Safetensors
vita-mixtral
ZhitaoGao commited on
Commit
1153a29
·
verified ·
1 Parent(s): 0abf7b5

Upload adapter_config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. adapter_config.json +1127 -0
adapter_config.json ADDED
@@ -0,0 +1,1127 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "/path/to/your/VITA_ckpt",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layers_pattern": null,
10
+ "layers_to_transform": null,
11
+ "loftq_config": {},
12
+ "lora_alpha": 16,
13
+ "lora_dropout": 0.05,
14
+ "megatron_config": null,
15
+ "megatron_core": "megatron.core",
16
+ "modules_to_save": null,
17
+ "peft_type": "LORA",
18
+ "r": 64,
19
+ "rank_pattern": {},
20
+ "revision": null,
21
+ "target_modules": [
22
+ "model.layers.18.block_sparse_moe.experts.6.w3",
23
+ "model.layers.18.block_sparse_moe.experts.3.w1",
24
+ "model.layers.19.block_sparse_moe.experts.4.w1",
25
+ "model.layers.30.block_sparse_moe.experts.4.w2",
26
+ "model.audio_encoder.encoder.enc.1.encoders.15.self_attn.linear_q",
27
+ "model.layers.0.block_sparse_moe.experts.3.w3",
28
+ "model.layers.12.self_attn.v_proj",
29
+ "model.layers.27.block_sparse_moe.experts.0.w1",
30
+ "model.layers.8.block_sparse_moe.experts.5.w2",
31
+ "model.audio_encoder.encoder.enc.1.encoders.5.self_attn.linear_out",
32
+ "model.layers.2.block_sparse_moe.experts.4.w2",
33
+ "model.layers.29.self_attn.q_proj",
34
+ "model.layers.27.block_sparse_moe.experts.5.w1",
35
+ "model.layers.7.self_attn.q_proj",
36
+ "model.layers.17.block_sparse_moe.experts.1.w3",
37
+ "model.audio_encoder.encoder.enc.1.encoders.15.self_attn.linear_k",
38
+ "model.layers.23.block_sparse_moe.experts.7.w1",
39
+ "model.layers.30.block_sparse_moe.experts.5.w1",
40
+ "model.layers.20.block_sparse_moe.experts.0.w2",
41
+ "model.layers.17.block_sparse_moe.experts.2.w2",
42
+ "model.layers.15.block_sparse_moe.experts.5.w2",
43
+ "model.layers.2.block_sparse_moe.experts.2.w3",
44
+ "model.layers.22.block_sparse_moe.experts.5.w2",
45
+ "model.layers.5.block_sparse_moe.experts.2.w3",
46
+ "model.layers.6.block_sparse_moe.experts.3.w3",
47
+ "model.layers.26.self_attn.o_proj",
48
+ "model.layers.10.block_sparse_moe.experts.2.w2",
49
+ "model.layers.31.block_sparse_moe.gate",
50
+ "model.layers.15.block_sparse_moe.experts.3.w3",
51
+ "model.layers.21.block_sparse_moe.experts.7.w1",
52
+ "model.layers.2.block_sparse_moe.experts.6.w2",
53
+ "model.layers.14.block_sparse_moe.experts.6.w3",
54
+ "model.layers.25.self_attn.o_proj",
55
+ "model.layers.26.block_sparse_moe.experts.0.w1",
56
+ "model.layers.27.block_sparse_moe.experts.4.w1",
57
+ "model.audio_encoder.encoder.enc.1.encoders.4.feed_forward.w_1",
58
+ "model.layers.4.block_sparse_moe.experts.1.w2",
59
+ "model.layers.14.block_sparse_moe.experts.7.w1",
60
+ "model.layers.20.self_attn.k_proj",
61
+ "model.audio_encoder.encoder.enc.1.encoders.23.self_attn.linear_k",
62
+ "model.layers.10.block_sparse_moe.experts.3.w3",
63
+ "model.layers.25.block_sparse_moe.experts.4.w3",
64
+ "model.audio_encoder.encoder.enc.1.encoders.5.self_attn.linear_k",
65
+ "model.layers.13.self_attn.o_proj",
66
+ "model.layers.4.block_sparse_moe.experts.6.w3",
67
+ "model.layers.21.block_sparse_moe.experts.3.w2",
68
+ "model.layers.31.block_sparse_moe.experts.0.w1",
69
+ "model.layers.3.block_sparse_moe.experts.2.w2",
70
+ "model.layers.4.block_sparse_moe.experts.4.w1",
71
+ "model.layers.16.block_sparse_moe.experts.0.w1",
72
+ "model.layers.7.self_attn.k_proj",
73
+ "model.layers.13.block_sparse_moe.experts.6.w3",
74
+ "model.layers.27.block_sparse_moe.experts.2.w3",
75
+ "model.layers.5.block_sparse_moe.experts.5.w3",
76
+ "model.layers.14.block_sparse_moe.experts.5.w2",
77
+ "model.layers.31.block_sparse_moe.experts.6.w3",
78
+ "model.layers.22.block_sparse_moe.experts.4.w3",
79
+ "model.layers.4.block_sparse_moe.experts.3.w3",
80
+ "model.layers.23.self_attn.o_proj",
81
+ "model.layers.30.block_sparse_moe.experts.4.w3",
82
+ "model.layers.17.self_attn.q_proj",
83
+ "model.layers.23.block_sparse_moe.experts.4.w3",
84
+ "model.layers.18.block_sparse_moe.experts.5.w3",
85
+ "model.layers.18.block_sparse_moe.experts.1.w1",
86
+ "model.layers.26.block_sparse_moe.experts.6.w2",
87
+ "model.audio_encoder.encoder.enc.1.encoders.3.self_attn.linear_k",
88
+ "model.layers.26.block_sparse_moe.experts.0.w2",
89
+ "model.layers.6.block_sparse_moe.experts.0.w3",
90
+ "model.layers.12.block_sparse_moe.experts.7.w2",
91
+ "model.layers.21.block_sparse_moe.experts.5.w2",
92
+ "model.audio_encoder.encoder.enc.1.encoders.14.feed_forward.w_1",
93
+ "model.layers.0.block_sparse_moe.experts.3.w1",
94
+ "model.layers.2.block_sparse_moe.experts.0.w3",
95
+ "model.audio_encoder.encoder.enc.1.encoders.4.self_attn.linear_q",
96
+ "model.layers.11.block_sparse_moe.experts.5.w2",
97
+ "model.layers.21.block_sparse_moe.gate",
98
+ "model.layers.27.block_sparse_moe.experts.4.w3",
99
+ "model.layers.29.self_attn.v_proj",
100
+ "model.audio_encoder.encoder.enc.1.encoders.12.self_attn.linear_q",
101
+ "model.layers.25.block_sparse_moe.experts.7.w3",
102
+ "model.layers.25.block_sparse_moe.experts.7.w1",
103
+ "model.layers.13.block_sparse_moe.experts.0.w2",
104
+ "model.layers.6.block_sparse_moe.experts.2.w3",
105
+ "model.layers.30.block_sparse_moe.experts.6.w2",
106
+ "model.layers.18.block_sparse_moe.experts.4.w2",
107
+ "model.layers.1.block_sparse_moe.experts.4.w2",
108
+ "model.layers.1.block_sparse_moe.experts.5.w3",
109
+ "model.layers.30.block_sparse_moe.experts.0.w1",
110
+ "model.layers.16.self_attn.q_proj",
111
+ "model.layers.0.block_sparse_moe.experts.1.w3",
112
+ "model.audio_encoder.encoder.enc.1.encoders.15.self_attn.linear_out",
113
+ "model.layers.26.block_sparse_moe.experts.5.w1",
114
+ "model.layers.7.self_attn.o_proj",
115
+ "model.layers.10.block_sparse_moe.experts.2.w1",
116
+ "model.layers.29.block_sparse_moe.experts.5.w2",
117
+ "model.layers.17.block_sparse_moe.experts.4.w2",
118
+ "model.layers.5.block_sparse_moe.gate",
119
+ "model.layers.26.block_sparse_moe.gate",
120
+ "model.layers.1.block_sparse_moe.experts.4.w1",
121
+ "model.layers.17.block_sparse_moe.experts.6.w2",
122
+ "model.layers.1.block_sparse_moe.experts.3.w2",
123
+ "model.layers.15.block_sparse_moe.experts.7.w3",
124
+ "model.layers.30.block_sparse_moe.experts.2.w1",
125
+ "model.layers.31.block_sparse_moe.experts.1.w3",
126
+ "model.layers.0.block_sparse_moe.experts.2.w1",
127
+ "model.layers.8.block_sparse_moe.experts.0.w1",
128
+ "model.layers.10.self_attn.k_proj",
129
+ "model.layers.24.block_sparse_moe.experts.2.w2",
130
+ "model.layers.12.block_sparse_moe.gate",
131
+ "model.layers.23.block_sparse_moe.experts.4.w1",
132
+ "model.layers.0.block_sparse_moe.experts.4.w1",
133
+ "model.layers.24.block_sparse_moe.experts.1.w2",
134
+ "model.layers.6.block_sparse_moe.experts.7.w2",
135
+ "model.layers.30.block_sparse_moe.experts.0.w2",
136
+ "model.layers.22.block_sparse_moe.experts.0.w3",
137
+ "model.layers.24.block_sparse_moe.experts.4.w1",
138
+ "model.layers.26.block_sparse_moe.experts.4.w2",
139
+ "model.layers.9.block_sparse_moe.experts.2.w1",
140
+ "model.layers.22.block_sparse_moe.experts.2.w3",
141
+ "model.layers.24.block_sparse_moe.experts.2.w1",
142
+ "model.layers.8.block_sparse_moe.experts.2.w2",
143
+ "model.layers.28.block_sparse_moe.experts.0.w3",
144
+ "model.layers.29.block_sparse_moe.experts.5.w1",
145
+ "model.layers.29.block_sparse_moe.experts.5.w3",
146
+ "model.audio_encoder.encoder.enc.1.encoders.22.self_attn.linear_q",
147
+ "model.layers.28.block_sparse_moe.experts.1.w2",
148
+ "model.layers.0.self_attn.k_proj",
149
+ "model.layers.23.block_sparse_moe.experts.5.w3",
150
+ "model.layers.19.block_sparse_moe.experts.4.w3",
151
+ "model.layers.2.self_attn.o_proj",
152
+ "model.layers.25.self_attn.v_proj",
153
+ "model.audio_encoder.encoder.enc.1.encoders.0.self_attn.linear_q",
154
+ "model.audio_encoder.encoder.enc.1.encoders.11.self_attn.linear_v",
155
+ "model.layers.27.block_sparse_moe.experts.3.w3",
156
+ "model.layers.22.block_sparse_moe.experts.5.w3",
157
+ "model.layers.8.block_sparse_moe.experts.7.w1",
158
+ "model.layers.4.block_sparse_moe.experts.5.w1",
159
+ "model.audio_encoder.encoder.enc.1.encoders.17.feed_forward.w_2",
160
+ "model.layers.9.self_attn.q_proj",
161
+ "model.layers.13.block_sparse_moe.gate",
162
+ "model.layers.15.block_sparse_moe.experts.3.w2",
163
+ "model.layers.0.block_sparse_moe.experts.1.w1",
164
+ "model.layers.27.block_sparse_moe.experts.0.w3",
165
+ "model.audio_encoder.encoder.enc.1.encoders.23.self_attn.linear_q",
166
+ "model.audio_encoder.encoder.enc.1.encoders.10.self_attn.linear_pos",
167
+ "model.layers.12.block_sparse_moe.experts.5.w1",
168
+ "model.audio_encoder.encoder.enc.1.encoders.2.feed_forward.w_2",
169
+ "model.layers.8.block_sparse_moe.experts.1.w1",
170
+ "model.layers.7.block_sparse_moe.experts.1.w2",
171
+ "model.layers.19.block_sparse_moe.experts.5.w2",
172
+ "model.layers.8.block_sparse_moe.gate",
173
+ "model.layers.15.block_sparse_moe.experts.5.w1",
174
+ "model.layers.22.self_attn.v_proj",
175
+ "model.layers.23.block_sparse_moe.gate",
176
+ "model.layers.4.block_sparse_moe.experts.3.w2",
177
+ "model.audio_encoder.encoder.enc.1.encoders.20.self_attn.linear_k",
178
+ "model.layers.13.block_sparse_moe.experts.2.w3",
179
+ "model.layers.15.block_sparse_moe.experts.3.w1",
180
+ "model.layers.21.block_sparse_moe.experts.1.w2",
181
+ "model.audio_encoder.encoder.enc.1.encoders.21.feed_forward.w_2",
182
+ "model.audio_encoder.encoder.enc.1.encoders.16.self_attn.linear_out",
183
+ "model.audio_encoder.encoder.enc.1.encoders.8.feed_forward.w_2",
184
+ "model.layers.1.block_sparse_moe.experts.3.w1",
185
+ "model.layers.1.block_sparse_moe.experts.7.w3",
186
+ "model.layers.26.block_sparse_moe.experts.3.w2",
187
+ "model.layers.31.block_sparse_moe.experts.7.w2",
188
+ "model.audio_encoder.encoder.enc.1.encoders.19.self_attn.linear_v",
189
+ "model.layers.1.block_sparse_moe.experts.6.w3",
190
+ "model.layers.6.block_sparse_moe.experts.1.w1",
191
+ "model.layers.18.block_sparse_moe.experts.5.w1",
192
+ "model.layers.3.self_attn.k_proj",
193
+ "model.layers.2.block_sparse_moe.experts.4.w1",
194
+ "model.layers.30.block_sparse_moe.experts.4.w1",
195
+ "model.layers.30.self_attn.v_proj",
196
+ "model.layers.20.self_attn.o_proj",
197
+ "model.layers.22.block_sparse_moe.experts.6.w3",
198
+ "model.layers.0.block_sparse_moe.experts.4.w3",
199
+ "model.layers.15.block_sparse_moe.experts.6.w3",
200
+ "model.layers.24.block_sparse_moe.experts.3.w1",
201
+ "model.layers.2.self_attn.q_proj",
202
+ "model.layers.0.block_sparse_moe.experts.5.w3",
203
+ "model.layers.10.block_sparse_moe.gate",
204
+ "model.layers.14.block_sparse_moe.experts.5.w3",
205
+ "model.layers.19.block_sparse_moe.experts.7.w2",
206
+ "model.layers.28.block_sparse_moe.experts.6.w1",
207
+ "model.layers.27.block_sparse_moe.experts.6.w1",
208
+ "model.audio_encoder.encoder.enc.1.encoders.10.self_attn.linear_out",
209
+ "model.layers.31.block_sparse_moe.experts.6.w1",
210
+ "model.layers.17.block_sparse_moe.experts.0.w1",
211
+ "model.layers.28.block_sparse_moe.experts.7.w2",
212
+ "model.layers.16.block_sparse_moe.experts.6.w1",
213
+ "model.mm_projector.0",
214
+ "model.layers.3.block_sparse_moe.experts.1.w3",
215
+ "model.layers.0.block_sparse_moe.experts.7.w1",
216
+ "model.layers.13.block_sparse_moe.experts.3.w1",
217
+ "model.layers.9.block_sparse_moe.experts.3.w3",
218
+ "model.layers.21.self_attn.q_proj",
219
+ "model.layers.27.block_sparse_moe.experts.3.w2",
220
+ "model.layers.29.block_sparse_moe.experts.2.w2",
221
+ "model.audio_encoder.encoder.enc.1.encoders.20.self_attn.linear_v",
222
+ "model.layers.1.block_sparse_moe.experts.1.w3",
223
+ "model.layers.14.block_sparse_moe.experts.2.w3",
224
+ "model.layers.15.block_sparse_moe.experts.0.w3",
225
+ "model.audio_encoder.encoder.enc.1.encoders.14.self_attn.linear_out",
226
+ "model.layers.25.block_sparse_moe.experts.5.w2",
227
+ "model.layers.15.block_sparse_moe.experts.1.w1",
228
+ "model.layers.12.block_sparse_moe.experts.1.w2",
229
+ "model.layers.20.block_sparse_moe.experts.1.w1",
230
+ "model.layers.22.block_sparse_moe.experts.7.w3",
231
+ "model.layers.2.block_sparse_moe.experts.4.w3",
232
+ "model.layers.26.block_sparse_moe.experts.7.w1",
233
+ "model.layers.18.block_sparse_moe.experts.3.w2",
234
+ "model.layers.12.block_sparse_moe.experts.4.w1",
235
+ "model.audio_encoder.encoder.enc.1.encoders.5.feed_forward.w_1",
236
+ "model.layers.17.block_sparse_moe.experts.3.w2",
237
+ "model.layers.21.self_attn.k_proj",
238
+ "model.audio_encoder.encoder.enc.1.encoders.15.self_attn.linear_v",
239
+ "model.layers.15.block_sparse_moe.experts.0.w1",
240
+ "model.layers.20.block_sparse_moe.experts.7.w3",
241
+ "model.layers.3.block_sparse_moe.experts.7.w2",
242
+ "model.layers.27.block_sparse_moe.experts.2.w1",
243
+ "model.layers.14.block_sparse_moe.experts.4.w3",
244
+ "model.layers.17.block_sparse_moe.experts.5.w3",
245
+ "model.layers.10.block_sparse_moe.experts.5.w2",
246
+ "model.layers.7.block_sparse_moe.experts.7.w2",
247
+ "model.layers.5.block_sparse_moe.experts.4.w3",
248
+ "model.layers.5.block_sparse_moe.experts.5.w2",
249
+ "model.layers.7.block_sparse_moe.experts.3.w3",
250
+ "model.layers.10.block_sparse_moe.experts.5.w3",
251
+ "model.layers.10.block_sparse_moe.experts.5.w1",
252
+ "model.layers.14.block_sparse_moe.experts.1.w2",
253
+ "model.layers.25.block_sparse_moe.experts.6.w1",
254
+ "model.audio_encoder.encoder.enc.1.encoders.3.feed_forward.w_2",
255
+ "model.audio_encoder.encoder.enc.1.encoders.13.feed_forward.w_1",
256
+ "model.layers.10.block_sparse_moe.experts.6.w2",
257
+ "model.layers.9.block_sparse_moe.experts.5.w1",
258
+ "model.layers.8.block_sparse_moe.experts.6.w2",
259
+ "model.audio_encoder.encoder.enc.1.encoders.11.self_attn.linear_q",
260
+ "model.layers.24.block_sparse_moe.experts.7.w2",
261
+ "model.layers.17.block_sparse_moe.experts.0.w2",
262
+ "model.layers.22.block_sparse_moe.experts.7.w1",
263
+ "model.layers.22.block_sparse_moe.experts.5.w1",
264
+ "model.layers.3.block_sparse_moe.experts.5.w2",
265
+ "model.layers.1.block_sparse_moe.experts.5.w1",
266
+ "model.layers.11.block_sparse_moe.experts.2.w3",
267
+ "model.layers.24.self_attn.o_proj",
268
+ "model.layers.14.block_sparse_moe.experts.0.w3",
269
+ "model.layers.18.block_sparse_moe.experts.5.w2",
270
+ "model.layers.1.block_sparse_moe.experts.2.w3",
271
+ "model.layers.28.block_sparse_moe.gate",
272
+ "model.layers.20.block_sparse_moe.experts.3.w1",
273
+ "model.layers.11.block_sparse_moe.experts.6.w2",
274
+ "model.layers.6.block_sparse_moe.experts.7.w1",
275
+ "model.layers.6.block_sparse_moe.experts.1.w2",
276
+ "model.layers.20.block_sparse_moe.experts.5.w1",
277
+ "model.audio_encoder.encoder.enc.1.encoders.1.self_attn.linear_v",
278
+ "model.audio_encoder.encoder.enc.1.encoders.13.feed_forward.w_2",
279
+ "model.layers.22.block_sparse_moe.experts.3.w1",
280
+ "model.layers.21.block_sparse_moe.experts.5.w1",
281
+ "model.layers.28.block_sparse_moe.experts.4.w1",
282
+ "model.audio_encoder.encoder.enc.1.encoders.1.self_attn.linear_q",
283
+ "model.layers.23.self_attn.q_proj",
284
+ "model.audio_encoder.encoder.enc.1.encoders.2.self_attn.linear_q",
285
+ "model.layers.31.block_sparse_moe.experts.3.w1",
286
+ "model.layers.31.block_sparse_moe.experts.3.w2",
287
+ "model.layers.4.block_sparse_moe.experts.5.w2",
288
+ "model.layers.18.block_sparse_moe.experts.6.w2",
289
+ "model.layers.27.block_sparse_moe.experts.2.w2",
290
+ "model.layers.1.block_sparse_moe.experts.5.w2",
291
+ "model.layers.12.block_sparse_moe.experts.2.w2",
292
+ "model.layers.1.block_sparse_moe.experts.7.w1",
293
+ "model.layers.25.self_attn.k_proj",
294
+ "model.layers.9.self_attn.k_proj",
295
+ "model.layers.23.block_sparse_moe.experts.0.w3",
296
+ "model.layers.28.block_sparse_moe.experts.7.w1",
297
+ "model.layers.24.block_sparse_moe.experts.5.w2",
298
+ "model.layers.9.block_sparse_moe.experts.6.w1",
299
+ "model.layers.18.self_attn.q_proj",
300
+ "model.audio_encoder.encoder.enc.1.encoders.15.feed_forward.w_1",
301
+ "model.layers.31.self_attn.o_proj",
302
+ "model.layers.26.block_sparse_moe.experts.4.w1",
303
+ "model.layers.8.block_sparse_moe.experts.7.w3",
304
+ "model.layers.24.block_sparse_moe.experts.7.w1",
305
+ "model.layers.21.block_sparse_moe.experts.6.w1",
306
+ "model.layers.19.block_sparse_moe.experts.2.w2",
307
+ "model.layers.19.block_sparse_moe.experts.6.w2",
308
+ "model.layers.28.self_attn.k_proj",
309
+ "model.audio_encoder.encoder.enc.1.encoders.15.feed_forward.w_2",
310
+ "model.layers.6.block_sparse_moe.gate",
311
+ "model.audio_encoder.encoder.enc.1.encoders.16.feed_forward.w_1",
312
+ "model.layers.9.block_sparse_moe.gate",
313
+ "model.audio_encoder.encoder.enc.1.encoders.1.self_attn.linear_out",
314
+ "model.layers.12.block_sparse_moe.experts.3.w1",
315
+ "model.layers.16.block_sparse_moe.experts.1.w1",
316
+ "model.layers.6.block_sparse_moe.experts.3.w2",
317
+ "model.layers.31.block_sparse_moe.experts.2.w1",
318
+ "model.audio_encoder.encoder.enc.1.encoders.10.self_attn.linear_q",
319
+ "model.layers.0.block_sparse_moe.experts.2.w2",
320
+ "model.layers.23.block_sparse_moe.experts.6.w1",
321
+ "model.layers.18.block_sparse_moe.experts.1.w2",
322
+ "model.layers.7.block_sparse_moe.experts.6.w2",
323
+ "model.layers.5.block_sparse_moe.experts.6.w1",
324
+ "model.layers.3.self_attn.v_proj",
325
+ "model.layers.10.block_sparse_moe.experts.4.w3",
326
+ "model.audio_encoder.encoder.enc.1.encoders.12.feed_forward.w_2",
327
+ "model.layers.22.self_attn.k_proj",
328
+ "model.layers.7.block_sparse_moe.experts.0.w1",
329
+ "model.layers.12.block_sparse_moe.experts.2.w3",
330
+ "model.layers.4.block_sparse_moe.experts.3.w1",
331
+ "model.audio_encoder.encoder.enc.1.encoders.9.self_attn.linear_pos",
332
+ "model.layers.26.block_sparse_moe.experts.1.w2",
333
+ "model.layers.25.block_sparse_moe.experts.1.w2",
334
+ "model.layers.29.self_attn.o_proj",
335
+ "model.layers.6.self_attn.k_proj",
336
+ "model.layers.8.block_sparse_moe.experts.3.w1",
337
+ "model.layers.12.block_sparse_moe.experts.0.w3",
338
+ "model.layers.12.block_sparse_moe.experts.3.w2",
339
+ "model.layers.11.self_attn.o_proj",
340
+ "model.layers.10.block_sparse_moe.experts.6.w1",
341
+ "model.layers.5.block_sparse_moe.experts.4.w1",
342
+ "model.layers.19.block_sparse_moe.experts.7.w1",
343
+ "model.layers.25.block_sparse_moe.experts.5.w3",
344
+ "model.layers.31.block_sparse_moe.experts.1.w1",
345
+ "model.layers.16.self_attn.k_proj",
346
+ "model.layers.26.block_sparse_moe.experts.2.w1",
347
+ "model.layers.2.block_sparse_moe.experts.5.w2",
348
+ "model.layers.0.block_sparse_moe.experts.5.w1",
349
+ "model.layers.9.block_sparse_moe.experts.0.w1",
350
+ "model.audio_encoder.encoder.enc.1.encoders.5.self_attn.linear_pos",
351
+ "model.audio_encoder.encoder.enc.1.encoders.0.self_attn.linear_v",
352
+ "model.layers.10.block_sparse_moe.experts.1.w2",
353
+ "model.layers.1.block_sparse_moe.experts.3.w3",
354
+ "model.layers.13.block_sparse_moe.experts.7.w2",
355
+ "model.layers.27.block_sparse_moe.experts.0.w2",
356
+ "model.layers.21.block_sparse_moe.experts.6.w3",
357
+ "model.layers.11.block_sparse_moe.experts.1.w2",
358
+ "model.layers.20.self_attn.q_proj",
359
+ "model.layers.21.block_sparse_moe.experts.7.w2",
360
+ "model.layers.30.block_sparse_moe.experts.6.w3",
361
+ "model.audio_encoder.encoder.enc.1.encoders.7.self_attn.linear_v",
362
+ "model.audio_encoder.encoder.enc.1.encoders.22.feed_forward.w_2",
363
+ "model.audio_encoder.encoder.enc.1.encoders.10.self_attn.linear_v",
364
+ "model.layers.16.block_sparse_moe.experts.4.w3",
365
+ "model.audio_encoder.encoder.enc.1.encoders.2.self_attn.linear_pos",
366
+ "model.layers.0.block_sparse_moe.experts.7.w2",
367
+ "model.audio_encoder.encoder.enc.1.encoders.21.feed_forward.w_1",
368
+ "model.layers.19.self_attn.q_proj",
369
+ "model.layers.13.block_sparse_moe.experts.5.w1",
370
+ "model.layers.12.block_sparse_moe.experts.7.w1",
371
+ "model.layers.8.block_sparse_moe.experts.3.w2",
372
+ "model.layers.4.block_sparse_moe.experts.7.w3",
373
+ "model.layers.14.block_sparse_moe.experts.6.w2",
374
+ "model.layers.15.block_sparse_moe.gate",
375
+ "model.layers.23.block_sparse_moe.experts.3.w3",
376
+ "model.layers.30.self_attn.o_proj",
377
+ "model.layers.5.block_sparse_moe.experts.2.w1",
378
+ "model.layers.11.block_sparse_moe.experts.3.w1",
379
+ "model.layers.30.block_sparse_moe.experts.1.w1",
380
+ "model.audio_encoder.encoder.enc.1.encoders.21.self_attn.linear_pos",
381
+ "model.layers.27.block_sparse_moe.experts.5.w2",
382
+ "model.layers.8.block_sparse_moe.experts.4.w3",
383
+ "model.layers.18.self_attn.v_proj",
384
+ "model.layers.22.block_sparse_moe.experts.6.w1",
385
+ "model.layers.19.block_sparse_moe.experts.2.w3",
386
+ "model.layers.28.block_sparse_moe.experts.1.w3",
387
+ "model.audio_encoder.encoder.enc.1.encoders.12.self_attn.linear_v",
388
+ "model.layers.24.self_attn.v_proj",
389
+ "model.layers.23.block_sparse_moe.experts.2.w1",
390
+ "model.layers.25.self_attn.q_proj",
391
+ "model.layers.28.block_sparse_moe.experts.2.w3",
392
+ "model.layers.31.block_sparse_moe.experts.5.w2",
393
+ "model.layers.14.self_attn.o_proj",
394
+ "model.audio_encoder.encoder.enc.1.encoders.7.self_attn.linear_out",
395
+ "model.layers.16.block_sparse_moe.experts.1.w2",
396
+ "model.layers.25.block_sparse_moe.experts.3.w1",
397
+ "model.layers.5.block_sparse_moe.experts.1.w2",
398
+ "model.audio_encoder.encoder.enc.1.encoders.16.self_attn.linear_q",
399
+ "model.layers.4.block_sparse_moe.experts.1.w1",
400
+ "model.layers.14.block_sparse_moe.experts.2.w1",
401
+ "model.layers.0.block_sparse_moe.experts.1.w2",
402
+ "model.layers.23.block_sparse_moe.experts.3.w2",
403
+ "model.layers.30.block_sparse_moe.experts.1.w2",
404
+ "model.layers.27.block_sparse_moe.experts.7.w1",
405
+ "model.audio_encoder.encoder.enc.1.encoders.8.self_attn.linear_out",
406
+ "model.layers.21.block_sparse_moe.experts.0.w1",
407
+ "model.audio_encoder.encoder.enc.1.encoders.12.feed_forward.w_1",
408
+ "model.layers.14.block_sparse_moe.experts.6.w1",
409
+ "model.layers.16.block_sparse_moe.experts.3.w1",
410
+ "model.audio_encoder.encoder.enc.1.encoders.13.self_attn.linear_q",
411
+ "model.layers.27.block_sparse_moe.gate",
412
+ "model.layers.5.self_attn.q_proj",
413
+ "model.layers.15.block_sparse_moe.experts.2.w1",
414
+ "model.layers.7.block_sparse_moe.experts.6.w3",
415
+ "model.layers.19.block_sparse_moe.experts.5.w3",
416
+ "model.layers.0.block_sparse_moe.experts.6.w2",
417
+ "model.layers.7.block_sparse_moe.experts.3.w1",
418
+ "model.layers.8.block_sparse_moe.experts.1.w2",
419
+ "model.layers.13.block_sparse_moe.experts.2.w2",
420
+ "model.layers.22.block_sparse_moe.experts.6.w2",
421
+ "model.audio_encoder.encoder.enc.1.encoders.11.feed_forward.w_2",
422
+ "model.layers.6.block_sparse_moe.experts.0.w2",
423
+ "model.audio_encoder.encoder.enc.1.encoders.6.self_attn.linear_v",
424
+ "model.layers.7.block_sparse_moe.experts.7.w3",
425
+ "model.audio_encoder.encoder.enc.1.encoders.18.self_attn.linear_out",
426
+ "model.audio_encoder.encoder.enc.1.encoders.19.self_attn.linear_k",
427
+ "model.layers.2.block_sparse_moe.experts.3.w1",
428
+ "model.layers.24.block_sparse_moe.experts.6.w3",
429
+ "model.layers.9.block_sparse_moe.experts.6.w3",
430
+ "model.layers.21.block_sparse_moe.experts.1.w3",
431
+ "model.layers.19.block_sparse_moe.experts.4.w2",
432
+ "model.layers.8.block_sparse_moe.experts.5.w3",
433
+ "model.audio_encoder.encoder.enc.1.encoders.2.feed_forward.w_1",
434
+ "model.layers.12.block_sparse_moe.experts.1.w1",
435
+ "model.audio_encoder.encoder.enc.1.encoders.12.self_attn.linear_k",
436
+ "model.layers.23.block_sparse_moe.experts.1.w2",
437
+ "model.layers.0.self_attn.o_proj",
438
+ "model.layers.16.block_sparse_moe.experts.5.w3",
439
+ "model.layers.18.block_sparse_moe.experts.4.w3",
440
+ "model.layers.26.block_sparse_moe.experts.6.w3",
441
+ "model.audio_encoder.encoder.enc.1.encoders.9.self_attn.linear_q",
442
+ "model.layers.21.block_sparse_moe.experts.3.w1",
443
+ "model.layers.11.block_sparse_moe.experts.0.w1",
444
+ "model.layers.3.block_sparse_moe.experts.1.w1",
445
+ "model.audio_encoder.encoder.enc.1.encoders.8.feed_forward.w_1",
446
+ "model.layers.5.block_sparse_moe.experts.2.w2",
447
+ "model.audio_encoder.encoder.enc.1.encoders.2.self_attn.linear_k",
448
+ "model.layers.30.self_attn.k_proj",
449
+ "model.layers.4.block_sparse_moe.experts.0.w3",
450
+ "model.layers.14.block_sparse_moe.experts.3.w1",
451
+ "model.layers.8.self_attn.v_proj",
452
+ "model.layers.17.block_sparse_moe.experts.5.w2",
453
+ "model.layers.22.block_sparse_moe.experts.0.w2",
454
+ "model.layers.25.block_sparse_moe.experts.0.w3",
455
+ "model.layers.16.block_sparse_moe.experts.5.w1",
456
+ "model.layers.19.block_sparse_moe.experts.6.w1",
457
+ "model.layers.11.block_sparse_moe.experts.4.w1",
458
+ "model.layers.10.block_sparse_moe.experts.6.w3",
459
+ "model.audio_encoder.encoder.enc.1.encoders.3.self_attn.linear_v",
460
+ "model.layers.28.self_attn.v_proj",
461
+ "model.layers.1.self_attn.k_proj",
462
+ "model.layers.14.block_sparse_moe.experts.7.w2",
463
+ "model.layers.24.block_sparse_moe.experts.7.w3",
464
+ "model.layers.0.block_sparse_moe.experts.6.w1",
465
+ "model.layers.10.self_attn.o_proj",
466
+ "model.layers.21.block_sparse_moe.experts.4.w3",
467
+ "model.layers.25.block_sparse_moe.experts.2.w2",
468
+ "model.layers.31.block_sparse_moe.experts.7.w1",
469
+ "model.layers.14.block_sparse_moe.experts.0.w1",
470
+ "model.layers.31.self_attn.q_proj",
471
+ "model.audio_encoder.encoder.enc.1.encoders.6.feed_forward.w_2",
472
+ "model.audio_encoder.encoder.enc.1.encoders.10.feed_forward.w_1",
473
+ "model.layers.11.block_sparse_moe.experts.6.w1",
474
+ "model.layers.28.block_sparse_moe.experts.2.w2",
475
+ "model.layers.12.block_sparse_moe.experts.2.w1",
476
+ "model.layers.12.self_attn.o_proj",
477
+ "model.layers.26.block_sparse_moe.experts.5.w3",
478
+ "model.layers.20.self_attn.v_proj",
479
+ "model.layers.10.block_sparse_moe.experts.1.w3",
480
+ "model.audio_encoder.encoder.enc.1.encoders.13.self_attn.linear_out",
481
+ "model.audio_encoder.encoder.enc.1.encoders.19.feed_forward.w_1",
482
+ "model.audio_encoder.encoder.enc.1.encoders.3.self_attn.linear_out",
483
+ "model.layers.0.block_sparse_moe.experts.2.w3",
484
+ "model.layers.5.self_attn.o_proj",
485
+ "model.audio_encoder.encoder.enc.1.encoders.3.self_attn.linear_q",
486
+ "model.layers.8.block_sparse_moe.experts.6.w1",
487
+ "model.layers.28.block_sparse_moe.experts.1.w1",
488
+ "model.layers.29.block_sparse_moe.experts.7.w3",
489
+ "model.layers.12.block_sparse_moe.experts.0.w1",
490
+ "model.layers.6.block_sparse_moe.experts.7.w3",
491
+ "model.layers.27.block_sparse_moe.experts.7.w2",
492
+ "model.audio_encoder.encoder.enc.1.encoders.0.feed_forward.w_1",
493
+ "model.layers.9.block_sparse_moe.experts.1.w3",
494
+ "model.layers.9.block_sparse_moe.experts.6.w2",
495
+ "model.audio_encoder.encoder.enc.1.encoders.15.self_attn.linear_pos",
496
+ "model.layers.25.block_sparse_moe.experts.1.w3",
497
+ "model.layers.16.block_sparse_moe.experts.2.w3",
498
+ "model.layers.18.block_sparse_moe.experts.2.w1",
499
+ "model.layers.18.block_sparse_moe.experts.3.w3",
500
+ "model.layers.18.block_sparse_moe.experts.7.w1",
501
+ "model.layers.17.block_sparse_moe.experts.7.w3",
502
+ "model.audio_encoder.encoder.enc.1.encoders.17.self_attn.linear_k",
503
+ "model.layers.29.block_sparse_moe.experts.6.w3",
504
+ "model.audio_encoder.encoder.enc.1.encoders.0.self_attn.linear_out",
505
+ "model.layers.27.block_sparse_moe.experts.6.w2",
506
+ "model.layers.1.block_sparse_moe.experts.7.w2",
507
+ "model.layers.6.block_sparse_moe.experts.4.w2",
508
+ "model.layers.2.self_attn.k_proj",
509
+ "model.layers.4.self_attn.k_proj",
510
+ "model.layers.23.block_sparse_moe.experts.0.w2",
511
+ "model.audio_encoder.encoder.enc.1.encoders.17.self_attn.linear_q",
512
+ "model.audio_encoder.encoder.enc.1.encoders.21.self_attn.linear_k",
513
+ "model.layers.1.self_attn.o_proj",
514
+ "model.layers.5.block_sparse_moe.experts.3.w1",
515
+ "model.layers.17.block_sparse_moe.experts.3.w1",
516
+ "model.layers.3.block_sparse_moe.experts.7.w1",
517
+ "model.layers.27.block_sparse_moe.experts.3.w1",
518
+ "model.layers.1.block_sparse_moe.experts.6.w1",
519
+ "model.layers.2.block_sparse_moe.experts.6.w1",
520
+ "model.audio_encoder.encoder.enc.1.encoders.0.self_attn.linear_pos",
521
+ "model.layers.14.self_attn.v_proj",
522
+ "model.layers.16.block_sparse_moe.experts.6.w2",
523
+ "model.layers.28.self_attn.q_proj",
524
+ "model.layers.17.self_attn.k_proj",
525
+ "model.layers.29.block_sparse_moe.experts.7.w2",
526
+ "model.layers.31.block_sparse_moe.experts.0.w2",
527
+ "model.layers.30.block_sparse_moe.experts.6.w1",
528
+ "model.layers.11.block_sparse_moe.experts.6.w3",
529
+ "model.layers.11.block_sparse_moe.experts.1.w1",
530
+ "model.layers.15.block_sparse_moe.experts.6.w1",
531
+ "model.layers.18.block_sparse_moe.experts.2.w2",
532
+ "model.layers.13.block_sparse_moe.experts.6.w2",
533
+ "model.layers.13.block_sparse_moe.experts.0.w3",
534
+ "model.layers.14.block_sparse_moe.experts.7.w3",
535
+ "model.layers.6.block_sparse_moe.experts.6.w2",
536
+ "model.layers.31.self_attn.k_proj",
537
+ "model.audio_encoder.encoder.enc.1.encoders.6.self_attn.linear_q",
538
+ "model.layers.0.block_sparse_moe.experts.4.w2",
539
+ "model.layers.8.block_sparse_moe.experts.4.w2",
540
+ "model.layers.23.block_sparse_moe.experts.1.w1",
541
+ "model.audio_encoder.encoder.enc.1.encoders.8.self_attn.linear_v",
542
+ "model.layers.13.block_sparse_moe.experts.5.w3",
543
+ "model.layers.9.self_attn.v_proj",
544
+ "model.layers.27.self_attn.q_proj",
545
+ "model.layers.20.block_sparse_moe.gate",
546
+ "model.layers.8.block_sparse_moe.experts.2.w3",
547
+ "model.layers.23.block_sparse_moe.experts.2.w3",
548
+ "model.layers.30.block_sparse_moe.experts.0.w3",
549
+ "model.audio_encoder.encoder.enc.1.encoders.17.self_attn.linear_v",
550
+ "model.layers.24.self_attn.k_proj",
551
+ "model.layers.17.block_sparse_moe.experts.2.w1",
552
+ "model.layers.15.block_sparse_moe.experts.4.w3",
553
+ "model.layers.6.self_attn.q_proj",
554
+ "model.layers.2.block_sparse_moe.experts.3.w2",
555
+ "model.layers.10.block_sparse_moe.experts.4.w2",
556
+ "model.layers.13.block_sparse_moe.experts.1.w3",
557
+ "model.layers.3.block_sparse_moe.experts.6.w3",
558
+ "model.layers.30.block_sparse_moe.experts.7.w1",
559
+ "model.layers.15.block_sparse_moe.experts.4.w2",
560
+ "model.layers.19.block_sparse_moe.experts.0.w3",
561
+ "model.layers.31.block_sparse_moe.experts.0.w3",
562
+ "model.layers.4.self_attn.q_proj",
563
+ "model.audio_encoder.encoder.enc.1.encoders.14.self_attn.linear_q",
564
+ "model.layers.8.block_sparse_moe.experts.5.w1",
565
+ "model.layers.19.block_sparse_moe.experts.6.w3",
566
+ "model.layers.10.block_sparse_moe.experts.7.w1",
567
+ "model.layers.28.block_sparse_moe.experts.3.w3",
568
+ "model.layers.5.block_sparse_moe.experts.0.w3",
569
+ "model.layers.31.block_sparse_moe.experts.5.w1",
570
+ "model.audio_encoder.encoder.enc.1.encoders.17.feed_forward.w_1",
571
+ "model.layers.1.block_sparse_moe.experts.0.w2",
572
+ "model.layers.6.self_attn.o_proj",
573
+ "model.layers.30.block_sparse_moe.experts.7.w3",
574
+ "model.layers.2.block_sparse_moe.experts.0.w1",
575
+ "model.layers.24.block_sparse_moe.experts.0.w1",
576
+ "model.layers.0.block_sparse_moe.experts.6.w3",
577
+ "model.layers.5.self_attn.k_proj",
578
+ "model.layers.22.block_sparse_moe.experts.4.w1",
579
+ "model.layers.23.block_sparse_moe.experts.7.w2",
580
+ "model.audio_encoder.encoder.enc.1.encoders.20.self_attn.linear_out",
581
+ "model.layers.20.block_sparse_moe.experts.0.w1",
582
+ "model.layers.3.block_sparse_moe.experts.2.w3",
583
+ "model.layers.3.block_sparse_moe.experts.5.w1",
584
+ "model.layers.21.block_sparse_moe.experts.1.w1",
585
+ "model.layers.3.block_sparse_moe.experts.6.w2",
586
+ "model.layers.14.block_sparse_moe.experts.5.w1",
587
+ "model.audio_encoder.encoder.enc.1.encoders.2.self_attn.linear_v",
588
+ "model.audio_encoder.encoder.enc.1.encoders.6.self_attn.linear_k",
589
+ "model.layers.20.block_sparse_moe.experts.3.w3",
590
+ "model.layers.6.block_sparse_moe.experts.4.w1",
591
+ "model.audio_encoder.adpter.project",
592
+ "model.layers.5.block_sparse_moe.experts.6.w2",
593
+ "model.layers.17.block_sparse_moe.experts.0.w3",
594
+ "model.layers.12.block_sparse_moe.experts.6.w3",
595
+ "model.layers.11.block_sparse_moe.experts.7.w3",
596
+ "model.layers.31.self_attn.v_proj",
597
+ "model.layers.31.block_sparse_moe.experts.4.w1",
598
+ "model.layers.28.block_sparse_moe.experts.4.w3",
599
+ "model.layers.29.block_sparse_moe.experts.7.w1",
600
+ "model.layers.15.block_sparse_moe.experts.2.w2",
601
+ "model.layers.14.self_attn.q_proj",
602
+ "model.layers.15.block_sparse_moe.experts.0.w2",
603
+ "model.layers.21.self_attn.o_proj",
604
+ "model.audio_encoder.encoder.enc.1.encoders.22.feed_forward.w_1",
605
+ "model.layers.14.block_sparse_moe.experts.3.w2",
606
+ "model.layers.3.self_attn.o_proj",
607
+ "model.layers.13.block_sparse_moe.experts.1.w2",
608
+ "model.layers.10.block_sparse_moe.experts.0.w3",
609
+ "model.layers.11.block_sparse_moe.experts.3.w2",
610
+ "model.layers.2.block_sparse_moe.experts.6.w3",
611
+ "model.layers.3.block_sparse_moe.experts.7.w3",
612
+ "model.layers.24.block_sparse_moe.experts.1.w3",
613
+ "model.audio_encoder.encoder.enc.1.encoders.6.self_attn.linear_pos",
614
+ "model.layers.16.block_sparse_moe.experts.0.w2",
615
+ "model.layers.20.block_sparse_moe.experts.4.w2",
616
+ "model.layers.3.block_sparse_moe.experts.6.w1",
617
+ "model.layers.13.block_sparse_moe.experts.6.w1",
618
+ "model.layers.28.block_sparse_moe.experts.5.w3",
619
+ "model.layers.21.block_sparse_moe.experts.6.w2",
620
+ "model.audio_encoder.encoder.enc.1.encoders.4.feed_forward.w_2",
621
+ "model.layers.29.block_sparse_moe.experts.0.w1",
622
+ "model.layers.28.block_sparse_moe.experts.3.w2",
623
+ "model.layers.19.block_sparse_moe.experts.5.w1",
624
+ "model.layers.8.block_sparse_moe.experts.4.w1",
625
+ "model.layers.24.block_sparse_moe.experts.4.w2",
626
+ "model.layers.1.block_sparse_moe.experts.1.w1",
627
+ "model.layers.0.block_sparse_moe.experts.7.w3",
628
+ "model.layers.10.block_sparse_moe.experts.7.w3",
629
+ "model.layers.3.block_sparse_moe.experts.0.w3",
630
+ "model.layers.15.block_sparse_moe.experts.2.w3",
631
+ "model.layers.10.block_sparse_moe.experts.1.w1",
632
+ "model.layers.9.block_sparse_moe.experts.7.w2",
633
+ "model.layers.19.block_sparse_moe.experts.0.w2",
634
+ "model.layers.10.block_sparse_moe.experts.4.w1",
635
+ "model.audio_encoder.encoder.enc.1.encoders.1.self_attn.linear_pos",
636
+ "model.layers.4.block_sparse_moe.experts.6.w2",
637
+ "model.layers.0.block_sparse_moe.gate",
638
+ "model.layers.22.block_sparse_moe.experts.1.w3",
639
+ "model.layers.5.block_sparse_moe.experts.0.w2",
640
+ "model.layers.29.block_sparse_moe.experts.2.w1",
641
+ "model.layers.31.block_sparse_moe.experts.2.w3",
642
+ "model.audio_encoder.encoder.enc.1.encoders.22.self_attn.linear_v",
643
+ "model.audio_encoder.encoder.enc.1.encoders.20.self_attn.linear_q",
644
+ "model.audio_encoder.encoder.enc.1.encoders.19.feed_forward.w_2",
645
+ "model.layers.10.block_sparse_moe.experts.3.w1",
646
+ "model.layers.20.block_sparse_moe.experts.2.w3",
647
+ "model.layers.2.block_sparse_moe.experts.1.w2",
648
+ "model.audio_encoder.encoder.enc.1.encoders.7.self_attn.linear_k",
649
+ "model.layers.22.block_sparse_moe.experts.3.w3",
650
+ "model.layers.23.self_attn.k_proj",
651
+ "model.layers.27.block_sparse_moe.experts.7.w3",
652
+ "model.audio_encoder.encoder.enc.1.encoders.23.feed_forward.w_2",
653
+ "model.layers.30.block_sparse_moe.experts.5.w3",
654
+ "model.layers.23.block_sparse_moe.experts.4.w2",
655
+ "model.layers.4.self_attn.v_proj",
656
+ "model.layers.1.self_attn.q_proj",
657
+ "model.layers.13.block_sparse_moe.experts.7.w3",
658
+ "model.layers.2.block_sparse_moe.experts.3.w3",
659
+ "model.layers.7.block_sparse_moe.gate",
660
+ "model.audio_encoder.encoder.enc.1.encoders.14.self_attn.linear_k",
661
+ "model.layers.11.block_sparse_moe.experts.0.w2",
662
+ "model.audio_encoder.encoder.enc.1.encoders.16.self_attn.linear_k",
663
+ "model.audio_encoder.encoder.enc.1.encoders.20.feed_forward.w_2",
664
+ "model.layers.20.block_sparse_moe.experts.4.w1",
665
+ "model.layers.9.block_sparse_moe.experts.2.w2",
666
+ "model.layers.15.block_sparse_moe.experts.5.w3",
667
+ "model.audio_encoder.encoder.enc.1.encoders.4.self_attn.linear_k",
668
+ "model.layers.10.block_sparse_moe.experts.2.w3",
669
+ "model.layers.14.block_sparse_moe.experts.2.w2",
670
+ "model.audio_encoder.encoder.enc.1.encoders.0.feed_forward.w_2",
671
+ "model.layers.3.block_sparse_moe.gate",
672
+ "model.layers.12.self_attn.q_proj",
673
+ "model.layers.26.block_sparse_moe.experts.3.w1",
674
+ "model.layers.6.block_sparse_moe.experts.2.w1",
675
+ "model.layers.7.block_sparse_moe.experts.2.w3",
676
+ "model.layers.12.block_sparse_moe.experts.4.w3",
677
+ "model.layers.9.block_sparse_moe.experts.1.w2",
678
+ "model.layers.25.block_sparse_moe.experts.0.w1",
679
+ "model.layers.26.block_sparse_moe.experts.2.w3",
680
+ "model.audio_encoder.encoder.enc.1.encoders.2.self_attn.linear_out",
681
+ "model.audio_encoder.encoder.enc.1.encoders.3.feed_forward.w_1",
682
+ "model.layers.7.block_sparse_moe.experts.0.w3",
683
+ "model.layers.27.self_attn.k_proj",
684
+ "model.layers.31.block_sparse_moe.experts.5.w3",
685
+ "model.layers.27.block_sparse_moe.experts.1.w2",
686
+ "model.layers.28.block_sparse_moe.experts.6.w2",
687
+ "model.layers.3.block_sparse_moe.experts.4.w2",
688
+ "model.layers.3.block_sparse_moe.experts.4.w3",
689
+ "model.audio_encoder.encoder.enc.1.encoders.7.self_attn.linear_pos",
690
+ "model.layers.4.block_sparse_moe.experts.1.w3",
691
+ "model.layers.7.block_sparse_moe.experts.1.w3",
692
+ "model.layers.12.block_sparse_moe.experts.6.w1",
693
+ "model.layers.24.block_sparse_moe.experts.0.w2",
694
+ "model.layers.24.block_sparse_moe.experts.0.w3",
695
+ "model.layers.29.block_sparse_moe.experts.3.w2",
696
+ "model.layers.16.block_sparse_moe.experts.0.w3",
697
+ "model.layers.11.block_sparse_moe.experts.7.w2",
698
+ "model.audio_encoder.encoder.enc.1.encoders.11.self_attn.linear_pos",
699
+ "model.layers.7.block_sparse_moe.experts.5.w1",
700
+ "model.layers.22.block_sparse_moe.experts.0.w1",
701
+ "model.layers.30.self_attn.q_proj",
702
+ "model.layers.3.self_attn.q_proj",
703
+ "model.audio_encoder.encoder.enc.1.encoders.11.feed_forward.w_1",
704
+ "model.layers.16.block_sparse_moe.experts.4.w1",
705
+ "model.audio_encoder.encoder.enc.1.encoders.9.self_attn.linear_out",
706
+ "model.layers.5.block_sparse_moe.experts.0.w1",
707
+ "model.layers.27.block_sparse_moe.experts.4.w2",
708
+ "model.audio_encoder.encoder.enc.1.encoders.23.feed_forward.w_1",
709
+ "model.audio_encoder.encoder.enc.1.encoders.18.self_attn.linear_k",
710
+ "model.audio_encoder.encoder.enc.1.encoders.11.self_attn.linear_k",
711
+ "model.layers.3.block_sparse_moe.experts.3.w3",
712
+ "model.layers.22.block_sparse_moe.experts.2.w1",
713
+ "model.layers.25.block_sparse_moe.experts.3.w2",
714
+ "model.layers.24.block_sparse_moe.experts.5.w3",
715
+ "model.layers.30.block_sparse_moe.gate",
716
+ "model.layers.30.block_sparse_moe.experts.5.w2",
717
+ "model.layers.18.block_sparse_moe.experts.7.w2",
718
+ "model.layers.21.block_sparse_moe.experts.4.w1",
719
+ "model.audio_encoder.encoder.enc.1.encoders.5.self_attn.linear_q",
720
+ "model.layers.29.block_sparse_moe.experts.6.w2",
721
+ "model.layers.10.block_sparse_moe.experts.3.w2",
722
+ "model.layers.2.block_sparse_moe.experts.0.w2",
723
+ "model.layers.15.self_attn.q_proj",
724
+ "model.layers.25.block_sparse_moe.experts.2.w3",
725
+ "model.layers.6.block_sparse_moe.experts.0.w1",
726
+ "model.layers.16.block_sparse_moe.experts.7.w3",
727
+ "model.audio_encoder.encoder.enc.1.encoders.6.self_attn.linear_out",
728
+ "model.audio_encoder.encoder.enc.1.encoders.20.self_attn.linear_pos",
729
+ "model.layers.21.block_sparse_moe.experts.0.w2",
730
+ "model.layers.22.block_sparse_moe.experts.7.w2",
731
+ "model.layers.23.block_sparse_moe.experts.7.w3",
732
+ "model.layers.4.block_sparse_moe.experts.4.w3",
733
+ "model.layers.3.block_sparse_moe.experts.1.w2",
734
+ "model.layers.7.block_sparse_moe.experts.2.w2",
735
+ "model.layers.22.self_attn.q_proj",
736
+ "model.audio_encoder.encoder.enc.1.encoders.9.feed_forward.w_1",
737
+ "model.audio_encoder.encoder.enc.1.encoders.7.feed_forward.w_1",
738
+ "model.layers.10.block_sparse_moe.experts.7.w2",
739
+ "model.layers.24.self_attn.q_proj",
740
+ "model.layers.4.block_sparse_moe.experts.0.w1",
741
+ "model.layers.14.self_attn.k_proj",
742
+ "model.layers.30.block_sparse_moe.experts.2.w2",
743
+ "model.layers.12.block_sparse_moe.experts.3.w3",
744
+ "model.audio_encoder.encoder.enc.1.encoders.18.self_attn.linear_pos",
745
+ "model.layers.20.block_sparse_moe.experts.3.w2",
746
+ "model.layers.9.block_sparse_moe.experts.4.w2",
747
+ "model.layers.15.block_sparse_moe.experts.7.w2",
748
+ "model.layers.28.block_sparse_moe.experts.2.w1",
749
+ "model.layers.6.block_sparse_moe.experts.4.w3",
750
+ "model.layers.17.block_sparse_moe.experts.2.w3",
751
+ "model.layers.29.block_sparse_moe.experts.4.w2",
752
+ "model.layers.22.block_sparse_moe.gate",
753
+ "model.layers.26.self_attn.v_proj",
754
+ "model.audio_encoder.encoder.enc.1.encoders.4.self_attn.linear_pos",
755
+ "model.layers.6.block_sparse_moe.experts.5.w3",
756
+ "model.layers.10.self_attn.q_proj",
757
+ "model.layers.25.block_sparse_moe.experts.4.w2",
758
+ "model.layers.3.block_sparse_moe.experts.0.w1",
759
+ "model.audio_encoder.encoder.enc.1.encoders.1.self_attn.linear_k",
760
+ "model.layers.8.block_sparse_moe.experts.6.w3",
761
+ "model.layers.24.block_sparse_moe.experts.3.w2",
762
+ "model.layers.5.block_sparse_moe.experts.3.w3",
763
+ "model.layers.1.block_sparse_moe.gate",
764
+ "model.layers.4.block_sparse_moe.experts.0.w2",
765
+ "model.layers.12.block_sparse_moe.experts.1.w3",
766
+ "model.audio_encoder.encoder.enc.1.encoders.8.self_attn.linear_pos",
767
+ "model.layers.5.block_sparse_moe.experts.1.w3",
768
+ "model.layers.28.block_sparse_moe.experts.3.w1",
769
+ "model.layers.11.self_attn.q_proj",
770
+ "model.layers.19.block_sparse_moe.experts.1.w1",
771
+ "model.layers.26.block_sparse_moe.experts.7.w2",
772
+ "model.layers.8.block_sparse_moe.experts.3.w3",
773
+ "model.layers.2.block_sparse_moe.experts.5.w3",
774
+ "model.layers.5.block_sparse_moe.experts.3.w2",
775
+ "model.layers.13.block_sparse_moe.experts.2.w1",
776
+ "model.layers.15.self_attn.v_proj",
777
+ "model.layers.21.block_sparse_moe.experts.7.w3",
778
+ "model.layers.25.block_sparse_moe.experts.4.w1",
779
+ "model.layers.16.self_attn.v_proj",
780
+ "model.layers.17.self_attn.o_proj",
781
+ "model.layers.29.block_sparse_moe.experts.0.w3",
782
+ "model.layers.29.block_sparse_moe.experts.6.w1",
783
+ "model.layers.16.self_attn.o_proj",
784
+ "model.layers.3.block_sparse_moe.experts.3.w2",
785
+ "model.layers.19.block_sparse_moe.experts.3.w2",
786
+ "model.layers.6.block_sparse_moe.experts.2.w2",
787
+ "model.layers.29.block_sparse_moe.experts.1.w1",
788
+ "model.layers.5.block_sparse_moe.experts.7.w1",
789
+ "model.layers.25.block_sparse_moe.experts.3.w3",
790
+ "model.layers.11.self_attn.v_proj",
791
+ "model.layers.7.block_sparse_moe.experts.4.w1",
792
+ "model.layers.12.block_sparse_moe.experts.5.w3",
793
+ "model.layers.17.block_sparse_moe.experts.3.w3",
794
+ "model.layers.19.self_attn.v_proj",
795
+ "model.layers.20.block_sparse_moe.experts.7.w1",
796
+ "model.layers.2.block_sparse_moe.gate",
797
+ "model.layers.28.block_sparse_moe.experts.6.w3",
798
+ "model.layers.2.block_sparse_moe.experts.1.w1",
799
+ "model.layers.16.block_sparse_moe.experts.1.w3",
800
+ "model.layers.31.block_sparse_moe.experts.2.w2",
801
+ "model.layers.29.block_sparse_moe.experts.4.w3",
802
+ "model.layers.9.block_sparse_moe.experts.1.w1",
803
+ "model.layers.26.block_sparse_moe.experts.5.w2",
804
+ "model.layers.13.block_sparse_moe.experts.0.w1",
805
+ "model.layers.13.block_sparse_moe.experts.4.w2",
806
+ "model.layers.1.block_sparse_moe.experts.6.w2",
807
+ "model.layers.9.block_sparse_moe.experts.0.w3",
808
+ "model.layers.17.block_sparse_moe.experts.6.w3",
809
+ "model.layers.30.block_sparse_moe.experts.3.w2",
810
+ "model.layers.7.block_sparse_moe.experts.6.w1",
811
+ "model.layers.25.block_sparse_moe.experts.1.w1",
812
+ "model.layers.1.block_sparse_moe.experts.4.w3",
813
+ "model.layers.23.self_attn.v_proj",
814
+ "model.layers.15.self_attn.o_proj",
815
+ "model.layers.21.block_sparse_moe.experts.2.w1",
816
+ "model.layers.12.block_sparse_moe.experts.0.w2",
817
+ "model.layers.29.block_sparse_moe.experts.0.w2",
818
+ "model.layers.9.block_sparse_moe.experts.3.w2",
819
+ "model.layers.17.block_sparse_moe.experts.5.w1",
820
+ "model.layers.27.block_sparse_moe.experts.1.w3",
821
+ "model.layers.29.block_sparse_moe.experts.1.w3",
822
+ "model.layers.1.block_sparse_moe.experts.1.w2",
823
+ "model.layers.31.block_sparse_moe.experts.7.w3",
824
+ "model.layers.19.block_sparse_moe.experts.0.w1",
825
+ "model.layers.18.block_sparse_moe.experts.0.w3",
826
+ "model.audio_encoder.encoder.enc.1.encoders.7.feed_forward.w_2",
827
+ "model.audio_encoder.encoder.enc.1.encoders.18.feed_forward.w_2",
828
+ "model.layers.0.block_sparse_moe.experts.0.w2",
829
+ "model.layers.6.block_sparse_moe.experts.5.w2",
830
+ "model.layers.12.block_sparse_moe.experts.4.w2",
831
+ "model.audio_encoder.encoder.enc.1.encoders.18.feed_forward.w_1",
832
+ "model.audio_encoder.encoder.enc.1.encoders.19.self_attn.linear_pos",
833
+ "model.layers.23.block_sparse_moe.experts.6.w3",
834
+ "model.layers.7.block_sparse_moe.experts.4.w2",
835
+ "model.layers.26.block_sparse_moe.experts.1.w3",
836
+ "model.layers.27.block_sparse_moe.experts.5.w3",
837
+ "model.audio_encoder.encoder.enc.1.encoders.23.self_attn.linear_out",
838
+ "model.layers.27.block_sparse_moe.experts.1.w1",
839
+ "model.layers.16.block_sparse_moe.experts.3.w3",
840
+ "model.layers.21.block_sparse_moe.experts.3.w3",
841
+ "model.audio_encoder.encoder.enc.1.encoders.13.self_attn.linear_v",
842
+ "model.layers.10.block_sparse_moe.experts.0.w2",
843
+ "model.layers.20.block_sparse_moe.experts.6.w1",
844
+ "model.layers.21.block_sparse_moe.experts.2.w2",
845
+ "model.layers.2.block_sparse_moe.experts.1.w3",
846
+ "model.layers.26.self_attn.k_proj",
847
+ "model.layers.7.block_sparse_moe.experts.0.w2",
848
+ "model.layers.0.self_attn.q_proj",
849
+ "model.layers.23.block_sparse_moe.experts.0.w1",
850
+ "model.layers.8.self_attn.q_proj",
851
+ "model.layers.13.self_attn.k_proj",
852
+ "model.layers.5.self_attn.v_proj",
853
+ "model.layers.11.self_attn.k_proj",
854
+ "model.layers.17.block_sparse_moe.experts.4.w1",
855
+ "model.layers.21.block_sparse_moe.experts.5.w3",
856
+ "model.layers.31.block_sparse_moe.experts.1.w2",
857
+ "model.layers.25.block_sparse_moe.gate",
858
+ "model.audio_encoder.encoder.enc.1.encoders.18.self_attn.linear_q",
859
+ "model.layers.18.block_sparse_moe.experts.0.w2",
860
+ "model.layers.19.block_sparse_moe.experts.3.w1",
861
+ "model.layers.30.block_sparse_moe.experts.3.w3",
862
+ "model.layers.31.block_sparse_moe.experts.6.w2",
863
+ "model.layers.24.block_sparse_moe.experts.3.w3",
864
+ "model.layers.11.block_sparse_moe.experts.4.w2",
865
+ "model.layers.14.block_sparse_moe.experts.3.w3",
866
+ "model.layers.7.block_sparse_moe.experts.1.w1",
867
+ "model.layers.25.block_sparse_moe.experts.5.w1",
868
+ "model.mm_projector.2",
869
+ "model.layers.4.block_sparse_moe.gate",
870
+ "model.audio_encoder.encoder.enc.1.encoders.8.self_attn.linear_k",
871
+ "model.layers.22.block_sparse_moe.experts.1.w1",
872
+ "model.layers.15.block_sparse_moe.experts.1.w2",
873
+ "model.layers.18.block_sparse_moe.experts.6.w1",
874
+ "model.layers.11.block_sparse_moe.experts.4.w3",
875
+ "model.layers.17.block_sparse_moe.experts.7.w2",
876
+ "model.layers.22.block_sparse_moe.experts.2.w2",
877
+ "model.layers.23.block_sparse_moe.experts.1.w3",
878
+ "model.layers.18.self_attn.k_proj",
879
+ "model.layers.16.block_sparse_moe.experts.6.w3",
880
+ "model.layers.18.block_sparse_moe.experts.2.w3",
881
+ "model.layers.1.self_attn.v_proj",
882
+ "model.layers.20.block_sparse_moe.experts.5.w2",
883
+ "model.layers.15.self_attn.k_proj",
884
+ "model.layers.14.block_sparse_moe.gate",
885
+ "model.layers.29.self_attn.k_proj",
886
+ "model.layers.14.block_sparse_moe.experts.1.w3",
887
+ "model.layers.11.block_sparse_moe.gate",
888
+ "model.layers.13.block_sparse_moe.experts.4.w3",
889
+ "model.layers.27.self_attn.o_proj",
890
+ "model.layers.28.block_sparse_moe.experts.7.w3",
891
+ "model.audio_encoder.encoder.enc.1.encoders.4.self_attn.linear_out",
892
+ "model.layers.8.block_sparse_moe.experts.7.w2",
893
+ "model.layers.12.block_sparse_moe.experts.7.w3",
894
+ "model.layers.2.block_sparse_moe.experts.2.w1",
895
+ "model.layers.21.block_sparse_moe.experts.4.w2",
896
+ "model.layers.29.block_sparse_moe.gate",
897
+ "model.layers.16.block_sparse_moe.experts.3.w2",
898
+ "model.layers.24.block_sparse_moe.experts.6.w2",
899
+ "model.layers.19.block_sparse_moe.experts.3.w3",
900
+ "model.layers.0.self_attn.v_proj",
901
+ "model.layers.1.block_sparse_moe.experts.2.w1",
902
+ "model.layers.8.block_sparse_moe.experts.2.w1",
903
+ "model.layers.1.block_sparse_moe.experts.0.w1",
904
+ "model.layers.26.block_sparse_moe.experts.1.w1",
905
+ "model.layers.3.block_sparse_moe.experts.2.w1",
906
+ "model.layers.25.block_sparse_moe.experts.2.w1",
907
+ "model.layers.5.block_sparse_moe.experts.7.w3",
908
+ "model.layers.16.block_sparse_moe.experts.5.w2",
909
+ "model.layers.25.block_sparse_moe.experts.7.w2",
910
+ "model.layers.26.block_sparse_moe.experts.3.w3",
911
+ "model.layers.27.block_sparse_moe.experts.6.w3",
912
+ "model.audio_encoder.encoder.enc.1.encoders.22.self_attn.linear_pos",
913
+ "model.layers.9.block_sparse_moe.experts.5.w3",
914
+ "model.audio_encoder.encoder.enc.1.encoders.8.self_attn.linear_q",
915
+ "model.layers.26.block_sparse_moe.experts.7.w3",
916
+ "model.layers.0.block_sparse_moe.experts.0.w1",
917
+ "model.audio_encoder.encoder.enc.1.encoders.17.self_attn.linear_out",
918
+ "model.layers.4.block_sparse_moe.experts.2.w1",
919
+ "model.layers.2.block_sparse_moe.experts.7.w3",
920
+ "model.layers.15.block_sparse_moe.experts.6.w2",
921
+ "model.layers.19.self_attn.o_proj",
922
+ "model.layers.18.block_sparse_moe.experts.4.w1",
923
+ "model.layers.17.block_sparse_moe.gate",
924
+ "model.audio_encoder.encoder.enc.1.encoders.5.self_attn.linear_v",
925
+ "model.layers.6.block_sparse_moe.experts.1.w3",
926
+ "model.layers.15.block_sparse_moe.experts.7.w1",
927
+ "model.audio_encoder.encoder.enc.1.encoders.9.self_attn.linear_k",
928
+ "model.layers.4.block_sparse_moe.experts.7.w2",
929
+ "model.layers.13.block_sparse_moe.experts.1.w1",
930
+ "model.audio_encoder.encoder.enc.1.encoders.19.self_attn.linear_out",
931
+ "model.layers.10.block_sparse_moe.experts.0.w1",
932
+ "model.layers.6.block_sparse_moe.experts.3.w1",
933
+ "model.layers.0.block_sparse_moe.experts.3.w2",
934
+ "model.layers.30.block_sparse_moe.experts.3.w1",
935
+ "model.layers.25.block_sparse_moe.experts.0.w2",
936
+ "model.layers.29.block_sparse_moe.experts.2.w3",
937
+ "model.layers.10.self_attn.v_proj",
938
+ "model.layers.25.block_sparse_moe.experts.6.w3",
939
+ "model.layers.11.block_sparse_moe.experts.5.w3",
940
+ "model.layers.9.block_sparse_moe.experts.5.w2",
941
+ "model.audio_encoder.encoder.enc.1.encoders.16.self_attn.linear_pos",
942
+ "model.layers.22.block_sparse_moe.experts.3.w2",
943
+ "model.layers.7.block_sparse_moe.experts.2.w1",
944
+ "model.layers.9.block_sparse_moe.experts.3.w1",
945
+ "model.layers.11.block_sparse_moe.experts.7.w1",
946
+ "model.audio_encoder.encoder.enc.1.embed.0",
947
+ "model.layers.26.block_sparse_moe.experts.4.w3",
948
+ "model.layers.8.block_sparse_moe.experts.0.w2",
949
+ "model.layers.29.block_sparse_moe.experts.3.w3",
950
+ "model.layers.11.block_sparse_moe.experts.2.w1",
951
+ "model.layers.3.block_sparse_moe.experts.5.w3",
952
+ "model.layers.9.block_sparse_moe.experts.2.w3",
953
+ "model.audio_encoder.encoder.enc.1.encoders.14.self_attn.linear_pos",
954
+ "model.layers.19.block_sparse_moe.experts.1.w2",
955
+ "model.layers.23.block_sparse_moe.experts.5.w1",
956
+ "model.layers.14.block_sparse_moe.experts.4.w2",
957
+ "model.layers.17.block_sparse_moe.experts.4.w3",
958
+ "model.layers.12.block_sparse_moe.experts.6.w2",
959
+ "model.layers.25.block_sparse_moe.experts.6.w2",
960
+ "model.layers.7.block_sparse_moe.experts.5.w2",
961
+ "model.audio_encoder.encoder.enc.1.encoders.13.self_attn.linear_pos",
962
+ "model.layers.7.block_sparse_moe.experts.7.w1",
963
+ "model.layers.20.block_sparse_moe.experts.0.w3",
964
+ "model.layers.31.block_sparse_moe.experts.4.w2",
965
+ "model.audio_encoder.encoder.enc.1.encoders.5.feed_forward.w_2",
966
+ "model.layers.6.self_attn.v_proj",
967
+ "model.layers.21.block_sparse_moe.experts.0.w3",
968
+ "model.audio_encoder.encoder.enc.1.encoders.23.self_attn.linear_v",
969
+ "model.layers.17.self_attn.v_proj",
970
+ "model.layers.12.block_sparse_moe.experts.5.w2",
971
+ "model.audio_encoder.encoder.enc.1.encoders.12.self_attn.linear_out",
972
+ "model.layers.13.block_sparse_moe.experts.5.w2",
973
+ "model.layers.7.self_attn.v_proj",
974
+ "model.layers.2.block_sparse_moe.experts.5.w1",
975
+ "model.layers.20.block_sparse_moe.experts.2.w1",
976
+ "model.layers.16.block_sparse_moe.gate",
977
+ "model.layers.24.block_sparse_moe.experts.5.w1",
978
+ "model.audio_encoder.encoder.enc.1.encoders.6.feed_forward.w_1",
979
+ "model.layers.1.block_sparse_moe.experts.2.w2",
980
+ "model.layers.8.block_sparse_moe.experts.0.w3",
981
+ "model.layers.11.block_sparse_moe.experts.5.w1",
982
+ "model.layers.9.block_sparse_moe.experts.7.w3",
983
+ "model.layers.19.block_sparse_moe.gate",
984
+ "model.layers.20.block_sparse_moe.experts.1.w3",
985
+ "model.layers.23.block_sparse_moe.experts.3.w1",
986
+ "model.layers.11.block_sparse_moe.experts.0.w3",
987
+ "model.layers.8.self_attn.o_proj",
988
+ "model.audio_encoder.encoder.enc.1.encoders.16.self_attn.linear_v",
989
+ "model.layers.28.block_sparse_moe.experts.0.w1",
990
+ "model.layers.11.block_sparse_moe.experts.2.w2",
991
+ "model.layers.14.block_sparse_moe.experts.1.w1",
992
+ "model.audio_encoder.encoder.enc.1.encoders.14.self_attn.linear_v",
993
+ "model.layers.29.block_sparse_moe.experts.4.w1",
994
+ "model.layers.8.self_attn.k_proj",
995
+ "model.audio_encoder.encoder.enc.1.encoders.21.self_attn.linear_out",
996
+ "model.layers.8.block_sparse_moe.experts.1.w3",
997
+ "model.layers.19.block_sparse_moe.experts.2.w1",
998
+ "model.layers.2.block_sparse_moe.experts.7.w2",
999
+ "model.layers.5.block_sparse_moe.experts.5.w1",
1000
+ "model.layers.13.block_sparse_moe.experts.3.w2",
1001
+ "model.layers.16.block_sparse_moe.experts.2.w2",
1002
+ "model.layers.19.block_sparse_moe.experts.1.w3",
1003
+ "model.layers.21.self_attn.v_proj",
1004
+ "model.layers.29.block_sparse_moe.experts.3.w1",
1005
+ "model.layers.18.block_sparse_moe.gate",
1006
+ "model.audio_encoder.encoder.enc.1.encoders.23.self_attn.linear_pos",
1007
+ "model.layers.24.block_sparse_moe.gate",
1008
+ "model.layers.17.block_sparse_moe.experts.6.w1",
1009
+ "model.audio_encoder.encoder.enc.1.encoders.21.self_attn.linear_v",
1010
+ "model.layers.24.block_sparse_moe.experts.6.w1",
1011
+ "model.audio_encoder.encoder.enc.1.encoders.21.self_attn.linear_q",
1012
+ "model.layers.19.self_attn.k_proj",
1013
+ "model.layers.9.block_sparse_moe.experts.4.w1",
1014
+ "model.layers.20.block_sparse_moe.experts.7.w2",
1015
+ "model.layers.4.block_sparse_moe.experts.2.w2",
1016
+ "model.audio_encoder.encoder.enc.1.encoders.17.self_attn.linear_pos",
1017
+ "model.layers.14.block_sparse_moe.experts.4.w1",
1018
+ "model.layers.27.self_attn.v_proj",
1019
+ "model.audio_encoder.encoder.enc.1.encoders.11.self_attn.linear_out",
1020
+ "model.audio_encoder.encoder.enc.1.encoders.16.feed_forward.w_2",
1021
+ "model.layers.4.block_sparse_moe.experts.7.w1",
1022
+ "model.layers.24.block_sparse_moe.experts.1.w1",
1023
+ "model.layers.31.block_sparse_moe.experts.4.w3",
1024
+ "model.layers.24.block_sparse_moe.experts.2.w3",
1025
+ "model.layers.17.block_sparse_moe.experts.1.w1",
1026
+ "model.layers.26.self_attn.q_proj",
1027
+ "model.layers.26.block_sparse_moe.experts.0.w3",
1028
+ "model.audio_encoder.encoder.enc.1.encoders.10.self_attn.linear_k",
1029
+ "model.layers.13.block_sparse_moe.experts.4.w1",
1030
+ "model.audio_encoder.encoder.enc.1.encoders.20.feed_forward.w_1",
1031
+ "model.layers.29.block_sparse_moe.experts.1.w2",
1032
+ "model.layers.6.block_sparse_moe.experts.6.w3",
1033
+ "model.audio_encoder.encoder.enc.1.encoders.14.feed_forward.w_2",
1034
+ "model.layers.13.block_sparse_moe.experts.3.w3",
1035
+ "model.layers.16.block_sparse_moe.experts.7.w2",
1036
+ "model.audio_encoder.encoder.enc.1.encoders.4.self_attn.linear_v",
1037
+ "model.layers.2.block_sparse_moe.experts.2.w2",
1038
+ "model.layers.4.self_attn.o_proj",
1039
+ "model.layers.11.block_sparse_moe.experts.1.w3",
1040
+ "model.layers.20.block_sparse_moe.experts.5.w3",
1041
+ "model.layers.13.self_attn.v_proj",
1042
+ "model.layers.12.self_attn.k_proj",
1043
+ "model.layers.23.block_sparse_moe.experts.6.w2",
1044
+ "model.layers.3.block_sparse_moe.experts.3.w1",
1045
+ "model.layers.7.block_sparse_moe.experts.3.w2",
1046
+ "model.layers.30.block_sparse_moe.experts.7.w2",
1047
+ "model.audio_encoder.encoder.enc.1.encoders.13.self_attn.linear_k",
1048
+ "model.layers.5.block_sparse_moe.experts.4.w2",
1049
+ "model.audio_encoder.encoder.enc.1.encoders.22.self_attn.linear_out",
1050
+ "model.layers.18.self_attn.o_proj",
1051
+ "model.layers.6.block_sparse_moe.experts.6.w1",
1052
+ "model.layers.3.block_sparse_moe.experts.4.w1",
1053
+ "model.audio_encoder.encoder.enc.0.core.out.0",
1054
+ "model.layers.1.block_sparse_moe.experts.0.w3",
1055
+ "model.layers.7.block_sparse_moe.experts.5.w3",
1056
+ "model.layers.2.self_attn.v_proj",
1057
+ "model.layers.15.block_sparse_moe.experts.1.w3",
1058
+ "model.layers.9.block_sparse_moe.experts.4.w3",
1059
+ "model.audio_encoder.encoder.enc.1.encoders.19.self_attn.linear_q",
1060
+ "model.layers.0.block_sparse_moe.experts.0.w3",
1061
+ "model.layers.9.block_sparse_moe.experts.7.w1",
1062
+ "model.layers.16.block_sparse_moe.experts.4.w2",
1063
+ "model.audio_encoder.encoder.enc.1.encoders.9.self_attn.linear_v",
1064
+ "model.audio_encoder.encoder.enc.1.encoders.3.self_attn.linear_pos",
1065
+ "model.layers.28.block_sparse_moe.experts.4.w2",
1066
+ "model.audio_encoder.encoder.enc.1.encoders.22.self_attn.linear_k",
1067
+ "model.layers.4.block_sparse_moe.experts.5.w3",
1068
+ "model.audio_encoder.encoder.enc.1.encoders.0.self_attn.linear_k",
1069
+ "model.layers.16.block_sparse_moe.experts.2.w1",
1070
+ "model.audio_encoder.encoder.enc.1.encoders.1.feed_forward.w_2",
1071
+ "model.audio_encoder.encoder.enc.1.encoders.12.self_attn.linear_pos",
1072
+ "model.layers.26.block_sparse_moe.experts.6.w1",
1073
+ "model.layers.3.block_sparse_moe.experts.0.w2",
1074
+ "model.layers.18.block_sparse_moe.experts.7.w3",
1075
+ "model.layers.5.block_sparse_moe.experts.7.w2",
1076
+ "model.layers.20.block_sparse_moe.experts.6.w3",
1077
+ "model.layers.22.self_attn.o_proj",
1078
+ "model.layers.20.block_sparse_moe.experts.4.w3",
1079
+ "model.layers.17.block_sparse_moe.experts.1.w2",
1080
+ "model.layers.23.block_sparse_moe.experts.5.w2",
1081
+ "model.layers.7.block_sparse_moe.experts.4.w3",
1082
+ "model.layers.30.block_sparse_moe.experts.1.w3",
1083
+ "model.layers.28.block_sparse_moe.experts.0.w2",
1084
+ "model.layers.15.block_sparse_moe.experts.4.w1",
1085
+ "model.audio_encoder.encoder.enc.1.encoders.18.self_attn.linear_v",
1086
+ "model.layers.16.block_sparse_moe.experts.7.w1",
1087
+ "model.layers.18.block_sparse_moe.experts.1.w3",
1088
+ "model.layers.24.block_sparse_moe.experts.4.w3",
1089
+ "model.layers.28.block_sparse_moe.experts.5.w2",
1090
+ "model.layers.22.block_sparse_moe.experts.4.w2",
1091
+ "model.layers.17.block_sparse_moe.experts.7.w1",
1092
+ "model.layers.28.block_sparse_moe.experts.5.w1",
1093
+ "model.layers.6.block_sparse_moe.experts.5.w1",
1094
+ "model.layers.20.block_sparse_moe.experts.1.w2",
1095
+ "model.layers.11.block_sparse_moe.experts.3.w3",
1096
+ "model.layers.13.block_sparse_moe.experts.7.w1",
1097
+ "model.layers.23.block_sparse_moe.experts.2.w2",
1098
+ "model.layers.4.block_sparse_moe.experts.2.w3",
1099
+ "model.audio_encoder.encoder.enc.1.encoders.1.feed_forward.w_1",
1100
+ "model.layers.5.block_sparse_moe.experts.6.w3",
1101
+ "model.layers.28.self_attn.o_proj",
1102
+ "model.layers.4.block_sparse_moe.experts.4.w2",
1103
+ "model.layers.19.block_sparse_moe.experts.7.w3",
1104
+ "model.layers.18.block_sparse_moe.experts.0.w1",
1105
+ "model.layers.20.block_sparse_moe.experts.2.w2",
1106
+ "model.layers.9.block_sparse_moe.experts.0.w2",
1107
+ "model.layers.21.block_sparse_moe.experts.2.w3",
1108
+ "model.layers.22.block_sparse_moe.experts.1.w2",
1109
+ "model.layers.14.block_sparse_moe.experts.0.w2",
1110
+ "model.layers.5.block_sparse_moe.experts.1.w1",
1111
+ "model.layers.4.block_sparse_moe.experts.6.w1",
1112
+ "model.layers.2.block_sparse_moe.experts.7.w1",
1113
+ "model.layers.20.block_sparse_moe.experts.6.w2",
1114
+ "model.layers.30.block_sparse_moe.experts.2.w3",
1115
+ "model.audio_encoder.encoder.enc.1.encoders.7.self_attn.linear_q",
1116
+ "model.audio_encoder.encoder.enc.1.encoders.10.feed_forward.w_2",
1117
+ "model.audio_encoder.encoder.enc.1.encoders.9.feed_forward.w_2",
1118
+ "model.layers.13.self_attn.q_proj",
1119
+ "model.layers.0.block_sparse_moe.experts.5.w2",
1120
+ "model.layers.26.block_sparse_moe.experts.2.w2",
1121
+ "model.layers.31.block_sparse_moe.experts.3.w3",
1122
+ "model.layers.9.self_attn.o_proj"
1123
+ ],
1124
+ "task_type": "CAUSAL_LM",
1125
+ "use_dora": false,
1126
+ "use_rslora": false
1127
+ }