Hugging Face's logo

nie10
/
multi_Stage1

PEFT
bunny-phi
custom_code
Model card Files
xet
Community
nie10 commited on
Commit
85efc86
·
verified ·
1 Parent(s): 488f9b4

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. README.md +9 -0
  2. adapter_config.json +367 -0
  3. adapter_model.bin +3 -0
  4. config.json +47 -0
  5. non_lora_trainables.bin +3 -0
README.md ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+ ### Framework versions
7
+
8
+
9
+ - PEFT 0.4.0
adapter_config.json ADDED
@@ -0,0 +1,367 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "/mnt/workspace/workgroup/ustc/Bunny-v1/",
4
+ "bias": "none",
5
+ "fan_in_fan_out": false,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 256,
11
+ "lora_dropout": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 128,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.k_proj",
18
+ "model.layers.4.self_attn.dense",
19
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.21.mlp.fc1",
20
+ "model.layers.22.self_attn.dense",
21
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.v_proj",
22
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.17.mlp.fc1",
23
+ "model.layers.20.mlp.fc1",
24
+ "model.layers.27.self_attn.k_proj",
25
+ "model.layers.2.mlp.fc2",
26
+ "model.layers.17.self_attn.q_proj",
27
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.7.mlp.fc1",
28
+ "model.layers.31.self_attn.q_proj",
29
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.q_proj",
30
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.q_proj",
31
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.q_proj",
32
+ "model.layers.23.mlp.fc2",
33
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.2.mlp.fc1",
34
+ "model.layers.14.self_attn.dense",
35
+ "model.layers.31.mlp.fc2",
36
+ "model.layers.1.self_attn.k_proj",
37
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.v_proj",
38
+ "model.layers.21.self_attn.v_proj",
39
+ "model.layers.26.mlp.fc1",
40
+ "model.layers.0.self_attn.dense",
41
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.q_proj",
42
+ "model.layers.15.self_attn.v_proj",
43
+ "model.layers.1.self_attn.v_proj",
44
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.v_proj",
45
+ "model.layers.11.mlp.fc2",
46
+ "model.layers.9.mlp.fc2",
47
+ "model.layers.3.self_attn.k_proj",
48
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.out_proj",
49
+ "model.layers.3.mlp.fc2",
50
+ "model.layers.19.self_attn.k_proj",
51
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.18.mlp.fc1",
52
+ "model.layers.28.self_attn.v_proj",
53
+ "model.layers.11.self_attn.k_proj",
54
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.v_proj",
55
+ "model.layers.6.self_attn.q_proj",
56
+ "model.layers.8.mlp.fc1",
57
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.k_proj",
58
+ "model.layers.15.mlp.fc1",
59
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.16.mlp.fc2",
60
+ "model.layers.20.self_attn.q_proj",
61
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.0.mlp.fc1",
62
+ "model.layers.21.self_attn.q_proj",
63
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.v_proj",
64
+ "model.layers.28.mlp.fc1",
65
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.q_proj",
66
+ "model.layers.17.self_attn.dense",
67
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.v_proj",
68
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.20.mlp.fc2",
69
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.q_proj",
70
+ "model.layers.7.self_attn.q_proj",
71
+ "model.layers.30.self_attn.q_proj",
72
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj",
73
+ "model.layers.0.self_attn.q_proj",
74
+ "model.layers.24.self_attn.dense",
75
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.2.mlp.fc2",
76
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.q_proj",
77
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.k_proj",
78
+ "model.layers.20.mlp.fc2",
79
+ "model.layers.8.mlp.fc2",
80
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.13.mlp.fc2",
81
+ "model.layers.6.mlp.fc2",
82
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.q_proj",
83
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.k_proj",
84
+ "model.layers.11.self_attn.q_proj",
85
+ "model.layers.31.mlp.fc1",
86
+ "model.layers.1.mlp.fc1",
87
+ "model.layers.28.self_attn.k_proj",
88
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.q_proj",
89
+ "model.layers.12.self_attn.dense",
90
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.q_proj",
91
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.v_proj",
92
+ "model.layers.13.self_attn.v_proj",
93
+ "model.layers.24.mlp.fc2",
94
+ "model.layers.18.self_attn.v_proj",
95
+ "model.layers.3.mlp.fc1",
96
+ "model.layers.29.self_attn.q_proj",
97
+ "model.layers.23.self_attn.q_proj",
98
+ "model.layers.31.self_attn.k_proj",
99
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.k_proj",
100
+ "model.layers.14.mlp.fc2",
101
+ "model.layers.26.self_attn.dense",
102
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.out_proj",
103
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.21.mlp.fc2",
104
+ "model.layers.12.self_attn.v_proj",
105
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.10.mlp.fc1",
106
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.k_proj",
107
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.v_proj",
108
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.out_proj",
109
+ "model.layers.10.mlp.fc1",
110
+ "model.layers.12.self_attn.q_proj",
111
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.v_proj",
112
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.q_proj",
113
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.k_proj",
114
+ "model.layers.23.self_attn.k_proj",
115
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.0.self_attn.out_proj",
116
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.3.mlp.fc1",
117
+ "model.layers.27.self_attn.q_proj",
118
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.out_proj",
119
+ "model.layers.18.self_attn.dense",
120
+ "model.layers.28.mlp.fc2",
121
+ "model.layers.19.self_attn.q_proj",
122
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.1.mlp.fc1",
123
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.out_proj",
124
+ "model.layers.5.self_attn.q_proj",
125
+ "model.layers.3.self_attn.dense",
126
+ "model.layers.24.mlp.fc1",
127
+ "model.layers.6.self_attn.k_proj",
128
+ "model.layers.27.self_attn.v_proj",
129
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.9.mlp.fc1",
130
+ "model.layers.25.self_attn.q_proj",
131
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.v_proj",
132
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.14.mlp.fc1",
133
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.v_proj",
134
+ "model.layers.10.self_attn.k_proj",
135
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.v_proj",
136
+ "model.layers.4.self_attn.k_proj",
137
+ "model.layers.13.mlp.fc1",
138
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.k_proj",
139
+ "model.layers.4.self_attn.q_proj",
140
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.v_proj",
141
+ "model.layers.16.mlp.fc1",
142
+ "model.layers.30.self_attn.k_proj",
143
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.v_proj",
144
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.23.mlp.fc1",
145
+ "model.layers.22.mlp.fc2",
146
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.4.mlp.fc2",
147
+ "model.layers.21.mlp.fc1",
148
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.22.mlp.fc2",
149
+ "model.layers.15.self_attn.k_proj",
150
+ "model.layers.8.self_attn.q_proj",
151
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.10.mlp.fc2",
152
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.v_proj",
153
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.16.mlp.fc1",
154
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.v_proj",
155
+ "model.layers.14.self_attn.q_proj",
156
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.v_proj",
157
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.k_proj",
158
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.q_proj",
159
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.24.mlp.fc1",
160
+ "model.layers.30.mlp.fc2",
161
+ "model.layers.22.mlp.fc1",
162
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.v_proj",
163
+ "model.layers.22.self_attn.k_proj",
164
+ "model.layers.19.mlp.fc1",
165
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.5.mlp.fc1",
166
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.5.mlp.fc2",
167
+ "model.layers.14.mlp.fc1",
168
+ "model.layers.10.self_attn.q_proj",
169
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.17.mlp.fc2",
170
+ "model.layers.7.mlp.fc2",
171
+ "model.layers.23.mlp.fc1",
172
+ "model.layers.7.self_attn.k_proj",
173
+ "model.layers.13.self_attn.dense",
174
+ "model.layers.15.mlp.fc2",
175
+ "model.layers.15.self_attn.dense",
176
+ "model.layers.2.self_attn.dense",
177
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.1.self_attn.k_proj",
178
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.out_proj",
179
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.0.mlp.fc2",
180
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.k_proj",
181
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.v_proj",
182
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.q_proj",
183
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.v_proj",
184
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.k_proj",
185
+ "model.layers.24.self_attn.v_proj",
186
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.15.mlp.fc2",
187
+ "model.layers.17.mlp.fc2",
188
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.13.mlp.fc1",
189
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.k_proj",
190
+ "model.layers.28.self_attn.q_proj",
191
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.k_proj",
192
+ "model.layers.22.self_attn.v_proj",
193
+ "model.layers.29.self_attn.v_proj",
194
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.14.mlp.fc2",
195
+ "model.layers.30.mlp.fc1",
196
+ "model.layers.29.mlp.fc2",
197
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.3.mlp.fc2",
198
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.out_proj",
199
+ "model.layers.5.self_attn.dense",
200
+ "model.layers.20.self_attn.dense",
201
+ "model.layers.12.mlp.fc2",
202
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.18.self_attn.k_proj",
203
+ "model.layers.6.self_attn.v_proj",
204
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.out_proj",
205
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.q_proj",
206
+ "model.layers.20.self_attn.v_proj",
207
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.8.mlp.fc1",
208
+ "model.layers.3.self_attn.q_proj",
209
+ "model.layers.29.self_attn.dense",
210
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.k_proj",
211
+ "model.layers.30.self_attn.dense",
212
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.25.mlp.fc1",
213
+ "model.layers.8.self_attn.k_proj",
214
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.k_proj",
215
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.q_proj",
216
+ "model.layers.9.mlp.fc1",
217
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.11.self_attn.k_proj",
218
+ "model.layers.27.self_attn.dense",
219
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.7.mlp.fc2",
220
+ "model.layers.13.self_attn.k_proj",
221
+ "model.layers.26.mlp.fc2",
222
+ "model.layers.9.self_attn.dense",
223
+ "model.layers.1.self_attn.q_proj",
224
+ "model.layers.10.self_attn.dense",
225
+ "model.layers.14.self_attn.k_proj",
226
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.out_proj",
227
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.24.mlp.fc2",
228
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.17.self_attn.k_proj",
229
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.25.mlp.fc2",
230
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.23.mlp.fc2",
231
+ "model.layers.12.mlp.fc1",
232
+ "model.layers.13.self_attn.q_proj",
233
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.4.mlp.fc1",
234
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.k_proj",
235
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj",
236
+ "model.layers.26.self_attn.q_proj",
237
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.19.mlp.fc1",
238
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.out_proj",
239
+ "model.layers.4.mlp.fc1",
240
+ "model.layers.18.mlp.fc2",
241
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.7.self_attn.q_proj",
242
+ "model.layers.31.self_attn.v_proj",
243
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.v_proj",
244
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.18.mlp.fc2",
245
+ "model.layers.24.self_attn.k_proj",
246
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.out_proj",
247
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.q_proj",
248
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.out_proj",
249
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.11.mlp.fc2",
250
+ "model.layers.16.self_attn.dense",
251
+ "model.layers.25.self_attn.k_proj",
252
+ "model.layers.0.self_attn.v_proj",
253
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.21.self_attn.out_proj",
254
+ "model.layers.5.mlp.fc1",
255
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.12.mlp.fc2",
256
+ "model.layers.27.mlp.fc2",
257
+ "model.layers.23.self_attn.v_proj",
258
+ "model.layers.16.self_attn.v_proj",
259
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.out_proj",
260
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.v_proj",
261
+ "model.layers.11.mlp.fc1",
262
+ "model.layers.9.self_attn.k_proj",
263
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.11.mlp.fc1",
264
+ "model.layers.11.self_attn.v_proj",
265
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.10.self_attn.q_proj",
266
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.3.self_attn.out_proj",
267
+ "model.layers.6.self_attn.dense",
268
+ "model.layers.4.self_attn.v_proj",
269
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.15.mlp.fc1",
270
+ "model.layers.18.self_attn.k_proj",
271
+ "model.layers.7.mlp.fc1",
272
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.19.mlp.fc2",
273
+ "model.layers.22.self_attn.q_proj",
274
+ "model.layers.29.self_attn.k_proj",
275
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.12.self_attn.out_proj",
276
+ "model.layers.2.self_attn.q_proj",
277
+ "model.layers.17.self_attn.v_proj",
278
+ "model.layers.1.mlp.fc2",
279
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.13.self_attn.v_proj",
280
+ "model.layers.16.self_attn.k_proj",
281
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.6.mlp.fc1",
282
+ "model.layers.0.mlp.fc1",
283
+ "model.layers.5.mlp.fc2",
284
+ "model.layers.8.self_attn.dense",
285
+ "model.layers.21.self_attn.k_proj",
286
+ "model.layers.29.mlp.fc1",
287
+ "model.layers.9.self_attn.q_proj",
288
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.24.self_attn.k_proj",
289
+ "model.layers.5.self_attn.k_proj",
290
+ "model.layers.30.self_attn.v_proj",
291
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.k_proj",
292
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.22.mlp.fc1",
293
+ "model.layers.2.self_attn.v_proj",
294
+ "model.layers.3.self_attn.v_proj",
295
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.q_proj",
296
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.20.mlp.fc1",
297
+ "model.layers.19.self_attn.v_proj",
298
+ "model.layers.15.self_attn.q_proj",
299
+ "model.layers.6.mlp.fc1",
300
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.q_proj",
301
+ "model.layers.8.self_attn.v_proj",
302
+ "model.layers.28.self_attn.dense",
303
+ "model.layers.10.self_attn.v_proj",
304
+ "model.layers.21.mlp.fc2",
305
+ "model.layers.13.mlp.fc2",
306
+ "model.layers.10.mlp.fc2",
307
+ "model.layers.7.self_attn.v_proj",
308
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.19.self_attn.q_proj",
309
+ "model.layers.25.self_attn.v_proj",
310
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.v_proj",
311
+ "model.layers.11.self_attn.dense",
312
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.out_proj",
313
+ "model.layers.17.mlp.fc1",
314
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.14.self_attn.out_proj",
315
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.5.self_attn.out_proj",
316
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.22.self_attn.out_proj",
317
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.out_proj",
318
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.9.mlp.fc2",
319
+ "model.layers.2.self_attn.k_proj",
320
+ "model.layers.4.mlp.fc2",
321
+ "model.layers.21.self_attn.dense",
322
+ "model.layers.16.mlp.fc2",
323
+ "model.layers.18.self_attn.q_proj",
324
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj",
325
+ "model.layers.23.self_attn.dense",
326
+ "model.layers.12.self_attn.k_proj",
327
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.4.self_attn.out_proj",
328
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.6.mlp.fc2",
329
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.12.mlp.fc1",
330
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.v_proj",
331
+ "model.layers.0.mlp.fc2",
332
+ "model.layers.25.self_attn.dense",
333
+ "model.layers.25.mlp.fc1",
334
+ "model.layers.27.mlp.fc1",
335
+ "model.layers.18.mlp.fc1",
336
+ "model.layers.24.self_attn.q_proj",
337
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.8.mlp.fc2",
338
+ "model.layers.7.self_attn.dense",
339
+ "model.layers.25.mlp.fc2",
340
+ "model.layers.19.self_attn.dense",
341
+ "model.layers.16.self_attn.q_proj",
342
+ "model.layers.17.self_attn.k_proj",
343
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.15.self_attn.q_proj",
344
+ "model.layers.26.self_attn.v_proj",
345
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.9.self_attn.k_proj",
346
+ "model.layers.26.self_attn.k_proj",
347
+ "model.layers.1.self_attn.dense",
348
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.1.mlp.fc2",
349
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.20.self_attn.q_proj",
350
+ "model.layers.0.self_attn.k_proj",
351
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.6.self_attn.q_proj",
352
+ "model.layers.20.self_attn.k_proj",
353
+ "model.layers.2.mlp.fc1",
354
+ "model.layers.9.self_attn.v_proj",
355
+ "model.layers.5.self_attn.v_proj",
356
+ "model.layers.14.self_attn.v_proj",
357
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj",
358
+ "model.layers.19.mlp.fc2",
359
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.23.self_attn.k_proj",
360
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.16.self_attn.out_proj",
361
+ "model.layers.31.self_attn.dense",
362
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.out_proj",
363
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.2.self_attn.k_proj",
364
+ "model.vision_tower.vision_tower.vision_model.encoder.layers.8.self_attn.q_proj"
365
+ ],
366
+ "task_type": "CAUSAL_LM"
367
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68288f11209390933f0a4fef978ecf6dc8ff31780f0ffd6406aa77231e8bdb22
3
+ size 511714477
config.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/mnt/workspace/workgroup/ustc/Bunny-v1/",
3
+ "architectures": [
4
+ "BunnyPhiForCausalLM"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_bunny_phi.BunnyPhiConfig",
9
+ "AutoModelForCausalLM": "modeling_bunny_phi.BunnyPhiForCausalLM"
10
+ },
11
+ "bert_type": "qformer_pretrain",
12
+ "bos_token_id": 50256,
13
+ "compress_type": "mean",
14
+ "embd_pdrop": 0.0,
15
+ "eos_token_id": 50256,
16
+ "freeze_mm_mlp_adapter": false,
17
+ "hidden_act": "gelu_new",
18
+ "hidden_size": 2560,
19
+ "image_aspect_ratio": "pad",
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 10240,
22
+ "layer_norm_eps": 1e-05,
23
+ "max_position_embeddings": 2048,
24
+ "mm_hidden_size": 1152,
25
+ "mm_projector_lr": 2e-05,
26
+ "mm_projector_type": "mlp2x_gelu",
27
+ "mm_vision_tower": "/path/to/siglip-so400m-patch14-384",
28
+ "model_type": "bunny-phi",
29
+ "num_attention_heads": 32,
30
+ "num_hidden_layers": 32,
31
+ "num_key_value_heads": 32,
32
+ "pad_token_id": 50256,
33
+ "partial_rotary_factor": 0.4,
34
+ "qk_layernorm": false,
35
+ "resid_pdrop": 0.1,
36
+ "rope_scaling": null,
37
+ "rope_theta": 10000.0,
38
+ "tie_word_embeddings": false,
39
+ "tokenizer_model_max_length": 2048,
40
+ "tokenizer_padding_side": "right",
41
+ "torch_dtype": "float16",
42
+ "transformers_version": "4.39.3",
43
+ "tune_mm_mlp_adapter": false,
44
+ "use_cache": true,
45
+ "use_mm_proj": true,
46
+ "vocab_size": 50295
47
+ }
non_lora_trainables.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f4c4ea4a0a2792cb3aa2d05a992161c089f041cb1e23b739a8f2aebe3d975cc
3
+ size 19017447