Hugging Face's logo

mesolitica
/
malaysian-mistral-malaysian-whisper-small-audio-alignment

Feature Extraction
Transformers
Safetensors
mm_llms
custom_code
Model card Files Community
huseinzol05 commited on
Commit
fb8e179
•
1 Parent(s): 88126ab

Upload MM_LLMs

Browse files
Files changed (9) hide show
  1. README.md +201 -0
  2. config.json +293 -0
  3. generation_config.json +4 -0
  4. model-00001-of-00004.safetensors +3 -0
  5. model-00002-of-00004.safetensors +3 -0
  6. model-00003-of-00004.safetensors +3 -0
  7. model-00004-of-00004.safetensors +3 -0
  8. model.safetensors.index.json +781 -0
  9. modeling_audio.py +261 -0
README.md ADDED
@@ -0,0 +1,201 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ tags: []
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+ This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+
201
+
config.json ADDED
@@ -0,0 +1,293 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "audio-alignment-mistral/checkpoint-7800",
3
+ "architectures": [
4
+ "MM_LLMs"
5
+ ],
6
+ "audio_config": {
7
+ "_name_or_path": "mesolitica/malaysian-whisper-small",
8
+ "activation_dropout": 0.0,
9
+ "activation_function": "gelu",
10
+ "add_cross_attention": false,
11
+ "apply_spec_augment": false,
12
+ "architectures": [
13
+ "WhisperForConditionalGeneration"
14
+ ],
15
+ "attention_dropout": 0.0,
16
+ "bad_words_ids": null,
17
+ "begin_suppress_tokens": [
18
+ 220,
19
+ 50257
20
+ ],
21
+ "bos_token_id": 50257,
22
+ "chunk_size_feed_forward": 0,
23
+ "classifier_proj_size": 256,
24
+ "cross_attention_hidden_size": null,
25
+ "d_model": 768,
26
+ "decoder_attention_heads": 12,
27
+ "decoder_ffn_dim": 3072,
28
+ "decoder_layerdrop": 0.0,
29
+ "decoder_layers": 12,
30
+ "decoder_start_token_id": 50258,
31
+ "diversity_penalty": 0.0,
32
+ "do_sample": false,
33
+ "dropout": 0.0,
34
+ "early_stopping": false,
35
+ "encoder_attention_heads": 12,
36
+ "encoder_ffn_dim": 3072,
37
+ "encoder_layerdrop": 0.0,
38
+ "encoder_layers": 12,
39
+ "encoder_no_repeat_ngram_size": 0,
40
+ "eos_token_id": 50257,
41
+ "exponential_decay_length_penalty": null,
42
+ "finetuning_task": null,
43
+ "forced_bos_token_id": null,
44
+ "forced_decoder_ids": [
45
+ [
46
+ 1,
47
+ 50259
48
+ ],
49
+ [
50
+ 2,
51
+ 50359
52
+ ],
53
+ [
54
+ 3,
55
+ 50363
56
+ ]
57
+ ],
58
+ "forced_eos_token_id": null,
59
+ "id2label": {
60
+ "0": "LABEL_0",
61
+ "1": "LABEL_1"
62
+ },
63
+ "init_std": 0.02,
64
+ "is_decoder": false,
65
+ "is_encoder_decoder": true,
66
+ "label2id": {
67
+ "LABEL_0": 0,
68
+ "LABEL_1": 1
69
+ },
70
+ "length_penalty": 1.0,
71
+ "mask_feature_length": 10,
72
+ "mask_feature_min_masks": 0,
73
+ "mask_feature_prob": 0.0,
74
+ "mask_time_length": 10,
75
+ "mask_time_min_masks": 2,
76
+ "mask_time_prob": 0.05,
77
+ "max_length": 448,
78
+ "max_source_positions": 1500,
79
+ "max_target_positions": 448,
80
+ "median_filter_width": 7,
81
+ "min_length": 0,
82
+ "model_type": "whisper",
83
+ "no_repeat_ngram_size": 0,
84
+ "num_beam_groups": 1,
85
+ "num_beams": 1,
86
+ "num_hidden_layers": 12,
87
+ "num_mel_bins": 80,
88
+ "num_return_sequences": 1,
89
+ "output_attentions": false,
90
+ "output_hidden_states": false,
91
+ "output_scores": false,
92
+ "pad_token_id": 50257,
93
+ "prefix": null,
94
+ "problem_type": null,
95
+ "pruned_heads": {},
96
+ "remove_invalid_values": false,
97
+ "repetition_penalty": 1.0,
98
+ "return_dict": true,
99
+ "return_dict_in_generate": false,
100
+ "scale_embedding": false,
101
+ "sep_token_id": null,
102
+ "suppress_tokens": [
103
+ 1,
104
+ 2,
105
+ 7,
106
+ 8,
107
+ 9,
108
+ 10,
109
+ 14,
110
+ 25,
111
+ 26,
112
+ 27,
113
+ 28,
114
+ 29,
115
+ 31,
116
+ 58,
117
+ 59,
118
+ 60,
119
+ 61,
120
+ 62,
121
+ 63,
122
+ 90,
123
+ 91,
124
+ 92,
125
+ 93,
126
+ 359,
127
+ 503,
128
+ 522,
129
+ 542,
130
+ 873,
131
+ 893,
132
+ 902,
133
+ 918,
134
+ 922,
135
+ 931,
136
+ 1350,
137
+ 1853,
138
+ 1982,
139
+ 2460,
140
+ 2627,
141
+ 3246,
142
+ 3253,
143
+ 3268,
144
+ 3536,
145
+ 3846,
146
+ 3961,
147
+ 4183,
148
+ 4667,
149
+ 6585,
150
+ 6647,
151
+ 7273,
152
+ 9061,
153
+ 9383,
154
+ 10428,
155
+ 10929,
156
+ 11938,
157
+ 12033,
158
+ 12331,
159
+ 12562,
160
+ 13793,
161
+ 14157,
162
+ 14635,
163
+ 15265,
164
+ 15618,
165
+ 16553,
166
+ 16604,
167
+ 18362,
168
+ 18956,
169
+ 20075,
170
+ 21675,
171
+ 22520,
172
+ 26130,
173
+ 26161,
174
+ 26435,
175
+ 28279,
176
+ 29464,
177
+ 31650,
178
+ 32302,
179
+ 32470,
180
+ 36865,
181
+ 42863,
182
+ 47425,
183
+ 49870,
184
+ 50254,
185
+ 50258,
186
+ 50360,
187
+ 50361,
188
+ 50362
189
+ ],
190
+ "task_specific_params": null,
191
+ "temperature": 1.0,
192
+ "tf_legacy_loss": false,
193
+ "tie_encoder_decoder": false,
194
+ "tie_word_embeddings": true,
195
+ "tokenizer_class": null,
196
+ "top_k": 50,
197
+ "top_p": 1.0,
198
+ "torch_dtype": "bfloat16",
199
+ "torchscript": false,
200
+ "typical_p": 1.0,
201
+ "use_bfloat16": false,
202
+ "use_cache": true,
203
+ "use_weighted_layer_sum": false,
204
+ "vocab_size": 51865
205
+ },
206
+ "audio_select_layer": -2,
207
+ "auto_map": {
208
+ "AutoConfig": "modeling_audio.MM_LLMs_Config",
209
+ "AutoModel": "modeling_audio.MM_LLMs"
210
+ },
211
+ "llm_config": {
212
+ "_name_or_path": "mesolitica/malaysian-mistral-7b-32k-instructions-v4",
213
+ "add_cross_attention": false,
214
+ "architectures": [
215
+ "MistralForCausalLM"
216
+ ],
217
+ "attention_dropout": 0.0,
218
+ "bad_words_ids": null,
219
+ "begin_suppress_tokens": null,
220
+ "bos_token_id": 1,
221
+ "chunk_size_feed_forward": 0,
222
+ "cross_attention_hidden_size": null,
223
+ "decoder_start_token_id": null,
224
+ "diversity_penalty": 0.0,
225
+ "do_sample": false,
226
+ "early_stopping": false,
227
+ "encoder_no_repeat_ngram_size": 0,
228
+ "eos_token_id": 2,
229
+ "exponential_decay_length_penalty": null,
230
+ "finetuning_task": null,
231
+ "forced_bos_token_id": null,
232
+ "forced_eos_token_id": null,
233
+ "hidden_act": "silu",
234
+ "hidden_size": 4096,
235
+ "id2label": {
236
+ "0": "LABEL_0",
237
+ "1": "LABEL_1"
238
+ },
239
+ "initializer_range": 0.02,
240
+ "intermediate_size": 14336,
241
+ "is_decoder": false,
242
+ "is_encoder_decoder": false,
243
+ "label2id": {
244
+ "LABEL_0": 0,
245
+ "LABEL_1": 1
246
+ },
247
+ "length_penalty": 1.0,
248
+ "max_length": 20,
249
+ "max_position_embeddings": 32768,
250
+ "min_length": 0,
251
+ "model_type": "mistral",
252
+ "no_repeat_ngram_size": 0,
253
+ "num_attention_heads": 32,
254
+ "num_beam_groups": 1,
255
+ "num_beams": 1,
256
+ "num_hidden_layers": 32,
257
+ "num_key_value_heads": 8,
258
+ "num_return_sequences": 1,
259
+ "output_attentions": false,
260
+ "output_hidden_states": false,
261
+ "output_scores": false,
262
+ "pad_token_id": null,
263
+ "prefix": null,
264
+ "problem_type": null,
265
+ "pruned_heads": {},
266
+ "remove_invalid_values": false,
267
+ "repetition_penalty": 1.0,
268
+ "return_dict": true,
269
+ "return_dict_in_generate": false,
270
+ "rms_norm_eps": 1e-05,
271
+ "rope_theta": 10000.0,
272
+ "sep_token_id": null,
273
+ "sliding_window": 4096,
274
+ "suppress_tokens": null,
275
+ "task_specific_params": null,
276
+ "temperature": 1.0,
277
+ "tf_legacy_loss": false,
278
+ "tie_encoder_decoder": false,
279
+ "tie_word_embeddings": false,
280
+ "tokenizer_class": null,
281
+ "top_k": 50,
282
+ "top_p": 1.0,
283
+ "torch_dtype": "bfloat16",
284
+ "torchscript": false,
285
+ "typical_p": 1.0,
286
+ "use_bfloat16": false,
287
+ "use_cache": true,
288
+ "vocab_size": 32004
289
+ },
290
+ "model_type": "mm_llms",
291
+ "torch_dtype": "bfloat16",
292
+ "transformers_version": "4.37.2"
293
+ }
generation_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.37.2"
4
+ }
model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d508b9eab757234bd5202f1b64fa0a8999b0e79db0e223c8a4ae825382e651d
3
+ size 4990499904
model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a39b0ffd0f59889e26ca762e9c99850fefe8ef26752b95eb10c3323ec636456a
3
+ size 4999819760
model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3073db90378497fbfe287e614d53b75ceecc9c808aca6eb70953c1fe085fdc5e
3
+ size 4976774552
model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec049e4b1b968ad4735fbcbfb901d9244d53a2d733dfb20a6cc4d2accd764ecf
3
+ size 285229504
model.safetensors.index.json ADDED
@@ -0,0 +1,781 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 15252228608
4
+ },
5
+ "weight_map": {
6
+ "audio_encoder.decoder.embed_positions.weight": "model-00001-of-00004.safetensors",
7
+ "audio_encoder.decoder.embed_tokens.weight": "model-00001-of-00004.safetensors",
8
+ "audio_encoder.decoder.layer_norm.bias": "model-00001-of-00004.safetensors",
9
+ "audio_encoder.decoder.layer_norm.weight": "model-00001-of-00004.safetensors",
10
+ "audio_encoder.decoder.layers.0.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
11
+ "audio_encoder.decoder.layers.0.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
12
+ "audio_encoder.decoder.layers.0.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
13
+ "audio_encoder.decoder.layers.0.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
14
+ "audio_encoder.decoder.layers.0.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
15
+ "audio_encoder.decoder.layers.0.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
16
+ "audio_encoder.decoder.layers.0.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
17
+ "audio_encoder.decoder.layers.0.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
18
+ "audio_encoder.decoder.layers.0.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
19
+ "audio_encoder.decoder.layers.0.fc1.bias": "model-00001-of-00004.safetensors",
20
+ "audio_encoder.decoder.layers.0.fc1.weight": "model-00001-of-00004.safetensors",
21
+ "audio_encoder.decoder.layers.0.fc2.bias": "model-00001-of-00004.safetensors",
22
+ "audio_encoder.decoder.layers.0.fc2.weight": "model-00001-of-00004.safetensors",
23
+ "audio_encoder.decoder.layers.0.final_layer_norm.bias": "model-00001-of-00004.safetensors",
24
+ "audio_encoder.decoder.layers.0.final_layer_norm.weight": "model-00001-of-00004.safetensors",
25
+ "audio_encoder.decoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
26
+ "audio_encoder.decoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
27
+ "audio_encoder.decoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
28
+ "audio_encoder.decoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
29
+ "audio_encoder.decoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
30
+ "audio_encoder.decoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
31
+ "audio_encoder.decoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
32
+ "audio_encoder.decoder.layers.0.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
33
+ "audio_encoder.decoder.layers.0.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
34
+ "audio_encoder.decoder.layers.1.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
35
+ "audio_encoder.decoder.layers.1.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
36
+ "audio_encoder.decoder.layers.1.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
37
+ "audio_encoder.decoder.layers.1.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
38
+ "audio_encoder.decoder.layers.1.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
39
+ "audio_encoder.decoder.layers.1.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
40
+ "audio_encoder.decoder.layers.1.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
41
+ "audio_encoder.decoder.layers.1.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
42
+ "audio_encoder.decoder.layers.1.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
43
+ "audio_encoder.decoder.layers.1.fc1.bias": "model-00001-of-00004.safetensors",
44
+ "audio_encoder.decoder.layers.1.fc1.weight": "model-00001-of-00004.safetensors",
45
+ "audio_encoder.decoder.layers.1.fc2.bias": "model-00001-of-00004.safetensors",
46
+ "audio_encoder.decoder.layers.1.fc2.weight": "model-00001-of-00004.safetensors",
47
+ "audio_encoder.decoder.layers.1.final_layer_norm.bias": "model-00001-of-00004.safetensors",
48
+ "audio_encoder.decoder.layers.1.final_layer_norm.weight": "model-00001-of-00004.safetensors",
49
+ "audio_encoder.decoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
50
+ "audio_encoder.decoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
51
+ "audio_encoder.decoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
52
+ "audio_encoder.decoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
53
+ "audio_encoder.decoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
54
+ "audio_encoder.decoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
55
+ "audio_encoder.decoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
56
+ "audio_encoder.decoder.layers.1.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
57
+ "audio_encoder.decoder.layers.1.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
58
+ "audio_encoder.decoder.layers.10.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
59
+ "audio_encoder.decoder.layers.10.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
60
+ "audio_encoder.decoder.layers.10.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
61
+ "audio_encoder.decoder.layers.10.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
62
+ "audio_encoder.decoder.layers.10.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
63
+ "audio_encoder.decoder.layers.10.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
64
+ "audio_encoder.decoder.layers.10.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
65
+ "audio_encoder.decoder.layers.10.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
66
+ "audio_encoder.decoder.layers.10.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
67
+ "audio_encoder.decoder.layers.10.fc1.bias": "model-00001-of-00004.safetensors",
68
+ "audio_encoder.decoder.layers.10.fc1.weight": "model-00001-of-00004.safetensors",
69
+ "audio_encoder.decoder.layers.10.fc2.bias": "model-00001-of-00004.safetensors",
70
+ "audio_encoder.decoder.layers.10.fc2.weight": "model-00001-of-00004.safetensors",
71
+ "audio_encoder.decoder.layers.10.final_layer_norm.bias": "model-00001-of-00004.safetensors",
72
+ "audio_encoder.decoder.layers.10.final_layer_norm.weight": "model-00001-of-00004.safetensors",
73
+ "audio_encoder.decoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
74
+ "audio_encoder.decoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
75
+ "audio_encoder.decoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
76
+ "audio_encoder.decoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
77
+ "audio_encoder.decoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
78
+ "audio_encoder.decoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
79
+ "audio_encoder.decoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
80
+ "audio_encoder.decoder.layers.10.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
81
+ "audio_encoder.decoder.layers.10.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
82
+ "audio_encoder.decoder.layers.11.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
83
+ "audio_encoder.decoder.layers.11.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
84
+ "audio_encoder.decoder.layers.11.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
85
+ "audio_encoder.decoder.layers.11.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
86
+ "audio_encoder.decoder.layers.11.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
87
+ "audio_encoder.decoder.layers.11.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
88
+ "audio_encoder.decoder.layers.11.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
89
+ "audio_encoder.decoder.layers.11.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
90
+ "audio_encoder.decoder.layers.11.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
91
+ "audio_encoder.decoder.layers.11.fc1.bias": "model-00001-of-00004.safetensors",
92
+ "audio_encoder.decoder.layers.11.fc1.weight": "model-00001-of-00004.safetensors",
93
+ "audio_encoder.decoder.layers.11.fc2.bias": "model-00001-of-00004.safetensors",
94
+ "audio_encoder.decoder.layers.11.fc2.weight": "model-00001-of-00004.safetensors",
95
+ "audio_encoder.decoder.layers.11.final_layer_norm.bias": "model-00001-of-00004.safetensors",
96
+ "audio_encoder.decoder.layers.11.final_layer_norm.weight": "model-00001-of-00004.safetensors",
97
+ "audio_encoder.decoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
98
+ "audio_encoder.decoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
99
+ "audio_encoder.decoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
100
+ "audio_encoder.decoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
101
+ "audio_encoder.decoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
102
+ "audio_encoder.decoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
103
+ "audio_encoder.decoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
104
+ "audio_encoder.decoder.layers.11.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
105
+ "audio_encoder.decoder.layers.11.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
106
+ "audio_encoder.decoder.layers.2.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
107
+ "audio_encoder.decoder.layers.2.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
108
+ "audio_encoder.decoder.layers.2.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
109
+ "audio_encoder.decoder.layers.2.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
110
+ "audio_encoder.decoder.layers.2.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
111
+ "audio_encoder.decoder.layers.2.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
112
+ "audio_encoder.decoder.layers.2.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
113
+ "audio_encoder.decoder.layers.2.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
114
+ "audio_encoder.decoder.layers.2.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
115
+ "audio_encoder.decoder.layers.2.fc1.bias": "model-00001-of-00004.safetensors",
116
+ "audio_encoder.decoder.layers.2.fc1.weight": "model-00001-of-00004.safetensors",
117
+ "audio_encoder.decoder.layers.2.fc2.bias": "model-00001-of-00004.safetensors",
118
+ "audio_encoder.decoder.layers.2.fc2.weight": "model-00001-of-00004.safetensors",
119
+ "audio_encoder.decoder.layers.2.final_layer_norm.bias": "model-00001-of-00004.safetensors",
120
+ "audio_encoder.decoder.layers.2.final_layer_norm.weight": "model-00001-of-00004.safetensors",
121
+ "audio_encoder.decoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
122
+ "audio_encoder.decoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
123
+ "audio_encoder.decoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
124
+ "audio_encoder.decoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
125
+ "audio_encoder.decoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
126
+ "audio_encoder.decoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
127
+ "audio_encoder.decoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
128
+ "audio_encoder.decoder.layers.2.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
129
+ "audio_encoder.decoder.layers.2.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
130
+ "audio_encoder.decoder.layers.3.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
131
+ "audio_encoder.decoder.layers.3.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
132
+ "audio_encoder.decoder.layers.3.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
133
+ "audio_encoder.decoder.layers.3.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
134
+ "audio_encoder.decoder.layers.3.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
135
+ "audio_encoder.decoder.layers.3.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
136
+ "audio_encoder.decoder.layers.3.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
137
+ "audio_encoder.decoder.layers.3.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
138
+ "audio_encoder.decoder.layers.3.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
139
+ "audio_encoder.decoder.layers.3.fc1.bias": "model-00001-of-00004.safetensors",
140
+ "audio_encoder.decoder.layers.3.fc1.weight": "model-00001-of-00004.safetensors",
141
+ "audio_encoder.decoder.layers.3.fc2.bias": "model-00001-of-00004.safetensors",
142
+ "audio_encoder.decoder.layers.3.fc2.weight": "model-00001-of-00004.safetensors",
143
+ "audio_encoder.decoder.layers.3.final_layer_norm.bias": "model-00001-of-00004.safetensors",
144
+ "audio_encoder.decoder.layers.3.final_layer_norm.weight": "model-00001-of-00004.safetensors",
145
+ "audio_encoder.decoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
146
+ "audio_encoder.decoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
147
+ "audio_encoder.decoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
148
+ "audio_encoder.decoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
149
+ "audio_encoder.decoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
150
+ "audio_encoder.decoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
151
+ "audio_encoder.decoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
152
+ "audio_encoder.decoder.layers.3.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
153
+ "audio_encoder.decoder.layers.3.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
154
+ "audio_encoder.decoder.layers.4.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
155
+ "audio_encoder.decoder.layers.4.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
156
+ "audio_encoder.decoder.layers.4.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
157
+ "audio_encoder.decoder.layers.4.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
158
+ "audio_encoder.decoder.layers.4.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
159
+ "audio_encoder.decoder.layers.4.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
160
+ "audio_encoder.decoder.layers.4.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
161
+ "audio_encoder.decoder.layers.4.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
162
+ "audio_encoder.decoder.layers.4.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
163
+ "audio_encoder.decoder.layers.4.fc1.bias": "model-00001-of-00004.safetensors",
164
+ "audio_encoder.decoder.layers.4.fc1.weight": "model-00001-of-00004.safetensors",
165
+ "audio_encoder.decoder.layers.4.fc2.bias": "model-00001-of-00004.safetensors",
166
+ "audio_encoder.decoder.layers.4.fc2.weight": "model-00001-of-00004.safetensors",
167
+ "audio_encoder.decoder.layers.4.final_layer_norm.bias": "model-00001-of-00004.safetensors",
168
+ "audio_encoder.decoder.layers.4.final_layer_norm.weight": "model-00001-of-00004.safetensors",
169
+ "audio_encoder.decoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
170
+ "audio_encoder.decoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
171
+ "audio_encoder.decoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
172
+ "audio_encoder.decoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
173
+ "audio_encoder.decoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
174
+ "audio_encoder.decoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
175
+ "audio_encoder.decoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
176
+ "audio_encoder.decoder.layers.4.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
177
+ "audio_encoder.decoder.layers.4.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
178
+ "audio_encoder.decoder.layers.5.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
179
+ "audio_encoder.decoder.layers.5.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
180
+ "audio_encoder.decoder.layers.5.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
181
+ "audio_encoder.decoder.layers.5.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
182
+ "audio_encoder.decoder.layers.5.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
183
+ "audio_encoder.decoder.layers.5.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
184
+ "audio_encoder.decoder.layers.5.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
185
+ "audio_encoder.decoder.layers.5.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
186
+ "audio_encoder.decoder.layers.5.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
187
+ "audio_encoder.decoder.layers.5.fc1.bias": "model-00001-of-00004.safetensors",
188
+ "audio_encoder.decoder.layers.5.fc1.weight": "model-00001-of-00004.safetensors",
189
+ "audio_encoder.decoder.layers.5.fc2.bias": "model-00001-of-00004.safetensors",
190
+ "audio_encoder.decoder.layers.5.fc2.weight": "model-00001-of-00004.safetensors",
191
+ "audio_encoder.decoder.layers.5.final_layer_norm.bias": "model-00001-of-00004.safetensors",
192
+ "audio_encoder.decoder.layers.5.final_layer_norm.weight": "model-00001-of-00004.safetensors",
193
+ "audio_encoder.decoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
194
+ "audio_encoder.decoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
195
+ "audio_encoder.decoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
196
+ "audio_encoder.decoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
197
+ "audio_encoder.decoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
198
+ "audio_encoder.decoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
199
+ "audio_encoder.decoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
200
+ "audio_encoder.decoder.layers.5.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
201
+ "audio_encoder.decoder.layers.5.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
202
+ "audio_encoder.decoder.layers.6.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
203
+ "audio_encoder.decoder.layers.6.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
204
+ "audio_encoder.decoder.layers.6.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
205
+ "audio_encoder.decoder.layers.6.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
206
+ "audio_encoder.decoder.layers.6.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
207
+ "audio_encoder.decoder.layers.6.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
208
+ "audio_encoder.decoder.layers.6.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
209
+ "audio_encoder.decoder.layers.6.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
210
+ "audio_encoder.decoder.layers.6.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
211
+ "audio_encoder.decoder.layers.6.fc1.bias": "model-00001-of-00004.safetensors",
212
+ "audio_encoder.decoder.layers.6.fc1.weight": "model-00001-of-00004.safetensors",
213
+ "audio_encoder.decoder.layers.6.fc2.bias": "model-00001-of-00004.safetensors",
214
+ "audio_encoder.decoder.layers.6.fc2.weight": "model-00001-of-00004.safetensors",
215
+ "audio_encoder.decoder.layers.6.final_layer_norm.bias": "model-00001-of-00004.safetensors",
216
+ "audio_encoder.decoder.layers.6.final_layer_norm.weight": "model-00001-of-00004.safetensors",
217
+ "audio_encoder.decoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
218
+ "audio_encoder.decoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
219
+ "audio_encoder.decoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
220
+ "audio_encoder.decoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
221
+ "audio_encoder.decoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
222
+ "audio_encoder.decoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
223
+ "audio_encoder.decoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
224
+ "audio_encoder.decoder.layers.6.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
225
+ "audio_encoder.decoder.layers.6.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
226
+ "audio_encoder.decoder.layers.7.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
227
+ "audio_encoder.decoder.layers.7.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
228
+ "audio_encoder.decoder.layers.7.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
229
+ "audio_encoder.decoder.layers.7.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
230
+ "audio_encoder.decoder.layers.7.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
231
+ "audio_encoder.decoder.layers.7.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
232
+ "audio_encoder.decoder.layers.7.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
233
+ "audio_encoder.decoder.layers.7.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
234
+ "audio_encoder.decoder.layers.7.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
235
+ "audio_encoder.decoder.layers.7.fc1.bias": "model-00001-of-00004.safetensors",
236
+ "audio_encoder.decoder.layers.7.fc1.weight": "model-00001-of-00004.safetensors",
237
+ "audio_encoder.decoder.layers.7.fc2.bias": "model-00001-of-00004.safetensors",
238
+ "audio_encoder.decoder.layers.7.fc2.weight": "model-00001-of-00004.safetensors",
239
+ "audio_encoder.decoder.layers.7.final_layer_norm.bias": "model-00001-of-00004.safetensors",
240
+ "audio_encoder.decoder.layers.7.final_layer_norm.weight": "model-00001-of-00004.safetensors",
241
+ "audio_encoder.decoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
242
+ "audio_encoder.decoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
243
+ "audio_encoder.decoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
244
+ "audio_encoder.decoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
245
+ "audio_encoder.decoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
246
+ "audio_encoder.decoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
247
+ "audio_encoder.decoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
248
+ "audio_encoder.decoder.layers.7.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
249
+ "audio_encoder.decoder.layers.7.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
250
+ "audio_encoder.decoder.layers.8.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
251
+ "audio_encoder.decoder.layers.8.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
252
+ "audio_encoder.decoder.layers.8.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
253
+ "audio_encoder.decoder.layers.8.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
254
+ "audio_encoder.decoder.layers.8.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
255
+ "audio_encoder.decoder.layers.8.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
256
+ "audio_encoder.decoder.layers.8.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
257
+ "audio_encoder.decoder.layers.8.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
258
+ "audio_encoder.decoder.layers.8.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
259
+ "audio_encoder.decoder.layers.8.fc1.bias": "model-00001-of-00004.safetensors",
260
+ "audio_encoder.decoder.layers.8.fc1.weight": "model-00001-of-00004.safetensors",
261
+ "audio_encoder.decoder.layers.8.fc2.bias": "model-00001-of-00004.safetensors",
262
+ "audio_encoder.decoder.layers.8.fc2.weight": "model-00001-of-00004.safetensors",
263
+ "audio_encoder.decoder.layers.8.final_layer_norm.bias": "model-00001-of-00004.safetensors",
264
+ "audio_encoder.decoder.layers.8.final_layer_norm.weight": "model-00001-of-00004.safetensors",
265
+ "audio_encoder.decoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
266
+ "audio_encoder.decoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
267
+ "audio_encoder.decoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
268
+ "audio_encoder.decoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
269
+ "audio_encoder.decoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
270
+ "audio_encoder.decoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
271
+ "audio_encoder.decoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
272
+ "audio_encoder.decoder.layers.8.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
273
+ "audio_encoder.decoder.layers.8.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
274
+ "audio_encoder.decoder.layers.9.encoder_attn.k_proj.weight": "model-00001-of-00004.safetensors",
275
+ "audio_encoder.decoder.layers.9.encoder_attn.out_proj.bias": "model-00001-of-00004.safetensors",
276
+ "audio_encoder.decoder.layers.9.encoder_attn.out_proj.weight": "model-00001-of-00004.safetensors",
277
+ "audio_encoder.decoder.layers.9.encoder_attn.q_proj.bias": "model-00001-of-00004.safetensors",
278
+ "audio_encoder.decoder.layers.9.encoder_attn.q_proj.weight": "model-00001-of-00004.safetensors",
279
+ "audio_encoder.decoder.layers.9.encoder_attn.v_proj.bias": "model-00001-of-00004.safetensors",
280
+ "audio_encoder.decoder.layers.9.encoder_attn.v_proj.weight": "model-00001-of-00004.safetensors",
281
+ "audio_encoder.decoder.layers.9.encoder_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
282
+ "audio_encoder.decoder.layers.9.encoder_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
283
+ "audio_encoder.decoder.layers.9.fc1.bias": "model-00001-of-00004.safetensors",
284
+ "audio_encoder.decoder.layers.9.fc1.weight": "model-00001-of-00004.safetensors",
285
+ "audio_encoder.decoder.layers.9.fc2.bias": "model-00001-of-00004.safetensors",
286
+ "audio_encoder.decoder.layers.9.fc2.weight": "model-00001-of-00004.safetensors",
287
+ "audio_encoder.decoder.layers.9.final_layer_norm.bias": "model-00001-of-00004.safetensors",
288
+ "audio_encoder.decoder.layers.9.final_layer_norm.weight": "model-00001-of-00004.safetensors",
289
+ "audio_encoder.decoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
290
+ "audio_encoder.decoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
291
+ "audio_encoder.decoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
292
+ "audio_encoder.decoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
293
+ "audio_encoder.decoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
294
+ "audio_encoder.decoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
295
+ "audio_encoder.decoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
296
+ "audio_encoder.decoder.layers.9.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
297
+ "audio_encoder.decoder.layers.9.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
298
+ "audio_encoder.encoder.conv1.bias": "model-00001-of-00004.safetensors",
299
+ "audio_encoder.encoder.conv1.weight": "model-00001-of-00004.safetensors",
300
+ "audio_encoder.encoder.conv2.bias": "model-00001-of-00004.safetensors",
301
+ "audio_encoder.encoder.conv2.weight": "model-00001-of-00004.safetensors",
302
+ "audio_encoder.encoder.embed_positions.weight": "model-00001-of-00004.safetensors",
303
+ "audio_encoder.encoder.layer_norm.bias": "model-00001-of-00004.safetensors",
304
+ "audio_encoder.encoder.layer_norm.weight": "model-00001-of-00004.safetensors",
305
+ "audio_encoder.encoder.layers.0.fc1.bias": "model-00001-of-00004.safetensors",
306
+ "audio_encoder.encoder.layers.0.fc1.weight": "model-00001-of-00004.safetensors",
307
+ "audio_encoder.encoder.layers.0.fc2.bias": "model-00001-of-00004.safetensors",
308
+ "audio_encoder.encoder.layers.0.fc2.weight": "model-00001-of-00004.safetensors",
309
+ "audio_encoder.encoder.layers.0.final_layer_norm.bias": "model-00001-of-00004.safetensors",
310
+ "audio_encoder.encoder.layers.0.final_layer_norm.weight": "model-00001-of-00004.safetensors",
311
+ "audio_encoder.encoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
312
+ "audio_encoder.encoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
313
+ "audio_encoder.encoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
314
+ "audio_encoder.encoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
315
+ "audio_encoder.encoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
316
+ "audio_encoder.encoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
317
+ "audio_encoder.encoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
318
+ "audio_encoder.encoder.layers.0.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
319
+ "audio_encoder.encoder.layers.0.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
320
+ "audio_encoder.encoder.layers.1.fc1.bias": "model-00001-of-00004.safetensors",
321
+ "audio_encoder.encoder.layers.1.fc1.weight": "model-00001-of-00004.safetensors",
322
+ "audio_encoder.encoder.layers.1.fc2.bias": "model-00001-of-00004.safetensors",
323
+ "audio_encoder.encoder.layers.1.fc2.weight": "model-00001-of-00004.safetensors",
324
+ "audio_encoder.encoder.layers.1.final_layer_norm.bias": "model-00001-of-00004.safetensors",
325
+ "audio_encoder.encoder.layers.1.final_layer_norm.weight": "model-00001-of-00004.safetensors",
326
+ "audio_encoder.encoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
327
+ "audio_encoder.encoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
328
+ "audio_encoder.encoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
329
+ "audio_encoder.encoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
330
+ "audio_encoder.encoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
331
+ "audio_encoder.encoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
332
+ "audio_encoder.encoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
333
+ "audio_encoder.encoder.layers.1.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
334
+ "audio_encoder.encoder.layers.1.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
335
+ "audio_encoder.encoder.layers.10.fc1.bias": "model-00001-of-00004.safetensors",
336
+ "audio_encoder.encoder.layers.10.fc1.weight": "model-00001-of-00004.safetensors",
337
+ "audio_encoder.encoder.layers.10.fc2.bias": "model-00001-of-00004.safetensors",
338
+ "audio_encoder.encoder.layers.10.fc2.weight": "model-00001-of-00004.safetensors",
339
+ "audio_encoder.encoder.layers.10.final_layer_norm.bias": "model-00001-of-00004.safetensors",
340
+ "audio_encoder.encoder.layers.10.final_layer_norm.weight": "model-00001-of-00004.safetensors",
341
+ "audio_encoder.encoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
342
+ "audio_encoder.encoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
343
+ "audio_encoder.encoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
344
+ "audio_encoder.encoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
345
+ "audio_encoder.encoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
346
+ "audio_encoder.encoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
347
+ "audio_encoder.encoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
348
+ "audio_encoder.encoder.layers.10.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
349
+ "audio_encoder.encoder.layers.10.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
350
+ "audio_encoder.encoder.layers.11.fc1.bias": "model-00001-of-00004.safetensors",
351
+ "audio_encoder.encoder.layers.11.fc1.weight": "model-00001-of-00004.safetensors",
352
+ "audio_encoder.encoder.layers.11.fc2.bias": "model-00001-of-00004.safetensors",
353
+ "audio_encoder.encoder.layers.11.fc2.weight": "model-00001-of-00004.safetensors",
354
+ "audio_encoder.encoder.layers.11.final_layer_norm.bias": "model-00001-of-00004.safetensors",
355
+ "audio_encoder.encoder.layers.11.final_layer_norm.weight": "model-00001-of-00004.safetensors",
356
+ "audio_encoder.encoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
357
+ "audio_encoder.encoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
358
+ "audio_encoder.encoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
359
+ "audio_encoder.encoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
360
+ "audio_encoder.encoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
361
+ "audio_encoder.encoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
362
+ "audio_encoder.encoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
363
+ "audio_encoder.encoder.layers.11.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
364
+ "audio_encoder.encoder.layers.11.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
365
+ "audio_encoder.encoder.layers.2.fc1.bias": "model-00001-of-00004.safetensors",
366
+ "audio_encoder.encoder.layers.2.fc1.weight": "model-00001-of-00004.safetensors",
367
+ "audio_encoder.encoder.layers.2.fc2.bias": "model-00001-of-00004.safetensors",
368
+ "audio_encoder.encoder.layers.2.fc2.weight": "model-00001-of-00004.safetensors",
369
+ "audio_encoder.encoder.layers.2.final_layer_norm.bias": "model-00001-of-00004.safetensors",
370
+ "audio_encoder.encoder.layers.2.final_layer_norm.weight": "model-00001-of-00004.safetensors",
371
+ "audio_encoder.encoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
372
+ "audio_encoder.encoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
373
+ "audio_encoder.encoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
374
+ "audio_encoder.encoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
375
+ "audio_encoder.encoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
376
+ "audio_encoder.encoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
377
+ "audio_encoder.encoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
378
+ "audio_encoder.encoder.layers.2.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
379
+ "audio_encoder.encoder.layers.2.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
380
+ "audio_encoder.encoder.layers.3.fc1.bias": "model-00001-of-00004.safetensors",
381
+ "audio_encoder.encoder.layers.3.fc1.weight": "model-00001-of-00004.safetensors",
382
+ "audio_encoder.encoder.layers.3.fc2.bias": "model-00001-of-00004.safetensors",
383
+ "audio_encoder.encoder.layers.3.fc2.weight": "model-00001-of-00004.safetensors",
384
+ "audio_encoder.encoder.layers.3.final_layer_norm.bias": "model-00001-of-00004.safetensors",
385
+ "audio_encoder.encoder.layers.3.final_layer_norm.weight": "model-00001-of-00004.safetensors",
386
+ "audio_encoder.encoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
387
+ "audio_encoder.encoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
388
+ "audio_encoder.encoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
389
+ "audio_encoder.encoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
390
+ "audio_encoder.encoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
391
+ "audio_encoder.encoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
392
+ "audio_encoder.encoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
393
+ "audio_encoder.encoder.layers.3.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
394
+ "audio_encoder.encoder.layers.3.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
395
+ "audio_encoder.encoder.layers.4.fc1.bias": "model-00001-of-00004.safetensors",
396
+ "audio_encoder.encoder.layers.4.fc1.weight": "model-00001-of-00004.safetensors",
397
+ "audio_encoder.encoder.layers.4.fc2.bias": "model-00001-of-00004.safetensors",
398
+ "audio_encoder.encoder.layers.4.fc2.weight": "model-00001-of-00004.safetensors",
399
+ "audio_encoder.encoder.layers.4.final_layer_norm.bias": "model-00001-of-00004.safetensors",
400
+ "audio_encoder.encoder.layers.4.final_layer_norm.weight": "model-00001-of-00004.safetensors",
401
+ "audio_encoder.encoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
402
+ "audio_encoder.encoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
403
+ "audio_encoder.encoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
404
+ "audio_encoder.encoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
405
+ "audio_encoder.encoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
406
+ "audio_encoder.encoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
407
+ "audio_encoder.encoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
408
+ "audio_encoder.encoder.layers.4.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
409
+ "audio_encoder.encoder.layers.4.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
410
+ "audio_encoder.encoder.layers.5.fc1.bias": "model-00001-of-00004.safetensors",
411
+ "audio_encoder.encoder.layers.5.fc1.weight": "model-00001-of-00004.safetensors",
412
+ "audio_encoder.encoder.layers.5.fc2.bias": "model-00001-of-00004.safetensors",
413
+ "audio_encoder.encoder.layers.5.fc2.weight": "model-00001-of-00004.safetensors",
414
+ "audio_encoder.encoder.layers.5.final_layer_norm.bias": "model-00001-of-00004.safetensors",
415
+ "audio_encoder.encoder.layers.5.final_layer_norm.weight": "model-00001-of-00004.safetensors",
416
+ "audio_encoder.encoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
417
+ "audio_encoder.encoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
418
+ "audio_encoder.encoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
419
+ "audio_encoder.encoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
420
+ "audio_encoder.encoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
421
+ "audio_encoder.encoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
422
+ "audio_encoder.encoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
423
+ "audio_encoder.encoder.layers.5.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
424
+ "audio_encoder.encoder.layers.5.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
425
+ "audio_encoder.encoder.layers.6.fc1.bias": "model-00001-of-00004.safetensors",
426
+ "audio_encoder.encoder.layers.6.fc1.weight": "model-00001-of-00004.safetensors",
427
+ "audio_encoder.encoder.layers.6.fc2.bias": "model-00001-of-00004.safetensors",
428
+ "audio_encoder.encoder.layers.6.fc2.weight": "model-00001-of-00004.safetensors",
429
+ "audio_encoder.encoder.layers.6.final_layer_norm.bias": "model-00001-of-00004.safetensors",
430
+ "audio_encoder.encoder.layers.6.final_layer_norm.weight": "model-00001-of-00004.safetensors",
431
+ "audio_encoder.encoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
432
+ "audio_encoder.encoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
433
+ "audio_encoder.encoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
434
+ "audio_encoder.encoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
435
+ "audio_encoder.encoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
436
+ "audio_encoder.encoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
437
+ "audio_encoder.encoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
438
+ "audio_encoder.encoder.layers.6.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
439
+ "audio_encoder.encoder.layers.6.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
440
+ "audio_encoder.encoder.layers.7.fc1.bias": "model-00001-of-00004.safetensors",
441
+ "audio_encoder.encoder.layers.7.fc1.weight": "model-00001-of-00004.safetensors",
442
+ "audio_encoder.encoder.layers.7.fc2.bias": "model-00001-of-00004.safetensors",
443
+ "audio_encoder.encoder.layers.7.fc2.weight": "model-00001-of-00004.safetensors",
444
+ "audio_encoder.encoder.layers.7.final_layer_norm.bias": "model-00001-of-00004.safetensors",
445
+ "audio_encoder.encoder.layers.7.final_layer_norm.weight": "model-00001-of-00004.safetensors",
446
+ "audio_encoder.encoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
447
+ "audio_encoder.encoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
448
+ "audio_encoder.encoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
449
+ "audio_encoder.encoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
450
+ "audio_encoder.encoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
451
+ "audio_encoder.encoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
452
+ "audio_encoder.encoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
453
+ "audio_encoder.encoder.layers.7.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
454
+ "audio_encoder.encoder.layers.7.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
455
+ "audio_encoder.encoder.layers.8.fc1.bias": "model-00001-of-00004.safetensors",
456
+ "audio_encoder.encoder.layers.8.fc1.weight": "model-00001-of-00004.safetensors",
457
+ "audio_encoder.encoder.layers.8.fc2.bias": "model-00001-of-00004.safetensors",
458
+ "audio_encoder.encoder.layers.8.fc2.weight": "model-00001-of-00004.safetensors",
459
+ "audio_encoder.encoder.layers.8.final_layer_norm.bias": "model-00001-of-00004.safetensors",
460
+ "audio_encoder.encoder.layers.8.final_layer_norm.weight": "model-00001-of-00004.safetensors",
461
+ "audio_encoder.encoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
462
+ "audio_encoder.encoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
463
+ "audio_encoder.encoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
464
+ "audio_encoder.encoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
465
+ "audio_encoder.encoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
466
+ "audio_encoder.encoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
467
+ "audio_encoder.encoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
468
+ "audio_encoder.encoder.layers.8.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
469
+ "audio_encoder.encoder.layers.8.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
470
+ "audio_encoder.encoder.layers.9.fc1.bias": "model-00001-of-00004.safetensors",
471
+ "audio_encoder.encoder.layers.9.fc1.weight": "model-00001-of-00004.safetensors",
472
+ "audio_encoder.encoder.layers.9.fc2.bias": "model-00001-of-00004.safetensors",
473
+ "audio_encoder.encoder.layers.9.fc2.weight": "model-00001-of-00004.safetensors",
474
+ "audio_encoder.encoder.layers.9.final_layer_norm.bias": "model-00001-of-00004.safetensors",
475
+ "audio_encoder.encoder.layers.9.final_layer_norm.weight": "model-00001-of-00004.safetensors",
476
+ "audio_encoder.encoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
477
+ "audio_encoder.encoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00004.safetensors",
478
+ "audio_encoder.encoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00004.safetensors",
479
+ "audio_encoder.encoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
480
+ "audio_encoder.encoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
481
+ "audio_encoder.encoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
482
+ "audio_encoder.encoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
483
+ "audio_encoder.encoder.layers.9.self_attn_layer_norm.bias": "model-00001-of-00004.safetensors",
484
+ "audio_encoder.encoder.layers.9.self_attn_layer_norm.weight": "model-00001-of-00004.safetensors",
485
+ "audio_projector.linear_1.bias": "model-00004-of-00004.safetensors",
486
+ "audio_projector.linear_1.weight": "model-00004-of-00004.safetensors",
487
+ "audio_projector.linear_2.bias": "model-00004-of-00004.safetensors",
488
+ "audio_projector.linear_2.weight": "model-00004-of-00004.safetensors",
489
+ "llm.lm_head.weight": "model-00003-of-00004.safetensors",
490
+ "llm.model.embed_tokens.weight": "model-00001-of-00004.safetensors",
491
+ "llm.model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
492
+ "llm.model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
493
+ "llm.model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
494
+ "llm.model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
495
+ "llm.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
496
+ "llm.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
497
+ "llm.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
498
+ "llm.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
499
+ "llm.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
500
+ "llm.model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
501
+ "llm.model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
502
+ "llm.model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
503
+ "llm.model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
504
+ "llm.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
505
+ "llm.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
506
+ "llm.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
507
+ "llm.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
508
+ "llm.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
509
+ "llm.model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
510
+ "llm.model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
511
+ "llm.model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
512
+ "llm.model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
513
+ "llm.model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
514
+ "llm.model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
515
+ "llm.model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
516
+ "llm.model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
517
+ "llm.model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
518
+ "llm.model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
519
+ "llm.model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
520
+ "llm.model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
521
+ "llm.model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
522
+ "llm.model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
523
+ "llm.model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
524
+ "llm.model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
525
+ "llm.model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
526
+ "llm.model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
527
+ "llm.model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
528
+ "llm.model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
529
+ "llm.model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
530
+ "llm.model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
531
+ "llm.model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
532
+ "llm.model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
533
+ "llm.model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
534
+ "llm.model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
535
+ "llm.model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
536
+ "llm.model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
537
+ "llm.model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
538
+ "llm.model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
539
+ "llm.model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
540
+ "llm.model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
541
+ "llm.model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
542
+ "llm.model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
543
+ "llm.model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
544
+ "llm.model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
545
+ "llm.model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
546
+ "llm.model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
547
+ "llm.model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
548
+ "llm.model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
549
+ "llm.model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
550
+ "llm.model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
551
+ "llm.model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
552
+ "llm.model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
553
+ "llm.model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
554
+ "llm.model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
555
+ "llm.model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
556
+ "llm.model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
557
+ "llm.model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
558
+ "llm.model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
559
+ "llm.model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
560
+ "llm.model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
561
+ "llm.model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
562
+ "llm.model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
563
+ "llm.model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
564
+ "llm.model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
565
+ "llm.model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
566
+ "llm.model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
567
+ "llm.model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
568
+ "llm.model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
569
+ "llm.model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
570
+ "llm.model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
571
+ "llm.model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
572
+ "llm.model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
573
+ "llm.model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
574
+ "llm.model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
575
+ "llm.model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
576
+ "llm.model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
577
+ "llm.model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
578
+ "llm.model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
579
+ "llm.model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
580
+ "llm.model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
581
+ "llm.model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors",
582
+ "llm.model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
583
+ "llm.model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
584
+ "llm.model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
585
+ "llm.model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
586
+ "llm.model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
587
+ "llm.model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
588
+ "llm.model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
589
+ "llm.model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
590
+ "llm.model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors",
591
+ "llm.model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
592
+ "llm.model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
593
+ "llm.model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
594
+ "llm.model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
595
+ "llm.model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
596
+ "llm.model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
597
+ "llm.model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
598
+ "llm.model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
599
+ "llm.model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
600
+ "llm.model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
601
+ "llm.model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
602
+ "llm.model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
603
+ "llm.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
604
+ "llm.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
605
+ "llm.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
606
+ "llm.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
607
+ "llm.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
608
+ "llm.model.layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors",
609
+ "llm.model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
610
+ "llm.model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
611
+ "llm.model.layers.20.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
612
+ "llm.model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
613
+ "llm.model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
614
+ "llm.model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
615
+ "llm.model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
616
+ "llm.model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
617
+ "llm.model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
618
+ "llm.model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
619
+ "llm.model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
620
+ "llm.model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
621
+ "llm.model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
622
+ "llm.model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
623
+ "llm.model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
624
+ "llm.model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
625
+ "llm.model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
626
+ "llm.model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
627
+ "llm.model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
628
+ "llm.model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
629
+ "llm.model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
630
+ "llm.model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
631
+ "llm.model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
632
+ "llm.model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
633
+ "llm.model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
634
+ "llm.model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
635
+ "llm.model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
636
+ "llm.model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
637
+ "llm.model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
638
+ "llm.model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
639
+ "llm.model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
640
+ "llm.model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
641
+ "llm.model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
642
+ "llm.model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
643
+ "llm.model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
644
+ "llm.model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
645
+ "llm.model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
646
+ "llm.model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
647
+ "llm.model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
648
+ "llm.model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
649
+ "llm.model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
650
+ "llm.model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
651
+ "llm.model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
652
+ "llm.model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
653
+ "llm.model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
654
+ "llm.model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
655
+ "llm.model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
656
+ "llm.model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
657
+ "llm.model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
658
+ "llm.model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
659
+ "llm.model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
660
+ "llm.model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
661
+ "llm.model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
662
+ "llm.model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
663
+ "llm.model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
664
+ "llm.model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
665
+ "llm.model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
666
+ "llm.model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
667
+ "llm.model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
668
+ "llm.model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
669
+ "llm.model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
670
+ "llm.model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
671
+ "llm.model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
672
+ "llm.model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
673
+ "llm.model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
674
+ "llm.model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
675
+ "llm.model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
676
+ "llm.model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
677
+ "llm.model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
678
+ "llm.model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
679
+ "llm.model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
680
+ "llm.model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors",
681
+ "llm.model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
682
+ "llm.model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
683
+ "llm.model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
684
+ "llm.model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
685
+ "llm.model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
686
+ "llm.model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
687
+ "llm.model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
688
+ "llm.model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
689
+ "llm.model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors",
690
+ "llm.model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
691
+ "llm.model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
692
+ "llm.model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
693
+ "llm.model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
694
+ "llm.model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
695
+ "llm.model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
696
+ "llm.model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
697
+ "llm.model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
698
+ "llm.model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
699
+ "llm.model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
700
+ "llm.model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
701
+ "llm.model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
702
+ "llm.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
703
+ "llm.model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
704
+ "llm.model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
705
+ "llm.model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
706
+ "llm.model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
707
+ "llm.model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors",
708
+ "llm.model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
709
+ "llm.model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
710
+ "llm.model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
711
+ "llm.model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
712
+ "llm.model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
713
+ "llm.model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
714
+ "llm.model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
715
+ "llm.model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
716
+ "llm.model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors",
717
+ "llm.model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
718
+ "llm.model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
719
+ "llm.model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
720
+ "llm.model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
721
+ "llm.model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
722
+ "llm.model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
723
+ "llm.model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
724
+ "llm.model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
725
+ "llm.model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
726
+ "llm.model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
727
+ "llm.model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
728
+ "llm.model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
729
+ "llm.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
730
+ "llm.model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
731
+ "llm.model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
732
+ "llm.model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
733
+ "llm.model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
734
+ "llm.model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
735
+ "llm.model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
736
+ "llm.model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
737
+ "llm.model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
738
+ "llm.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
739
+ "llm.model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
740
+ "llm.model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
741
+ "llm.model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
742
+ "llm.model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
743
+ "llm.model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
744
+ "llm.model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
745
+ "llm.model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
746
+ "llm.model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
747
+ "llm.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
748
+ "llm.model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
749
+ "llm.model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
750
+ "llm.model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
751
+ "llm.model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
752
+ "llm.model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
753
+ "llm.model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
754
+ "llm.model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
755
+ "llm.model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
756
+ "llm.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
757
+ "llm.model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
758
+ "llm.model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
759
+ "llm.model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
760
+ "llm.model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
761
+ "llm.model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors",
762
+ "llm.model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
763
+ "llm.model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
764
+ "llm.model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
765
+ "llm.model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
766
+ "llm.model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
767
+ "llm.model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
768
+ "llm.model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
769
+ "llm.model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
770
+ "llm.model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
771
+ "llm.model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
772
+ "llm.model.layers.9.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
773
+ "llm.model.layers.9.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
774
+ "llm.model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
775
+ "llm.model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
776
+ "llm.model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
777
+ "llm.model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
778
+ "llm.model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
779
+ "llm.model.norm.weight": "model-00003-of-00004.safetensors"
780
+ }
781
+ }
modeling_audio.py ADDED
@@ -0,0 +1,261 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from collections import Counter, defaultdict
2
+ import numpy as np
3
+ import torch
4
+ import torch.nn.functional as F
5
+ from torch import Tensor
6
+ from torch import nn
7
+ from torch.nn import CrossEntropyLoss
8
+ import copy
9
+ import math
10
+ from transformers.activations import gelu
11
+ from typing import List, Optional, Tuple, Union
12
+ from transformers.modeling_utils import PreTrainedModel, PretrainedConfig
13
+ from transformers import CONFIG_MAPPING
14
+ from transformers.modeling_outputs import BaseModelOutput
15
+ from transformers import GenerationConfig
16
+ from transformers import CLIPConfig, CLIPProcessor, CLIPModel, AutoModel
17
+ from transformers import WhisperConfig, WhisperPreTrainedModel, WhisperModel
18
+ from transformers import AutoConfig, AutoModelForCausalLM, LlamaConfig
19
+
20
+
21
+ def most_frequent_element(tensor):
22
+ flattened_list = tensor.flatten().tolist()
23
+ counter = Counter(flattened_list)
24
+ most_common_element = counter.most_common(1)[0][1]
25
+
26
+ return most_common_element
27
+
28
+
29
+ class MM_LLMs_Config(PretrainedConfig):
30
+ model_type = 'mm_llms'
31
+ is_composition = True
32
+
33
+ def __init__(
34
+ self,
35
+ audio_config=None,
36
+ llm_config=None,
37
+ audio_select_layer=-2,
38
+ **kwargs
39
+ ):
40
+
41
+ self.audio_config = audio_config
42
+ self.llm_config = llm_config
43
+ self.audio_select_layer = audio_select_layer
44
+
45
+ if isinstance(self.audio_config, dict):
46
+ audio_config["model_type"] = (
47
+ audio_config["model_type"] if "model_type" in audio_config else "whisper"
48
+ )
49
+ self.audio_config = CONFIG_MAPPING[audio_config["model_type"]](**audio_config)
50
+ if isinstance(self.llm_config, dict):
51
+ llm_config["model_type"] = llm_config["model_type"] if "model_type" in llm_config else "llama"
52
+ self.llm_config = CONFIG_MAPPING[llm_config["model_type"]](**llm_config)
53
+
54
+ super().__init__(**kwargs)
55
+
56
+
57
+ class LlavaMultiModalProjector(nn.Module):
58
+ def __init__(self, in_hidden_size, out_hidden_size, conv_kernel=None, conv_stride=3):
59
+ super().__init__()
60
+
61
+ self.conv_kernel = conv_kernel
62
+
63
+ if conv_kernel:
64
+ self.linear_1 = nn.Conv1d(
65
+ in_hidden_size,
66
+ out_hidden_size,
67
+ kernel_size=conv_kernel,
68
+ stride=conv_stride)
69
+ else:
70
+ self.linear_1 = nn.Linear(
71
+ in_hidden_size,
72
+ out_hidden_size,
73
+ bias=True,
74
+ )
75
+ self.act = gelu
76
+ self.linear_2 = nn.Linear(
77
+ out_hidden_size,
78
+ out_hidden_size,
79
+ bias=True)
80
+
81
+ def forward(self, image_features):
82
+ hidden_states = self.linear_1(image_features)
83
+ if self.conv_kernel:
84
+ hidden_states = hidden_states.transpose(1, 2).contiguous()
85
+ hidden_states = self.act(hidden_states)
86
+ hidden_states = self.linear_2(hidden_states)
87
+ return hidden_states
88
+
89
+
90
+ class MM_LLMs(PreTrainedModel):
91
+ config_class = MM_LLMs_Config
92
+ supports_gradient_checkpointing = True
93
+ _supports_flash_attn_2 = True
94
+
95
+ def __init__(self, config, flash_attention=False, dtype=torch.float32):
96
+ super().__init__(config)
97
+ self.config = config
98
+
99
+ self.audio_encoder = AutoModel.from_config(config.audio_config)
100
+
101
+ self.llm = AutoModelForCausalLM.from_config(
102
+ config.llm_config,
103
+ use_flash_attention_2=flash_attention,
104
+ torch_dtype=dtype,
105
+ )
106
+
107
+ self.audio_projector = LlavaMultiModalProjector(
108
+ config.audio_config.d_model,
109
+ config.llm_config.hidden_size,
110
+ conv_kernel=40,
111
+ conv_stride=3,
112
+ )
113
+
114
+ def forward(self,
115
+ input_ids: torch.LongTensor = None,
116
+ image_index: torch.LongTensor = None,
117
+ audio_index: torch.LongTensor = None,
118
+ image_starts: torch.int = None,
119
+ image_ends: torch.int = None,
120
+ audio_starts: torch.int = None,
121
+ audio_ends: torch.int = None,
122
+ images: torch.FloatTensor = None,
123
+ audios: torch.FloatTensor = None,
124
+ attention_mask: Optional[torch.Tensor] = None,
125
+ position_ids: Optional[torch.LongTensor] = None,
126
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
127
+ inputs_embeds: Optional[torch.FloatTensor] = None,
128
+ labels: Optional[torch.LongTensor] = None,
129
+ output_attentions: Optional[bool] = None,
130
+ output_hidden_states: Optional[bool] = None,
131
+ use_cache: Optional[bool] = None,
132
+ return_dict: Optional[bool] = None, **kwargs):
133
+
134
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
135
+
136
+ audios = audios.type(self.audio_encoder.dtype) if audios is not None else None
137
+
138
+ model_inputs = self.prepare_inputs_for_generation(
139
+ input_ids=input_ids,
140
+ image_index=image_index,
141
+ audio_index=audio_index,
142
+ image_starts=image_starts,
143
+ image_ends=image_ends,
144
+ audio_starts=audio_starts,
145
+ audio_ends=audio_ends,
146
+ images=images,
147
+ audios=audios,
148
+ attention_mask=attention_mask,
149
+ labels=labels)
150
+
151
+ print(input_ids.shape, model_inputs['inputs_embeds'].shape)
152
+ outputs = self.llm(
153
+ inputs_embeds=model_inputs['inputs_embeds'],
154
+ attention_mask=model_inputs['attention_mask'],
155
+ labels=model_inputs['labels'],
156
+ return_dict=return_dict)
157
+
158
+ return outputs
159
+
160
+ def prepare_inputs_for_generation(
161
+ self,
162
+ input_ids,
163
+ past_key_values=None,
164
+ inputs_embeds=None,
165
+ images=None,
166
+ audios=None,
167
+ audio_starts=None,
168
+ audio_ends=None,
169
+ image_starts=None,
170
+ image_ends=None,
171
+ attention_mask=None,
172
+ labels=None,
173
+ audio_index=None,
174
+ image_index=None,
175
+ **kwargs):
176
+
177
+ audio_features = self.encode_audio(
178
+ audios) if audios is not None else None
179
+ embed_tokens = self.llm.model.embed_tokens
180
+ text_embeddings = embed_tokens(input_ids)
181
+ batch_size = text_embeddings.shape[0]
182
+ seq_len = text_embeddings.shape[1]
183
+ embed_dim = text_embeddings.shape[2]
184
+
185
+ if len(audio_index):
186
+ max_count_audio = most_frequent_element(audio_index)
187
+ else:
188
+ max_count_audio = 0
189
+
190
+ if audio_features is not None:
191
+ seq_audio = audio_features.shape[1]
192
+ else:
193
+ seq_audio = 0
194
+
195
+ audio_len = seq_audio * max_count_audio
196
+
197
+ new_len = text_embeddings.shape[1] + audio_len
198
+ final_embedding = torch.zeros(
199
+ batch_size, new_len, embed_dim,
200
+ device=text_embeddings.device,
201
+ dtype=text_embeddings.dtype
202
+ )
203
+ final_embedding[:, :seq_len] = text_embeddings
204
+ final_attention_mask = torch.zeros(
205
+ batch_size, new_len,
206
+ device=attention_mask.device,
207
+ dtype=attention_mask.dtype
208
+ )
209
+ final_attention_mask[:, :seq_len] = attention_mask
210
+ if labels is not None:
211
+ final_labels = torch.full(
212
+ (batch_size, new_len),
213
+ -100,
214
+ device=labels.device,
215
+ dtype=labels.dtype
216
+ )
217
+ final_labels[:, :seq_len] = labels
218
+ else:
219
+ final_labels = None
220
+
221
+ audio_id = int(audio_starts[0])
222
+
223
+ where_is = torch.where(input_ids == audio_id)
224
+ positions = defaultdict(int)
225
+ b_audio = 0
226
+
227
+ for i in range(len(where_is[0])):
228
+ b, k = where_is[0][i], where_is[1][i]
229
+ int_b = int(b)
230
+ int_k = int(k)
231
+ f = audio_features[b_audio]
232
+ b_audio += 1
233
+
234
+ c = torch.cat([final_embedding[b, :int_k + 1 + positions[int_b]],
235
+ f, text_embeddings[b, k + 1:]])
236
+ final_embedding[b, :len(c)] = c
237
+ final_attention_mask[b, :len(c)] = 1.0
238
+
239
+ if labels is not None:
240
+ ignore = torch.tensor([-100] * len(f), device=labels.device)
241
+ c_label = torch.cat(
242
+ [final_labels[b, :int_k + 1 + positions[int_b]], ignore, labels[b, k + 1:]])
243
+ final_labels[b, :len(c)] = c_label
244
+
245
+ positions[int_b] += len(f)
246
+
247
+ model_inputs = {
248
+ "input_ids": input_ids,
249
+ "inputs_embeds": final_embedding,
250
+ "use_cache": kwargs.get("use_cache"),
251
+ "attention_mask": final_attention_mask,
252
+ "labels": final_labels,
253
+ }
254
+ return model_inputs
255
+
256
+ def encode_audio(self, audios):
257
+
258
+ encoded = self.audio_encoder.encoder(audios, output_hidden_states=True)
259
+ encoded = encoded.hidden_states[self.config.audio_select_layer]
260
+ audio_features = self.audio_projector(encoded.transpose(1, 2).contiguous())
261
+ return audio_features