aspctu commited on
Commit
257fbd1
·
verified ·
1 Parent(s): 18e5f89

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +245 -245
config.json CHANGED
@@ -97,251 +97,251 @@
97
  }
98
  },
99
  "ignore": [
100
- "language_model.lm_head",
101
- "language_model.model.layers.0.self_attn.kv_a_proj_with_mqa",
102
- "language_model.model.layers.0.self_attn.kv_b_proj",
103
- "language_model.model.layers.0.self_attn.q_a_proj",
104
- "language_model.model.layers.0.self_attn.q_b_proj",
105
- "language_model.model.layers.1.self_attn.kv_a_proj_with_mqa",
106
- "language_model.model.layers.1.self_attn.kv_b_proj",
107
- "language_model.model.layers.1.self_attn.q_a_proj",
108
- "language_model.model.layers.1.self_attn.q_b_proj",
109
- "language_model.model.layers.10.self_attn.kv_a_proj_with_mqa",
110
- "language_model.model.layers.10.self_attn.kv_b_proj",
111
- "language_model.model.layers.10.self_attn.q_a_proj",
112
- "language_model.model.layers.10.self_attn.q_b_proj",
113
- "language_model.model.layers.11.self_attn.kv_a_proj_with_mqa",
114
- "language_model.model.layers.11.self_attn.kv_b_proj",
115
- "language_model.model.layers.11.self_attn.q_a_proj",
116
- "language_model.model.layers.11.self_attn.q_b_proj",
117
- "language_model.model.layers.12.self_attn.kv_a_proj_with_mqa",
118
- "language_model.model.layers.12.self_attn.kv_b_proj",
119
- "language_model.model.layers.12.self_attn.q_a_proj",
120
- "language_model.model.layers.12.self_attn.q_b_proj",
121
- "language_model.model.layers.13.self_attn.kv_a_proj_with_mqa",
122
- "language_model.model.layers.13.self_attn.kv_b_proj",
123
- "language_model.model.layers.13.self_attn.q_a_proj",
124
- "language_model.model.layers.13.self_attn.q_b_proj",
125
- "language_model.model.layers.14.self_attn.kv_a_proj_with_mqa",
126
- "language_model.model.layers.14.self_attn.kv_b_proj",
127
- "language_model.model.layers.14.self_attn.q_a_proj",
128
- "language_model.model.layers.14.self_attn.q_b_proj",
129
- "language_model.model.layers.15.self_attn.kv_a_proj_with_mqa",
130
- "language_model.model.layers.15.self_attn.kv_b_proj",
131
- "language_model.model.layers.15.self_attn.q_a_proj",
132
- "language_model.model.layers.15.self_attn.q_b_proj",
133
- "language_model.model.layers.16.self_attn.kv_a_proj_with_mqa",
134
- "language_model.model.layers.16.self_attn.kv_b_proj",
135
- "language_model.model.layers.16.self_attn.q_a_proj",
136
- "language_model.model.layers.16.self_attn.q_b_proj",
137
- "language_model.model.layers.17.self_attn.kv_a_proj_with_mqa",
138
- "language_model.model.layers.17.self_attn.kv_b_proj",
139
- "language_model.model.layers.17.self_attn.q_a_proj",
140
- "language_model.model.layers.17.self_attn.q_b_proj",
141
- "language_model.model.layers.18.self_attn.kv_a_proj_with_mqa",
142
- "language_model.model.layers.18.self_attn.kv_b_proj",
143
- "language_model.model.layers.18.self_attn.q_a_proj",
144
- "language_model.model.layers.18.self_attn.q_b_proj",
145
- "language_model.model.layers.19.self_attn.kv_a_proj_with_mqa",
146
- "language_model.model.layers.19.self_attn.kv_b_proj",
147
- "language_model.model.layers.19.self_attn.q_a_proj",
148
- "language_model.model.layers.19.self_attn.q_b_proj",
149
- "language_model.model.layers.2.self_attn.kv_a_proj_with_mqa",
150
- "language_model.model.layers.2.self_attn.kv_b_proj",
151
- "language_model.model.layers.2.self_attn.q_a_proj",
152
- "language_model.model.layers.2.self_attn.q_b_proj",
153
- "language_model.model.layers.20.self_attn.kv_a_proj_with_mqa",
154
- "language_model.model.layers.20.self_attn.kv_b_proj",
155
- "language_model.model.layers.20.self_attn.q_a_proj",
156
- "language_model.model.layers.20.self_attn.q_b_proj",
157
- "language_model.model.layers.21.self_attn.kv_a_proj_with_mqa",
158
- "language_model.model.layers.21.self_attn.kv_b_proj",
159
- "language_model.model.layers.21.self_attn.q_a_proj",
160
- "language_model.model.layers.21.self_attn.q_b_proj",
161
- "language_model.model.layers.22.self_attn.kv_a_proj_with_mqa",
162
- "language_model.model.layers.22.self_attn.kv_b_proj",
163
- "language_model.model.layers.22.self_attn.q_a_proj",
164
- "language_model.model.layers.22.self_attn.q_b_proj",
165
- "language_model.model.layers.23.self_attn.kv_a_proj_with_mqa",
166
- "language_model.model.layers.23.self_attn.kv_b_proj",
167
- "language_model.model.layers.23.self_attn.q_a_proj",
168
- "language_model.model.layers.23.self_attn.q_b_proj",
169
- "language_model.model.layers.24.self_attn.kv_a_proj_with_mqa",
170
- "language_model.model.layers.24.self_attn.kv_b_proj",
171
- "language_model.model.layers.24.self_attn.q_a_proj",
172
- "language_model.model.layers.24.self_attn.q_b_proj",
173
- "language_model.model.layers.25.self_attn.kv_a_proj_with_mqa",
174
- "language_model.model.layers.25.self_attn.kv_b_proj",
175
- "language_model.model.layers.25.self_attn.q_a_proj",
176
- "language_model.model.layers.25.self_attn.q_b_proj",
177
- "language_model.model.layers.26.self_attn.kv_a_proj_with_mqa",
178
- "language_model.model.layers.26.self_attn.kv_b_proj",
179
- "language_model.model.layers.26.self_attn.q_a_proj",
180
- "language_model.model.layers.26.self_attn.q_b_proj",
181
- "language_model.model.layers.27.self_attn.kv_a_proj_with_mqa",
182
- "language_model.model.layers.27.self_attn.kv_b_proj",
183
- "language_model.model.layers.27.self_attn.q_a_proj",
184
- "language_model.model.layers.27.self_attn.q_b_proj",
185
- "language_model.model.layers.28.self_attn.kv_a_proj_with_mqa",
186
- "language_model.model.layers.28.self_attn.kv_b_proj",
187
- "language_model.model.layers.28.self_attn.q_a_proj",
188
- "language_model.model.layers.28.self_attn.q_b_proj",
189
- "language_model.model.layers.29.self_attn.kv_a_proj_with_mqa",
190
- "language_model.model.layers.29.self_attn.kv_b_proj",
191
- "language_model.model.layers.29.self_attn.q_a_proj",
192
- "language_model.model.layers.29.self_attn.q_b_proj",
193
- "language_model.model.layers.3.self_attn.kv_a_proj_with_mqa",
194
- "language_model.model.layers.3.self_attn.kv_b_proj",
195
- "language_model.model.layers.3.self_attn.q_a_proj",
196
- "language_model.model.layers.3.self_attn.q_b_proj",
197
- "language_model.model.layers.30.self_attn.kv_a_proj_with_mqa",
198
- "language_model.model.layers.30.self_attn.kv_b_proj",
199
- "language_model.model.layers.30.self_attn.q_a_proj",
200
- "language_model.model.layers.30.self_attn.q_b_proj",
201
- "language_model.model.layers.31.self_attn.kv_a_proj_with_mqa",
202
- "language_model.model.layers.31.self_attn.kv_b_proj",
203
- "language_model.model.layers.31.self_attn.q_a_proj",
204
- "language_model.model.layers.31.self_attn.q_b_proj",
205
- "language_model.model.layers.32.self_attn.kv_a_proj_with_mqa",
206
- "language_model.model.layers.32.self_attn.kv_b_proj",
207
- "language_model.model.layers.32.self_attn.q_a_proj",
208
- "language_model.model.layers.32.self_attn.q_b_proj",
209
- "language_model.model.layers.33.self_attn.kv_a_proj_with_mqa",
210
- "language_model.model.layers.33.self_attn.kv_b_proj",
211
- "language_model.model.layers.33.self_attn.q_a_proj",
212
- "language_model.model.layers.33.self_attn.q_b_proj",
213
- "language_model.model.layers.34.self_attn.kv_a_proj_with_mqa",
214
- "language_model.model.layers.34.self_attn.kv_b_proj",
215
- "language_model.model.layers.34.self_attn.q_a_proj",
216
- "language_model.model.layers.34.self_attn.q_b_proj",
217
- "language_model.model.layers.35.self_attn.kv_a_proj_with_mqa",
218
- "language_model.model.layers.35.self_attn.kv_b_proj",
219
- "language_model.model.layers.35.self_attn.q_a_proj",
220
- "language_model.model.layers.35.self_attn.q_b_proj",
221
- "language_model.model.layers.36.self_attn.kv_a_proj_with_mqa",
222
- "language_model.model.layers.36.self_attn.kv_b_proj",
223
- "language_model.model.layers.36.self_attn.q_a_proj",
224
- "language_model.model.layers.36.self_attn.q_b_proj",
225
- "language_model.model.layers.37.self_attn.kv_a_proj_with_mqa",
226
- "language_model.model.layers.37.self_attn.kv_b_proj",
227
- "language_model.model.layers.37.self_attn.q_a_proj",
228
- "language_model.model.layers.37.self_attn.q_b_proj",
229
- "language_model.model.layers.38.self_attn.kv_a_proj_with_mqa",
230
- "language_model.model.layers.38.self_attn.kv_b_proj",
231
- "language_model.model.layers.38.self_attn.q_a_proj",
232
- "language_model.model.layers.38.self_attn.q_b_proj",
233
- "language_model.model.layers.39.self_attn.kv_a_proj_with_mqa",
234
- "language_model.model.layers.39.self_attn.kv_b_proj",
235
- "language_model.model.layers.39.self_attn.q_a_proj",
236
- "language_model.model.layers.39.self_attn.q_b_proj",
237
- "language_model.model.layers.4.self_attn.kv_a_proj_with_mqa",
238
- "language_model.model.layers.4.self_attn.kv_b_proj",
239
- "language_model.model.layers.4.self_attn.q_a_proj",
240
- "language_model.model.layers.4.self_attn.q_b_proj",
241
- "language_model.model.layers.40.self_attn.kv_a_proj_with_mqa",
242
- "language_model.model.layers.40.self_attn.kv_b_proj",
243
- "language_model.model.layers.40.self_attn.q_a_proj",
244
- "language_model.model.layers.40.self_attn.q_b_proj",
245
- "language_model.model.layers.41.self_attn.kv_a_proj_with_mqa",
246
- "language_model.model.layers.41.self_attn.kv_b_proj",
247
- "language_model.model.layers.41.self_attn.q_a_proj",
248
- "language_model.model.layers.41.self_attn.q_b_proj",
249
- "language_model.model.layers.42.self_attn.kv_a_proj_with_mqa",
250
- "language_model.model.layers.42.self_attn.kv_b_proj",
251
- "language_model.model.layers.42.self_attn.q_a_proj",
252
- "language_model.model.layers.42.self_attn.q_b_proj",
253
- "language_model.model.layers.43.self_attn.kv_a_proj_with_mqa",
254
- "language_model.model.layers.43.self_attn.kv_b_proj",
255
- "language_model.model.layers.43.self_attn.q_a_proj",
256
- "language_model.model.layers.43.self_attn.q_b_proj",
257
- "language_model.model.layers.44.self_attn.kv_a_proj_with_mqa",
258
- "language_model.model.layers.44.self_attn.kv_b_proj",
259
- "language_model.model.layers.44.self_attn.q_a_proj",
260
- "language_model.model.layers.44.self_attn.q_b_proj",
261
- "language_model.model.layers.45.self_attn.kv_a_proj_with_mqa",
262
- "language_model.model.layers.45.self_attn.kv_b_proj",
263
- "language_model.model.layers.45.self_attn.q_a_proj",
264
- "language_model.model.layers.45.self_attn.q_b_proj",
265
- "language_model.model.layers.46.self_attn.kv_a_proj_with_mqa",
266
- "language_model.model.layers.46.self_attn.kv_b_proj",
267
- "language_model.model.layers.46.self_attn.q_a_proj",
268
- "language_model.model.layers.46.self_attn.q_b_proj",
269
- "language_model.model.layers.47.self_attn.kv_a_proj_with_mqa",
270
- "language_model.model.layers.47.self_attn.kv_b_proj",
271
- "language_model.model.layers.47.self_attn.q_a_proj",
272
- "language_model.model.layers.47.self_attn.q_b_proj",
273
- "language_model.model.layers.48.self_attn.kv_a_proj_with_mqa",
274
- "language_model.model.layers.48.self_attn.kv_b_proj",
275
- "language_model.model.layers.48.self_attn.q_a_proj",
276
- "language_model.model.layers.48.self_attn.q_b_proj",
277
- "language_model.model.layers.49.self_attn.kv_a_proj_with_mqa",
278
- "language_model.model.layers.49.self_attn.kv_b_proj",
279
- "language_model.model.layers.49.self_attn.q_a_proj",
280
- "language_model.model.layers.49.self_attn.q_b_proj",
281
- "language_model.model.layers.5.self_attn.kv_a_proj_with_mqa",
282
- "language_model.model.layers.5.self_attn.kv_b_proj",
283
- "language_model.model.layers.5.self_attn.q_a_proj",
284
- "language_model.model.layers.5.self_attn.q_b_proj",
285
- "language_model.model.layers.50.self_attn.kv_a_proj_with_mqa",
286
- "language_model.model.layers.50.self_attn.kv_b_proj",
287
- "language_model.model.layers.50.self_attn.q_a_proj",
288
- "language_model.model.layers.50.self_attn.q_b_proj",
289
- "language_model.model.layers.51.self_attn.kv_a_proj_with_mqa",
290
- "language_model.model.layers.51.self_attn.kv_b_proj",
291
- "language_model.model.layers.51.self_attn.q_a_proj",
292
- "language_model.model.layers.51.self_attn.q_b_proj",
293
- "language_model.model.layers.52.self_attn.kv_a_proj_with_mqa",
294
- "language_model.model.layers.52.self_attn.kv_b_proj",
295
- "language_model.model.layers.52.self_attn.q_a_proj",
296
- "language_model.model.layers.52.self_attn.q_b_proj",
297
- "language_model.model.layers.53.self_attn.kv_a_proj_with_mqa",
298
- "language_model.model.layers.53.self_attn.kv_b_proj",
299
- "language_model.model.layers.53.self_attn.q_a_proj",
300
- "language_model.model.layers.53.self_attn.q_b_proj",
301
- "language_model.model.layers.54.self_attn.kv_a_proj_with_mqa",
302
- "language_model.model.layers.54.self_attn.kv_b_proj",
303
- "language_model.model.layers.54.self_attn.q_a_proj",
304
- "language_model.model.layers.54.self_attn.q_b_proj",
305
- "language_model.model.layers.55.self_attn.kv_a_proj_with_mqa",
306
- "language_model.model.layers.55.self_attn.kv_b_proj",
307
- "language_model.model.layers.55.self_attn.q_a_proj",
308
- "language_model.model.layers.55.self_attn.q_b_proj",
309
- "language_model.model.layers.56.self_attn.kv_a_proj_with_mqa",
310
- "language_model.model.layers.56.self_attn.kv_b_proj",
311
- "language_model.model.layers.56.self_attn.q_a_proj",
312
- "language_model.model.layers.56.self_attn.q_b_proj",
313
- "language_model.model.layers.57.self_attn.kv_a_proj_with_mqa",
314
- "language_model.model.layers.57.self_attn.kv_b_proj",
315
- "language_model.model.layers.57.self_attn.q_a_proj",
316
- "language_model.model.layers.57.self_attn.q_b_proj",
317
- "language_model.model.layers.58.self_attn.kv_a_proj_with_mqa",
318
- "language_model.model.layers.58.self_attn.kv_b_proj",
319
- "language_model.model.layers.58.self_attn.q_a_proj",
320
- "language_model.model.layers.58.self_attn.q_b_proj",
321
- "language_model.model.layers.59.self_attn.kv_a_proj_with_mqa",
322
- "language_model.model.layers.59.self_attn.kv_b_proj",
323
- "language_model.model.layers.59.self_attn.q_a_proj",
324
- "language_model.model.layers.59.self_attn.q_b_proj",
325
- "language_model.model.layers.6.self_attn.kv_a_proj_with_mqa",
326
- "language_model.model.layers.6.self_attn.kv_b_proj",
327
- "language_model.model.layers.6.self_attn.q_a_proj",
328
- "language_model.model.layers.6.self_attn.q_b_proj",
329
- "language_model.model.layers.60.self_attn.kv_a_proj_with_mqa",
330
- "language_model.model.layers.60.self_attn.kv_b_proj",
331
- "language_model.model.layers.60.self_attn.q_a_proj",
332
- "language_model.model.layers.60.self_attn.q_b_proj",
333
- "language_model.model.layers.7.self_attn.kv_a_proj_with_mqa",
334
- "language_model.model.layers.7.self_attn.kv_b_proj",
335
- "language_model.model.layers.7.self_attn.q_a_proj",
336
- "language_model.model.layers.7.self_attn.q_b_proj",
337
- "language_model.model.layers.8.self_attn.kv_a_proj_with_mqa",
338
- "language_model.model.layers.8.self_attn.kv_b_proj",
339
- "language_model.model.layers.8.self_attn.q_a_proj",
340
- "language_model.model.layers.8.self_attn.q_b_proj",
341
- "language_model.model.layers.9.self_attn.kv_a_proj_with_mqa",
342
- "language_model.model.layers.9.self_attn.kv_b_proj",
343
- "language_model.model.layers.9.self_attn.q_a_proj",
344
- "language_model.model.layers.9.self_attn.q_b_proj",
345
  "mm_projector*",
346
  "vision_tower*"
347
  ],
 
97
  }
98
  },
99
  "ignore": [
100
+ "lm_head",
101
+ "model.layers.0.self_attn.kv_a_proj_with_mqa",
102
+ "model.layers.0.self_attn.kv_b_proj",
103
+ "model.layers.0.self_attn.q_a_proj",
104
+ "model.layers.0.self_attn.q_b_proj",
105
+ "model.layers.1.self_attn.kv_a_proj_with_mqa",
106
+ "model.layers.1.self_attn.kv_b_proj",
107
+ "model.layers.1.self_attn.q_a_proj",
108
+ "model.layers.1.self_attn.q_b_proj",
109
+ "model.layers.10.self_attn.kv_a_proj_with_mqa",
110
+ "model.layers.10.self_attn.kv_b_proj",
111
+ "model.layers.10.self_attn.q_a_proj",
112
+ "model.layers.10.self_attn.q_b_proj",
113
+ "model.layers.11.self_attn.kv_a_proj_with_mqa",
114
+ "model.layers.11.self_attn.kv_b_proj",
115
+ "model.layers.11.self_attn.q_a_proj",
116
+ "model.layers.11.self_attn.q_b_proj",
117
+ "model.layers.12.self_attn.kv_a_proj_with_mqa",
118
+ "model.layers.12.self_attn.kv_b_proj",
119
+ "model.layers.12.self_attn.q_a_proj",
120
+ "model.layers.12.self_attn.q_b_proj",
121
+ "model.layers.13.self_attn.kv_a_proj_with_mqa",
122
+ "model.layers.13.self_attn.kv_b_proj",
123
+ "model.layers.13.self_attn.q_a_proj",
124
+ "model.layers.13.self_attn.q_b_proj",
125
+ "model.layers.14.self_attn.kv_a_proj_with_mqa",
126
+ "model.layers.14.self_attn.kv_b_proj",
127
+ "model.layers.14.self_attn.q_a_proj",
128
+ "model.layers.14.self_attn.q_b_proj",
129
+ "model.layers.15.self_attn.kv_a_proj_with_mqa",
130
+ "model.layers.15.self_attn.kv_b_proj",
131
+ "model.layers.15.self_attn.q_a_proj",
132
+ "model.layers.15.self_attn.q_b_proj",
133
+ "model.layers.16.self_attn.kv_a_proj_with_mqa",
134
+ "model.layers.16.self_attn.kv_b_proj",
135
+ "model.layers.16.self_attn.q_a_proj",
136
+ "model.layers.16.self_attn.q_b_proj",
137
+ "model.layers.17.self_attn.kv_a_proj_with_mqa",
138
+ "model.layers.17.self_attn.kv_b_proj",
139
+ "model.layers.17.self_attn.q_a_proj",
140
+ "model.layers.17.self_attn.q_b_proj",
141
+ "model.layers.18.self_attn.kv_a_proj_with_mqa",
142
+ "model.layers.18.self_attn.kv_b_proj",
143
+ "model.layers.18.self_attn.q_a_proj",
144
+ "model.layers.18.self_attn.q_b_proj",
145
+ "model.layers.19.self_attn.kv_a_proj_with_mqa",
146
+ "model.layers.19.self_attn.kv_b_proj",
147
+ "model.layers.19.self_attn.q_a_proj",
148
+ "model.layers.19.self_attn.q_b_proj",
149
+ "model.layers.2.self_attn.kv_a_proj_with_mqa",
150
+ "model.layers.2.self_attn.kv_b_proj",
151
+ "model.layers.2.self_attn.q_a_proj",
152
+ "model.layers.2.self_attn.q_b_proj",
153
+ "model.layers.20.self_attn.kv_a_proj_with_mqa",
154
+ "model.layers.20.self_attn.kv_b_proj",
155
+ "model.layers.20.self_attn.q_a_proj",
156
+ "model.layers.20.self_attn.q_b_proj",
157
+ "model.layers.21.self_attn.kv_a_proj_with_mqa",
158
+ "model.layers.21.self_attn.kv_b_proj",
159
+ "model.layers.21.self_attn.q_a_proj",
160
+ "model.layers.21.self_attn.q_b_proj",
161
+ "model.layers.22.self_attn.kv_a_proj_with_mqa",
162
+ "model.layers.22.self_attn.kv_b_proj",
163
+ "model.layers.22.self_attn.q_a_proj",
164
+ "model.layers.22.self_attn.q_b_proj",
165
+ "model.layers.23.self_attn.kv_a_proj_with_mqa",
166
+ "model.layers.23.self_attn.kv_b_proj",
167
+ "model.layers.23.self_attn.q_a_proj",
168
+ "model.layers.23.self_attn.q_b_proj",
169
+ "model.layers.24.self_attn.kv_a_proj_with_mqa",
170
+ "model.layers.24.self_attn.kv_b_proj",
171
+ "model.layers.24.self_attn.q_a_proj",
172
+ "model.layers.24.self_attn.q_b_proj",
173
+ "model.layers.25.self_attn.kv_a_proj_with_mqa",
174
+ "model.layers.25.self_attn.kv_b_proj",
175
+ "model.layers.25.self_attn.q_a_proj",
176
+ "model.layers.25.self_attn.q_b_proj",
177
+ "model.layers.26.self_attn.kv_a_proj_with_mqa",
178
+ "model.layers.26.self_attn.kv_b_proj",
179
+ "model.layers.26.self_attn.q_a_proj",
180
+ "model.layers.26.self_attn.q_b_proj",
181
+ "model.layers.27.self_attn.kv_a_proj_with_mqa",
182
+ "model.layers.27.self_attn.kv_b_proj",
183
+ "model.layers.27.self_attn.q_a_proj",
184
+ "model.layers.27.self_attn.q_b_proj",
185
+ "model.layers.28.self_attn.kv_a_proj_with_mqa",
186
+ "model.layers.28.self_attn.kv_b_proj",
187
+ "model.layers.28.self_attn.q_a_proj",
188
+ "model.layers.28.self_attn.q_b_proj",
189
+ "model.layers.29.self_attn.kv_a_proj_with_mqa",
190
+ "model.layers.29.self_attn.kv_b_proj",
191
+ "model.layers.29.self_attn.q_a_proj",
192
+ "model.layers.29.self_attn.q_b_proj",
193
+ "model.layers.3.self_attn.kv_a_proj_with_mqa",
194
+ "model.layers.3.self_attn.kv_b_proj",
195
+ "model.layers.3.self_attn.q_a_proj",
196
+ "model.layers.3.self_attn.q_b_proj",
197
+ "model.layers.30.self_attn.kv_a_proj_with_mqa",
198
+ "model.layers.30.self_attn.kv_b_proj",
199
+ "model.layers.30.self_attn.q_a_proj",
200
+ "model.layers.30.self_attn.q_b_proj",
201
+ "model.layers.31.self_attn.kv_a_proj_with_mqa",
202
+ "model.layers.31.self_attn.kv_b_proj",
203
+ "model.layers.31.self_attn.q_a_proj",
204
+ "model.layers.31.self_attn.q_b_proj",
205
+ "model.layers.32.self_attn.kv_a_proj_with_mqa",
206
+ "model.layers.32.self_attn.kv_b_proj",
207
+ "model.layers.32.self_attn.q_a_proj",
208
+ "model.layers.32.self_attn.q_b_proj",
209
+ "model.layers.33.self_attn.kv_a_proj_with_mqa",
210
+ "model.layers.33.self_attn.kv_b_proj",
211
+ "model.layers.33.self_attn.q_a_proj",
212
+ "model.layers.33.self_attn.q_b_proj",
213
+ "model.layers.34.self_attn.kv_a_proj_with_mqa",
214
+ "model.layers.34.self_attn.kv_b_proj",
215
+ "model.layers.34.self_attn.q_a_proj",
216
+ "model.layers.34.self_attn.q_b_proj",
217
+ "model.layers.35.self_attn.kv_a_proj_with_mqa",
218
+ "model.layers.35.self_attn.kv_b_proj",
219
+ "model.layers.35.self_attn.q_a_proj",
220
+ "model.layers.35.self_attn.q_b_proj",
221
+ "model.layers.36.self_attn.kv_a_proj_with_mqa",
222
+ "model.layers.36.self_attn.kv_b_proj",
223
+ "model.layers.36.self_attn.q_a_proj",
224
+ "model.layers.36.self_attn.q_b_proj",
225
+ "model.layers.37.self_attn.kv_a_proj_with_mqa",
226
+ "model.layers.37.self_attn.kv_b_proj",
227
+ "model.layers.37.self_attn.q_a_proj",
228
+ "model.layers.37.self_attn.q_b_proj",
229
+ "model.layers.38.self_attn.kv_a_proj_with_mqa",
230
+ "model.layers.38.self_attn.kv_b_proj",
231
+ "model.layers.38.self_attn.q_a_proj",
232
+ "model.layers.38.self_attn.q_b_proj",
233
+ "model.layers.39.self_attn.kv_a_proj_with_mqa",
234
+ "model.layers.39.self_attn.kv_b_proj",
235
+ "model.layers.39.self_attn.q_a_proj",
236
+ "model.layers.39.self_attn.q_b_proj",
237
+ "model.layers.4.self_attn.kv_a_proj_with_mqa",
238
+ "model.layers.4.self_attn.kv_b_proj",
239
+ "model.layers.4.self_attn.q_a_proj",
240
+ "model.layers.4.self_attn.q_b_proj",
241
+ "model.layers.40.self_attn.kv_a_proj_with_mqa",
242
+ "model.layers.40.self_attn.kv_b_proj",
243
+ "model.layers.40.self_attn.q_a_proj",
244
+ "model.layers.40.self_attn.q_b_proj",
245
+ "model.layers.41.self_attn.kv_a_proj_with_mqa",
246
+ "model.layers.41.self_attn.kv_b_proj",
247
+ "model.layers.41.self_attn.q_a_proj",
248
+ "model.layers.41.self_attn.q_b_proj",
249
+ "model.layers.42.self_attn.kv_a_proj_with_mqa",
250
+ "model.layers.42.self_attn.kv_b_proj",
251
+ "model.layers.42.self_attn.q_a_proj",
252
+ "model.layers.42.self_attn.q_b_proj",
253
+ "model.layers.43.self_attn.kv_a_proj_with_mqa",
254
+ "model.layers.43.self_attn.kv_b_proj",
255
+ "model.layers.43.self_attn.q_a_proj",
256
+ "model.layers.43.self_attn.q_b_proj",
257
+ "model.layers.44.self_attn.kv_a_proj_with_mqa",
258
+ "model.layers.44.self_attn.kv_b_proj",
259
+ "model.layers.44.self_attn.q_a_proj",
260
+ "model.layers.44.self_attn.q_b_proj",
261
+ "model.layers.45.self_attn.kv_a_proj_with_mqa",
262
+ "model.layers.45.self_attn.kv_b_proj",
263
+ "model.layers.45.self_attn.q_a_proj",
264
+ "model.layers.45.self_attn.q_b_proj",
265
+ "model.layers.46.self_attn.kv_a_proj_with_mqa",
266
+ "model.layers.46.self_attn.kv_b_proj",
267
+ "model.layers.46.self_attn.q_a_proj",
268
+ "model.layers.46.self_attn.q_b_proj",
269
+ "model.layers.47.self_attn.kv_a_proj_with_mqa",
270
+ "model.layers.47.self_attn.kv_b_proj",
271
+ "model.layers.47.self_attn.q_a_proj",
272
+ "model.layers.47.self_attn.q_b_proj",
273
+ "model.layers.48.self_attn.kv_a_proj_with_mqa",
274
+ "model.layers.48.self_attn.kv_b_proj",
275
+ "model.layers.48.self_attn.q_a_proj",
276
+ "model.layers.48.self_attn.q_b_proj",
277
+ "model.layers.49.self_attn.kv_a_proj_with_mqa",
278
+ "model.layers.49.self_attn.kv_b_proj",
279
+ "model.layers.49.self_attn.q_a_proj",
280
+ "model.layers.49.self_attn.q_b_proj",
281
+ "model.layers.5.self_attn.kv_a_proj_with_mqa",
282
+ "model.layers.5.self_attn.kv_b_proj",
283
+ "model.layers.5.self_attn.q_a_proj",
284
+ "model.layers.5.self_attn.q_b_proj",
285
+ "model.layers.50.self_attn.kv_a_proj_with_mqa",
286
+ "model.layers.50.self_attn.kv_b_proj",
287
+ "model.layers.50.self_attn.q_a_proj",
288
+ "model.layers.50.self_attn.q_b_proj",
289
+ "model.layers.51.self_attn.kv_a_proj_with_mqa",
290
+ "model.layers.51.self_attn.kv_b_proj",
291
+ "model.layers.51.self_attn.q_a_proj",
292
+ "model.layers.51.self_attn.q_b_proj",
293
+ "model.layers.52.self_attn.kv_a_proj_with_mqa",
294
+ "model.layers.52.self_attn.kv_b_proj",
295
+ "model.layers.52.self_attn.q_a_proj",
296
+ "model.layers.52.self_attn.q_b_proj",
297
+ "model.layers.53.self_attn.kv_a_proj_with_mqa",
298
+ "model.layers.53.self_attn.kv_b_proj",
299
+ "model.layers.53.self_attn.q_a_proj",
300
+ "model.layers.53.self_attn.q_b_proj",
301
+ "model.layers.54.self_attn.kv_a_proj_with_mqa",
302
+ "model.layers.54.self_attn.kv_b_proj",
303
+ "model.layers.54.self_attn.q_a_proj",
304
+ "model.layers.54.self_attn.q_b_proj",
305
+ "model.layers.55.self_attn.kv_a_proj_with_mqa",
306
+ "model.layers.55.self_attn.kv_b_proj",
307
+ "model.layers.55.self_attn.q_a_proj",
308
+ "model.layers.55.self_attn.q_b_proj",
309
+ "model.layers.56.self_attn.kv_a_proj_with_mqa",
310
+ "model.layers.56.self_attn.kv_b_proj",
311
+ "model.layers.56.self_attn.q_a_proj",
312
+ "model.layers.56.self_attn.q_b_proj",
313
+ "model.layers.57.self_attn.kv_a_proj_with_mqa",
314
+ "model.layers.57.self_attn.kv_b_proj",
315
+ "model.layers.57.self_attn.q_a_proj",
316
+ "model.layers.57.self_attn.q_b_proj",
317
+ "model.layers.58.self_attn.kv_a_proj_with_mqa",
318
+ "model.layers.58.self_attn.kv_b_proj",
319
+ "model.layers.58.self_attn.q_a_proj",
320
+ "model.layers.58.self_attn.q_b_proj",
321
+ "model.layers.59.self_attn.kv_a_proj_with_mqa",
322
+ "model.layers.59.self_attn.kv_b_proj",
323
+ "model.layers.59.self_attn.q_a_proj",
324
+ "model.layers.59.self_attn.q_b_proj",
325
+ "model.layers.6.self_attn.kv_a_proj_with_mqa",
326
+ "model.layers.6.self_attn.kv_b_proj",
327
+ "model.layers.6.self_attn.q_a_proj",
328
+ "model.layers.6.self_attn.q_b_proj",
329
+ "model.layers.60.self_attn.kv_a_proj_with_mqa",
330
+ "model.layers.60.self_attn.kv_b_proj",
331
+ "model.layers.60.self_attn.q_a_proj",
332
+ "model.layers.60.self_attn.q_b_proj",
333
+ "model.layers.7.self_attn.kv_a_proj_with_mqa",
334
+ "model.layers.7.self_attn.kv_b_proj",
335
+ "model.layers.7.self_attn.q_a_proj",
336
+ "model.layers.7.self_attn.q_b_proj",
337
+ "model.layers.8.self_attn.kv_a_proj_with_mqa",
338
+ "model.layers.8.self_attn.kv_b_proj",
339
+ "model.layers.8.self_attn.q_a_proj",
340
+ "model.layers.8.self_attn.q_b_proj",
341
+ "model.layers.9.self_attn.kv_a_proj_with_mqa",
342
+ "model.layers.9.self_attn.kv_b_proj",
343
+ "model.layers.9.self_attn.q_a_proj",
344
+ "model.layers.9.self_attn.q_b_proj",
345
  "mm_projector*",
346
  "vision_tower*"
347
  ],