group_size": null,
88
+ "num_bits": 8,
89
+ "observer": null,
90
+ "observer_kwargs": {},
91
+ "strategy": "token",
92
+ "symmetric": true,
93
+ "type": "float"
94
+ },
95
+ "output_activations": null,
96
+ "targets": [
97
+ "Linear"
98
+ ],
99
+ "weights": {
100
+ "actorder": null,
101
+ "block_structure": null,
102
+ "dynamic": false,
103
+ "group_size": null,
104
+ "num_bits": 8,
105
+ "observer": "minmax",
106
+ "observer_kwargs": {},
107
+ "strategy": "channel",
108
+ "symmetric": true,
109
+ "type": "float"
110
+ }
111
+ }
112
+ },
113
+ "format": "float-quantized",
114
+ "ignore": [
115
+ "model.layers.4.input_layernorm",
116
+ "model.layers.3.input_layernorm",
117
+ "model.layers.2.input_layernorm",
118
+ "model.layers.0.input_layernorm",
119
+ "model.layers.1.input_layernorm",
120
+ "model.layers.8.input_layernorm",
121
+ "model.layers.6.input_layernorm",
122
+ "model.layers.7.input_layernorm",
123
+ "model.layers.5.input_layernorm",
124
+ "model.layers.9.input_layernorm",
125
+ "visual.blocks.6.norm1",
126
+ "visual.blocks.7.norm1",
127
+ "visual.blocks.8.norm1",
128
+ "visual.blocks.9.norm1",
129
+ "visual.patch_embed.proj.bias",
130
+ "visual.downsample.bias",
131
+ "model.layers.38.mlp.gate",
132
+ "model.layers.39.mlp.gate",
133
+ "model.layers.40.mlp.gate",
134
+ "model.layers.41.mlp.gate",
135
+ "model.layers.42.mlp.gate",
136
+ "model.layers.28.mlp.gate",
137
+ "model.layers.29.mlp.gate",
138
+ "model.layers.30.mlp.gate",
139
+ "model.layers.31.mlp.gate",
140
+ "model.layers.32.mlp.gate",
141
+ "model.layers.31.self_attn.q_proj.bias",
142
+ "model.layers.32.self_attn.q_proj.bias",
143
+ "model.layers.33.self_attn.q_proj.bias",
144
+ "model.layers.43.self_attn.q_proj.bias",
145
+ "model.layers.44.self_attn.q_proj.bias",
146
+ "model.layers.45.self_attn.q_proj.bias",
147
+ "model.layers.20.self_attn.q_proj.bias",
148
+ "model.layers.22.self_attn.q_proj.bias",
149
+ "model.layers.21.self_attn.q_proj.bias",
150
+ "model.layers.23.self_attn.q_proj.bias",
151
+ "model.layers.10.self_attn.q_proj.bias",
152
+ "model.layers.11.self_attn.q_proj.bias",
153
+ "model.layers.12.self_attn.q_proj.bias",
154
+ "model.layers.14.self_attn.q_proj.bias",
155
+ "model.layers.38.self_attn.q_proj.bias",
156
+ "model.layers.39.self_attn.q_proj.bias",
157
+ "model.layers.37.post_attention_layernorm",
158
+ "model.layers.38.post_attention_layernorm",
159
+ "model.layers.39.post_attention_layernorm",
160
+ "model.layers.20.post_attention_layernorm",
161
+ "model.layers.22.post_attention_layernorm",
162
+ "model.layers.21.post_attention_layernorm",
163
+ "model.layers.23.post_attention_layernorm",
164
+ "model.layers.24.post_attention_layernorm",
165
+ "model.layers.25.post_attention_layernorm",
166
+ "model.layers.26.post_attention_layernorm",
167
+ "model.layers.43.self_attn.k_proj.bias",
168
+ "model.layers.44.self_attn.k_proj.bias",
169
+ "model.layers.45.self_attn.k_proj.bias",
170
+ "model.layers.5.self_attn.v_proj.bias",
171
+ "model.layers.9.self_attn.v_proj.bias",
172
+ "model.layers.33.mlp.gate",
173
+ "model.layers.34.mlp.gate",
174
+ "model.layers.35.mlp.gate",
175
+ "model.layers.36.mlp.gate",
176
+ "model.layers.37.mlp.gate",
177
+ "model.layers.10.mlp.gate",
178
+ "model.layers.11.mlp.gate",
179
+ "model.layers.12.mlp.gate",
180
+ "model.layers.14.mlp.gate",
181
+ "model.layers.13.mlp.gate",
182
+ "model.layers.6.mlp.gate",
183
+ "model.layers.7.mlp.gate",
184
+ "model.layers.5.mlp.gate",
185
+ "model.layers.3.mlp.gate",
186
+ "model.layers.2.mlp.gate",
187
+ "model.layers.4.mlp.gate",
188
+ "model.layers.1.mlp.gate",
189
+ "model.layers.8.mlp.gate",
190
+ "model.layers.9.mlp.gate",
191
+ "visual.blocks.15.mlp.gate_up_proj",
192
+ "visual.blocks.15.mlp.gate_proj",
193
+ "visual.blocks.15.mlp.up_proj",
194
+ "visual.blocks.16.mlp.down_proj",
195
+ "visual.blocks.16.mlp.gate_up_proj",
196
+ "visual.blocks.16.mlp.gate_proj",
197
+ "visual.blocks.16.mlp.up_proj",
198
+ "visual.blocks.17.mlp.down_proj",
199
+ "visual.blocks.17.mlp.gate_up_proj",
200
+ "visual.blocks.17.mlp.gate_proj",
201
+ "visual.blocks.17.mlp.up_proj",
202
+ "visual.blocks.18.mlp.down_proj",
203
+ "visual.blocks.18.mlp.gate_up_proj",
204
+ "visual.blocks.18.mlp.gate_proj",
205
+ "visual.blocks.18.mlp.up_proj",
206
+ "visual.blocks.19.mlp.down_proj",
207
+ "visual.blocks.19.mlp.gate_up_proj",
208
+ "visual.blocks.19.mlp.gate_proj",
209
+ "visual.blocks.19.mlp.up_proj",
210
+ "visual.blocks.20.mlp.down_proj",
211
+ "visual.post_layernorm",
212
+ "model.layers.40.mlp.gate.e_score_correction_bias",
213
+ "model.layers.41.mlp.gate.e_score_correction_bias",
214
+ "model.layers.42.mlp.gate.e_score_correction_bias",
215
+ "model.layers.29.mlp.gate.e_score_correction_bias",
216
+ "model.layers.30.mlp.gate.e_score_correction_bias",
217
+ "model.layers.31.mlp.gate.e_score_correction_bias",
218
+ "model.layers.32.mlp.gate.e_score_correction_bias",
219
+ "model.layers.33.mlp.gate.e_score_correction_bias",
220
+ "model.layers.34.mlp.gate.e_score_correction_bias",
221
+ "model.layers.35.mlp.gate.e_score_correction_bias",
222
+ "visual.blocks.12.norm1",
223
+ "visual.blocks.12.norm2",
224
+ "visual.blocks.13.norm1",
225
+ "visual.blocks.13.norm2",
226
+ "visual.blocks.14.norm1",
227
+ "visual.blocks.14.norm2",
228
+ "visual.blocks.15.norm1",
229
+ "visual.blocks.15.norm2",
230
+ "visual.blocks.16.norm2",
231
+ "visual.blocks.17.norm2",
232
+ "model.layers.45.mlp.gate.e_score_correction_bias",
233
+ "model.layers.17.post_attention_layernorm",
234
+ "model.layers.18.post_attention_layernorm",
235
+ "model.layers.19.post_attention_layernorm",
236
+ "model.layers.43.post_attention_layernorm",
237
+ "model.layers.44.post_attention_layernorm",
238
+ "model.layers.45.post_attention_layernorm",
239
+ "model.layers.20.self_attn.v_proj.bias",
240
+ "model.layers.22.self_attn.v_proj.bias",
241
+ "model.layers.21.self_attn.v_proj.bias",
242
+ "model.layers.23.self_attn.v_proj.bias",
243
+ "model.layers.24.self_attn.v_proj.bias",
244
+ "model.layers.25.self_attn.v_proj.bias",
245
+ "model.layers.26.self_attn.v_proj.bias",
246
+ "model.layers.27.self_attn.v_proj.bias",
247
+ "model.layers.28.self_attn.v_proj.bias",
248
+ "model.layers.29.self_attn.v_proj.bias",
249
+ "visual.blocks.11.mlp.gate_up_proj",
250
+ "visual.blocks.11.mlp.gate_proj",
251
+ "visual.blocks.11.mlp.up_proj",
252
+ "visual.blocks.5.mlp.gate_up_proj",
253
+ "visual.blocks.5.mlp.gate_proj",
254
+ "visual.blocks.5.mlp.up_proj",
255
+ "visual.blocks.5.norm1",
256
+ "visual.blocks.6.mlp.gate_up_proj",
257
+ "visual.blocks.6.mlp.gate_proj",
258
+ "visual.blocks.6.mlp.up_proj",
259
+ "visual.blocks.7.mlp.gate_up_proj",
260
+ "visual.blocks.7.mlp.gate_proj",
261
+ "visual.blocks.7.mlp.up_proj",
262
+ "visual.blocks.8.mlp.gate_up_proj",
263
+ "visual.blocks.8.mlp.gate_proj",
264
+ "visual.blocks.8.mlp.up_proj",
265
+ "visual.blocks.9.mlp.gate_up_proj",
266
+ "visual.blocks.9.mlp.gate_proj",
267
+ "visual.blocks.9.mlp.up_proj",
268
+ "model.layers.22.mlp.gate.e_score_correction_bias",
269
+ "model.layers.10.mlp.gate.e_score_correction_bias",
270
+ "model.layers.11.mlp.gate.e_score_correction_bias",
271
+ "model.layers.12.mlp.gate.e_score_correction_bias",
272
+ "visual.blocks.12.mlp.down_proj",
273
+ "visual.blocks.12.mlp.gate_up_proj",
274
+ "visual.blocks.12.mlp.gate_proj",
275
+ "visual.blocks.12.mlp.up_proj",
276
+ "model.layers.14.mlp.gate.e_score_correction_bias",
277
+ "model.layers.13.mlp.gate.e_score_correction_bias",
278
+ "model.layers.15.mlp.gate.e_score_correction_bias",
279
+ "visual.blocks.13.mlp.down_proj",
280
+ "visual.blocks.13.mlp.gate_up_proj",
281
+ "visual.blocks.13.mlp.gate_proj",
282
+ "visual.blocks.13.mlp.up_proj",
283
+ "visual.blocks.14.mlp.down_proj",
284
+ "visual.blocks.14.mlp.gate_up_proj",
285
+ "visual.blocks.14.mlp.gate_proj",
286
+ "visual.blocks.14.mlp.up_proj",
287
+ "visual.blocks.15.mlp.down_proj",
288
+ "model.embed_tokens",
289
+ "model.layers.10.input_layernorm",
290
+ "model.layers.11.input_layernorm",
291
+ "model.layers.12.input_layernorm",
292
+ "model.layers.14.input_layernorm",
293
+ "model.layers.39.input_layernorm",
294
+ "model.layers.13.input_layernorm",
295
+ "model.layers.15.input_layernorm",
296
+ "model.layers.16.input_layernorm",
297
+ "model.layers.17.input_layernorm",
298
+ "model.layers.18.input_layernorm",
299
+ "model.layers.27.post_attention_layernorm",
300
+ "model.layers.28.post_attention_layernorm",
301
+ "model.layers.29.post_attention_layernorm",
302
+ "model.layers.10.post_attention_layernorm",
303
+ "model.layers.11.post_attention_layernorm",
304
+ "model.layers.12.post_attention_layernorm",
305
+ "model.layers.14.post_attention_layernorm",
306
+ "model.layers.13.post_attention_layernorm",
307
+ "model.layers.15.post_attention_layernorm",
308
+ "model.layers.16.post_attention_layernorm",
309
+ "visual.blocks.19.norm2",
310
+ "visual.blocks.20.norm2",
311
+ "visual.blocks.21.norm2",
312
+ "visual.blocks.22.norm2",
313
+ "visual.blocks.23.norm2",
314
+ "visual.blocks.10.norm2",
315
+ "visual.blocks.11.norm2",
316
+ "model.layers.2.mlp.gate.e_score_correction_bias",
317
+ "model.layers.4.mlp.gate.e_score_correction_bias",
318
+ "model.layers.3.mlp.gate.e_score_correction_bias",
319
+ "model.layers.1.mlp.gate.e_score_correction_bias",
320
+ "model.layers.8.mlp.gate.e_score_correction_bias",
321
+ "model.layers.6.mlp.gate.e_score_correction_bias",
322
+ "model.layers.7.mlp.gate.e_score_correction_bias",
323
+ "model.layers.5.mlp.gate.e_score_correction_bias",
324
+ "model.layers.9.mlp.gate.e_score_correction_bias",
325
+ "model.layers.2.self_attn.k_proj.bias",
326
+ "model.layers.4.self_attn.k_proj.bias",
327
+ "model.layers.3.self_attn.k_proj.bias",
328
+ "model.layers.0.self_attn.k_proj.bias",
329
+ "model.layers.1.self_attn.k_proj.bias",
330
+ "model.layers.8.self_attn.k_proj.bias",
331
+ "model.layers.6.self_attn.k_proj.bias",
332
+ "model.layers.7.self_attn.k_proj.bias",
333
+ "model.layers.5.self_attn.k_proj.bias",
334
+ "model.layers.9.self_attn.k_proj.bias",
335
+ "model.layers.34.self_attn.q_proj.bias",
336
+ "model.layers.35.self_attn.q_proj.bias",
337
+ "model.layers.36.self_attn.q_proj.bias",
338
+ "model.layers.37.self_attn.q_proj.bias",
339
+ "model.layers.13.self_attn.q_proj.bias",
340
+ "model.layers.15.self_attn.q_proj.bias",
341
+ "model.layers.16.self_attn.q_proj.bias",
342
+ "model.layers.17.self_attn.q_proj.bias",
343
+ "model.layers.18.self_attn.q_proj.bias",
344
+ "model.layers.19.self_attn.q_proj.bias",
345
+ "model.layers.40.self_attn.q_proj.bias",
346
+ "model.layers.41.self_attn.q_proj.bias",
347
+ "model.layers.42.self_attn.q_proj.bias",
348
+ "model.layers.24.self_attn.q_proj.bias",
349
+ "model.layers.25.self_attn.q_proj.bias",
350
+ "model.layers.26.self_attn.q_proj.bias",
351
+ "model.layers.27.self_attn.q_proj.bias",
352
+ "model.layers.28.self_attn.q_proj.bias",
353
+ "model.layers.29.self_attn.q_proj.bias",
354
+ "model.layers.30.self_attn.q_proj.bias",
355
+ "model.layers.25.mlp.gate.e_score_correction_bias",
356
+ "model.layers.26.mlp.gate.e_score_correction_bias",
357
+ "model.layers.27.mlp.gate.e_score_correction_bias",
358
+ "model.layers.28.mlp.gate.e_score_correction_bias",
359
+ "model.layers.39.mlp.gate.e_score_correction_bias",
360
+ "model.layers.16.mlp.gate.e_score_correction_bias",
361
+ "model.layers.17.mlp.gate.e_score_correction_bias",
362
+ "model.layers.18.mlp.gate.e_score_correction_bias",
363
+ "model.layers.19.mlp.gate.e_score_correction_bias",
364
+ "model.layers.20.mlp.gate.e_score_correction_bias",
365
+ "model.layers.36.mlp.gate.e_score_correction_bias",
366
+ "model.layers.37.mlp.gate.e_score_correction_bias",
367
+ "model.layers.38.mlp.gate.e_score_correction_bias",
368
+ "model.layers.21.mlp.gate.e_score_correction_bias",
369
+ "model.layers.23.mlp.gate.e_score_correction_bias",
370
+ "model.layers.24.mlp.gate.e_score_correction_bias",
371
+ "model.layers.22.self_attn.k_proj.bias",
372
+ "model.layers.21.self_attn.k_proj.bias",
373
+ "model.layers.23.self_attn.k_proj.bias",
374
+ "model.layers.24.self_attn.k_proj.bias",
375
+ "model.layers.25.self_attn.k_proj.bias",
376
+ "model.layers.26.self_attn.k_proj.bias",
377
+ "model.layers.27.self_attn.k_proj.bias",
378
+ "model.layers.28.self_attn.k_proj.bias",
379
+ "model.layers.38.self_attn.k_proj.bias",
380
+ "model.layers.39.self_attn.k_proj.bias",
381
+ "visual.blocks.0.norm1",
382
+ "visual.blocks.1.norm1",
383
+ "visual.blocks.2.norm1",
384
+ "visual.blocks.0.norm2",
385
+ "visual.blocks.1.norm2",
386
+ "visual.blocks.2.norm2",
387
+ "visual.blocks.3.norm1",
388
+ "visual.blocks.3.norm2",
389
+ "visual.blocks.4.norm1",
390
+ "visual.blocks.19.attn.qkv_proj",
391
+ "visual.blocks.20.attn.qkv_proj",
392
+ "visual.blocks.21.attn.qkv_proj",
393
+ "visual.blocks.22.attn.qkv_proj",
394
+ "visual.blocks.23.attn.qkv_proj",
395
+ "visual.blocks.10.attn.qkv_proj",
396
+ "visual.blocks.11.attn.qkv_proj",
397
+ "visual.blocks.4.norm2",
398
+ "visual.blocks.5.norm2",
399
+ "visual.blocks.6.norm2",
400
+ "visual.blocks.7.norm2",
401
+ "visual.blocks.8.norm2",
402
+ "visual.blocks.9.norm2",
403
+ "visual.blocks.20.mlp.gate_up_proj",
404
+ "visual.blocks.20.mlp.gate_proj",
405
+ "visual.blocks.20.mlp.up_proj",
406
+ "visual.blocks.21.mlp.down_proj",
407
+ "visual.blocks.21.mlp.gate_up_proj",
408
+ "visual.blocks.21.mlp.gate_proj",
409
+ "visual.blocks.21.mlp.up_proj",
410
+ "visual.blocks.22.mlp.down_proj",
411
+ "visual.blocks.22.mlp.gate_up_proj",
412
+ "visual.blocks.22.mlp.gate_proj",
413
+ "visual.blocks.22.mlp.up_proj",
414
+ "visual.blocks.23.mlp.down_proj",
415
+ "visual.blocks.23.mlp.gate_up_proj",
416
+ "visual.blocks.23.mlp.gate_proj",
417
+ "visual.blocks.23.mlp.up_proj",
418
+ "visual.blocks.10.mlp.down_proj",
419
+ "visual.blocks.10.mlp.gate_up_proj",
420
+ "visual.blocks.10.mlp.gate_proj",
421
+ "visual.blocks.10.mlp.up_proj",
422
+ "visual.blocks.11.mlp.down_proj",
423
+ "visual.embeddings.position_embedding",
424
+ "model.layers.15.mlp.gate",
425
+ "model.layers.16.mlp.gate",
426
+ "model.layers.17.mlp.gate",
427
+ "visual.merger.proj",
428
+ "model.layers.43.mlp.gate",
429
+ "model.layers.44.mlp.gate",
430
+ "model.layers.45.mlp.gate",
431
+ "model.layers.32.input_layernorm",
432
+ "model.layers.33.input_layernorm",
433
+ "model.layers.34.input_layernorm",
434
+ "model.layers.35.input_layernorm",
435
+ "model.layers.36.input_layernorm",
436
+ "model.layers.37.input_layernorm",
437
+ "model.layers.38.input_layernorm",
438
+ "model.layers.40.input_layernorm",
439
+ "model.layers.41.input_layernorm",
440
+ "model.layers.42.input_layernorm",
441
+ "model.layers.10.self_attn.k_proj.bias",
442
+ "model.layers.11.self_attn.k_proj.bias",
443
+ "model.layers.12.self_attn.k_proj.bias",
444
+ "model.layers.14.self_attn.k_proj.bias",
445
+ "model.layers.13.self_attn.k_proj.bias",
446
+ "model.layers.15.self_attn.k_proj.bias",
447
+ "model.layers.16.self_attn.k_proj.bias",
448
+ "model.layers.17.self_attn.k_proj.bias",
449
+ "model.layers.18.self_attn.k_proj.bias",
450
+ "model.layers.19.self_attn.k_proj.bias",
451
+ "visual.patch_embed.proj",
452
+ "model.layers.32.self_attn.k_proj.bias",
453
+ "model.layers.33.self_attn.k_proj.bias",
454
+ "model.layers.34.self_attn.k_proj.bias",
455
+ "model.layers.35.self_attn.k_proj.bias",
456
+ "model.layers.36.self_attn.k_proj.bias",
457
+ "model.layers.37.self_attn.k_proj.bias",
458
+ "model.layers.40.self_attn.k_proj.bias",
459
+ "model.layers.41.self_attn.k_proj.bias",
460
+ "model.layers.42.self_attn.k_proj.bias",
461
+ "model.layers.29.self_attn.k_proj.bias",
462
+ "lm_head",
463
+ "model.layers.40.post_attention_layernorm",
464
+ "model.layers.41.post_attention_layernorm",
465
+ "model.layers.42.post_attention_layernorm",
466
+ "model.layers.30.post_attention_layernorm",
467
+ "model.layers.31.post_attention_layernorm",
468
+ "model.layers.32.post_attention_layernorm",
469
+ "model.layers.33.post_attention_layernorm",
470
+ "model.layers.34.post_attention_layernorm",
471
+ "model.layers.35.post_attention_layernorm",
472
+ "model.layers.36.post_attention_layernorm",
473
+ "model.layers.30.self_attn.k_proj.bias",
474
+ "model.layers.31.self_attn.k_proj.bias",
475
+ "model.layers.20.self_attn.k_proj.bias",
476
+ "model.layers.21.input_layernorm",
477
+ "model.layers.23.input_layernorm",
478
+ "model.layers.24.input_layernorm",
479
+ "model.layers.25.input_layernorm",
480
+ "model.layers.26.input_layernorm",
481
+ "model.layers.27.input_layernorm",
482
+ "model.layers.28.input_layernorm",
483
+ "model.layers.29.input_layernorm",
484
+ "model.layers.30.input_layernorm",
485
+ "model.layers.31.input_layernorm",
486
+ "visual.blocks.22.attn.proj",
487
+ "visual.blocks.23.attn.proj",
488
+ "visual.blocks.10.attn.proj",
489
+ "visual.blocks.11.attn.proj",
490
+ "visual.merger.down_proj",
491
+ "visual.merger.gate_up_proj",
492
+ "visual.merger.post_projection_norm.bias",
493
+ "model.norm",
494
+ "visual.blocks.16.norm1",
495
+ "visual.blocks.17.norm1",
496
+ "visual.blocks.18.norm1",
497
+ "visual.blocks.19.norm1",
498
+ "visual.blocks.20.norm1",
499
+ "visual.blocks.21.norm1",
500
+ "visual.blocks.22.norm1",
501
+ "visual.blocks.23.norm1",
502
+ "visual.blocks.10.norm1",
503
+ "visual.blocks.11.norm1",
504
+ "visual.blocks.18.norm2",
505
+ "model.layers.10.self_attn.v_proj.bias",
506
+ "model.layers.11.self_attn.v_proj.bias",
507
+ "model.layers.12.self_attn.v_proj.bias",
508
+ "model.layers.14.self_attn.v_proj.bias",
509
+ "model.layers.13.self_attn.v_proj.bias",
510
+ "model.layers.15.self_attn.v_proj.bias",
511
+ "model.layers.16.self_attn.v_proj.bias",
512
+ "model.layers.17.self_attn.v_proj.bias",
513
+ "model.layers.18.self_attn.v_proj.bias",
514
+ "model.layers.19.self_attn.v_proj.bias",
515
+ "visual.blocks.0.attn.qkv_proj",
516
+ "visual.blocks.1.attn.qkv_proj",
517
+ "visual.blocks.2.attn.qkv_proj",
518
+ "visual.blocks.3.attn.qkv_proj",
519
+ "visual.blocks.4.attn.qkv_proj",
520
+ "visual.blocks.5.attn.qkv_proj",
521
+ "visual.blocks.6.attn.qkv_proj",
522
+ "visual.blocks.7.attn.qkv_proj",
523
+ "visual.blocks.8.attn.qkv_proj",
524
+ "visual.blocks.9.attn.qkv_proj",
525
+ "model.layers.31.self_attn.v_proj.bias",
526
+ "model.layers.32.self_attn.v_proj.bias",
527
+ "model.layers.33.self_attn.v_proj.bias",
528
+ "model.layers.43.self_attn.v_proj.bias",
529
+ "model.layers.44.self_attn.v_proj.bias",
530
+ "model.layers.45.self_attn.v_proj.bias",
531
+ "visual.post_conv_layernorm",
532
+ "visual.blocks.0.mlp.down_proj",
533
+ "visual.blocks.0.mlp.gate_up_proj",
534
+ "visual.blocks.0.mlp.gate_proj",
535
+ "visual.blocks.0.mlp.up_proj",
536
+ "visual.blocks.1.mlp.down_proj",
537
+ "visual.blocks.1.mlp.gate_up_proj",
538
+ "visual.blocks.1.mlp.gate_proj",
539
+ "visual.blocks.1.mlp.up_proj",
540
+ "visual.blocks.2.mlp.down_proj",
541
+ "visual.blocks.2.mlp.gate_up_proj",
542
+ "visual.blocks.2.mlp.gate_proj",
543
+ "visual.blocks.2.mlp.up_proj",
544
+ "visual.blocks.3.mlp.down_proj",
545
+ "visual.blocks.3.mlp.gate_up_proj",
546
+ "visual.blocks.3.mlp.gate_proj",
547
+ "visual.blocks.3.mlp.up_proj",
548
+ "visual.blocks.4.mlp.down_proj",
549
+ "visual.blocks.4.mlp.gate_up_proj",
550
+ "visual.blocks.4.mlp.gate_proj",
551
+ "visual.blocks.4.mlp.up_proj",
552
+ "visual.merger.post_projection_norm",
553
+ "visual.blocks.12.attn.proj",
554
+ "visual.blocks.13.attn.proj",
555
+ "visual.blocks.14.attn.proj",
556
+ "visual.blocks.15.attn.proj",
557
+ "visual.blocks.16.attn.proj",
558
+ "visual.blocks.17.attn.proj",
559
+ "visual.blocks.18.attn.proj",
560
+ "visual.blocks.19.attn.proj",
561
+ "visual.blocks.20.attn.proj",
562
+ "visual.blocks.21.attn.proj",
563
+ "visual.blocks.5.mlp.down_proj",
564
+ "visual.blocks.6.mlp.down_proj",
565
+ "visual.blocks.7.mlp.down_proj",
566
+ "visual.blocks.8.mlp.down_proj",
567
+ "visual.blocks.9.mlp.down_proj",
568
+ "visual.blocks.0.attn.proj",
569
+ "visual.blocks.1.attn.proj",
570
+ "visual.blocks.2.attn.proj",
571
+ "visual.blocks.3.attn.proj",
572
+ "visual.blocks.4.attn.proj",
573
+ "visual.blocks.5.attn.proj",
574
+ "visual.blocks.6.attn.proj",
575
+ "visual.blocks.7.attn.proj",
576
+ "visual.blocks.8.attn.proj",
577
+ "visual.blocks.9.attn.proj",
578
+ "visual.blocks.12.attn.qkv_proj",
579
+ "visual.blocks.13.attn.qkv_proj",
580
+ "visual.blocks.14.attn.qkv_proj",
581
+ "visual.blocks.15.attn.qkv_proj",
582
+ "visual.blocks.16.attn.qkv_proj",
583
+ "visual.blocks.17.attn.qkv_proj",
584
+ "visual.blocks.18.attn.qkv_proj",
585
+ "model.layers.34.self_attn.v_proj.bias",
586
+ "model.layers.35.self_attn.v_proj.bias",
587
+ "model.layers.36.self_attn.v_proj.bias",
588
+ "model.layers.37.self_attn.v_proj.bias",
589
+ "model.layers.38.self_attn.v_proj.bias",
590
+ "model.layers.39.self_attn.v_proj.bias",
591
+ "model.layers.40.self_attn.v_proj.bias",
592
+ "model.layers.41.self_attn.v_proj.bias",
593
+ "model.layers.42.self_attn.v_proj.bias",
594
+ "model.layers.30.self_attn.v_proj.bias",
595
+ "model.layers.2.self_attn.q_proj.bias",
596
+ "model.layers.4.self_attn.q_proj.bias",
597
+ "model.layers.3.self_attn.q_proj.bias",
598
+ "model.layers.0.self_attn.q_proj.bias",
599
+ "model.layers.1.self_attn.q_proj.bias",
600
+ "model.layers.8.self_attn.q_proj.bias",
601
+ "model.layers.6.self_attn.q_proj.bias",
602
+ "model.layers.7.self_attn.q_proj.bias",
603
+ "model.layers.5.self_attn.q_proj.bias",
604
+ "model.layers.9.self_attn.q_proj.bias",
605
+ "model.layers.2.self_attn.v_proj.bias",
606
+ "model.layers.4.self_attn.v_proj.bias",
607
+ "model.layers.3.self_attn.v_proj.bias",
608
+ "model.layers.0.self_attn.v_proj.bias",
609
+ "model.layers.1.self_attn.v_proj.bias",
610
+ "model.layers.8.self_attn.v_proj.bias",
611
+ "model.layers.6.self_attn.v_proj.bias",
612
+ "model.layers.7.self_attn.v_proj.bias",
613
+ "model.layers.18.mlp.gate",
614
+ "model.layers.19.mlp.gate",
615
+ "model.layers.20.mlp.gate",
616
+ "model.layers.22.mlp.gate",
617
+ "model.layers.21.mlp.gate",
618
+ "model.layers.23.mlp.gate",
619
+ "model.layers.24.mlp.gate",
620
+ "model.layers.25.mlp.gate",
621
+ "model.layers.26.mlp.gate",
622
+ "model.layers.27.mlp.gate",
623
+ "model.layers.43.mlp.gate.e_score_correction_bias",
624
+ "model.layers.44.mlp.gate.e_score_correction_bias",
625
+ "model.layers.19.input_layernorm",
626
+ "model.layers.20.input_layernorm",
627
+ "model.layers.22.input_layernorm",
628
+ "model.layers.43.input_layernorm",
629
+ "model.layers.44.input_layernorm",
630
+ "model.layers.45.input_layernorm",
631
+ "model.layers.2.post_attention_layernorm",
632
+ "model.layers.4.post_attention_layernorm",
633
+ "model.layers.3.post_attention_layernorm",
634
+ "model.layers.0.post_attention_layernorm",
635
+ "model.layers.1.post_attention_layernorm",
636
+ "model.layers.8.post_attention_layernorm",
637
+ "model.layers.6.post_attention_layernorm",
638
+ "model.layers.7.post_attention_layernorm",
639
+ "model.layers.5.post_attention_layernorm",
640
+ "model.layers.9.post_attention_layernorm",
641
+ "visual.downsample"
642
+ ],
643
+ "quant_method": "compressed-tensors",
644
+ "quantization_status": "compressed"
645
+ }
646
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8986666efa26c951c93df9c917a703cb1d00e1d857c1fe8430bbd84a00680bb4
3
+ size 2714933720
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9340665016419c825c4bdabbcc9acc43b7ca2c68ce142724afa829abb1be5efd
3
+ size 19970699