Hugging Face's logo

kurosekurose
/
wav2vec2-base-960h-EMOPIA

Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
kurosekurose commited on
Commit
38d18fc
•
1 Parent(s): caef9a2

End of training

Browse files
Files changed (6) hide show
  1. README.md +77 -0
  2. config.json +125 -0
  3. model.safetensors +3 -0
  4. preprocessor_config.json +9 -0
  5. trainer_state.json +282 -0
  6. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: facebook/wav2vec2-base-960h
4
+ tags:
5
+ - generated_from_trainer
6
+ metrics:
7
+ - accuracy
8
+ model-index:
9
+ - name: wav2vec2-base-960h-EMOPIA
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # wav2vec2-base-960h-EMOPIA
17
+
18
+ This model is a fine-tuned version of [facebook/wav2vec2-base-960h](https://huggingface.co/facebook/wav2vec2-base-960h) on an unknown dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.9295
21
+ - Accuracy: 0.7183
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 1e-05
41
+ - train_batch_size: 1
42
+ - eval_batch_size: 1
43
+ - seed: 42
44
+ - gradient_accumulation_steps: 3
45
+ - total_train_batch_size: 3
46
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
+ - lr_scheduler_type: linear
48
+ - num_epochs: 15
49
+ - mixed_precision_training: Native AMP
50
+
51
+ ### Training results
52
+
53
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
54
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
55
+ | 1.2837 | 1.0 | 269 | 1.0678 | 0.4366 |
56
+ | 1.036 | 2.0 | 538 | 1.1445 | 0.3803 |
57
+ | 0.9878 | 3.0 | 807 | 1.0701 | 0.3944 |
58
+ | 0.9453 | 4.0 | 1076 | 1.1832 | 0.4225 |
59
+ | 0.9144 | 5.0 | 1345 | 1.2068 | 0.3944 |
60
+ | 0.8847 | 6.0 | 1614 | 1.0656 | 0.4648 |
61
+ | 0.8561 | 7.0 | 1883 | 1.1535 | 0.4930 |
62
+ | 0.8303 | 8.0 | 2152 | 1.0277 | 0.5634 |
63
+ | 0.8153 | 9.0 | 2421 | 1.0782 | 0.6056 |
64
+ | 0.7697 | 10.0 | 2690 | 0.8603 | 0.7042 |
65
+ | 0.7142 | 11.0 | 2959 | 0.9536 | 0.6761 |
66
+ | 0.7283 | 12.0 | 3228 | 0.8901 | 0.7183 |
67
+ | 0.7062 | 13.0 | 3497 | 0.9479 | 0.7183 |
68
+ | 0.655 | 14.0 | 3766 | 0.9102 | 0.7324 |
69
+ | 0.6567 | 15.0 | 4035 | 0.9295 | 0.7183 |
70
+
71
+
72
+ ### Framework versions
73
+
74
+ - Transformers 4.42.3
75
+ - Pytorch 2.3.1+cu118
76
+ - Datasets 2.20.0
77
+ - Tokenizers 0.19.1
config.json ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base-960h",
3
+ "activation_dropout": 0.1,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSpeechClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": false,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "sum",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": false,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "group",
53
+ "feat_proj_dropout": 0.1,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.1,
56
+ "finetuning_task": "facebook/wav2vec2-base-960h",
57
+ "gradient_checkpointing": false,
58
+ "hidden_act": "gelu",
59
+ "hidden_dropout": 0.1,
60
+ "hidden_dropout_prob": 0.1,
61
+ "hidden_size": 768,
62
+ "id2label": {
63
+ "0": "Q1",
64
+ "1": "Q2",
65
+ "2": "Q3",
66
+ "3": "Q4"
67
+ },
68
+ "initializer_range": 0.02,
69
+ "intermediate_size": 3072,
70
+ "label2id": {
71
+ "Q1": 0,
72
+ "Q2": 1,
73
+ "Q3": 2,
74
+ "Q4": 3
75
+ },
76
+ "layer_norm_eps": 1e-05,
77
+ "layerdrop": 0.1,
78
+ "mask_feature_length": 10,
79
+ "mask_feature_min_masks": 0,
80
+ "mask_feature_prob": 0.0,
81
+ "mask_time_length": 10,
82
+ "mask_time_min_masks": 2,
83
+ "mask_time_prob": 0.05,
84
+ "model_type": "wav2vec2",
85
+ "num_adapter_layers": 3,
86
+ "num_attention_heads": 12,
87
+ "num_codevector_groups": 2,
88
+ "num_codevectors_per_group": 320,
89
+ "num_conv_pos_embedding_groups": 16,
90
+ "num_conv_pos_embeddings": 128,
91
+ "num_feat_extract_layers": 7,
92
+ "num_hidden_layers": 12,
93
+ "num_negatives": 100,
94
+ "output_hidden_size": 768,
95
+ "pad_token_id": 0,
96
+ "pooling_mode": "mean",
97
+ "problem_type": "single_label_classification",
98
+ "proj_codevector_dim": 256,
99
+ "tdnn_dilation": [
100
+ 1,
101
+ 2,
102
+ 3,
103
+ 1,
104
+ 1
105
+ ],
106
+ "tdnn_dim": [
107
+ 512,
108
+ 512,
109
+ 512,
110
+ 512,
111
+ 1500
112
+ ],
113
+ "tdnn_kernel": [
114
+ 5,
115
+ 3,
116
+ 3,
117
+ 1,
118
+ 1
119
+ ],
120
+ "torch_dtype": "float32",
121
+ "transformers_version": "4.42.3",
122
+ "use_weighted_layer_sum": false,
123
+ "vocab_size": 32,
124
+ "xvector_output_dim": 512
125
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a737810b83b3f0a64d57296427d0d8f0c30ba0732837d6d8c84b02d2d10db6a
3
+ size 379887584
preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,282 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 15.0,
5
+ "eval_steps": 500,
6
+ "global_step": 4035,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "grad_norm": 6.223687171936035,
14
+ "learning_rate": 9.335811648079307e-06,
15
+ "loss": 1.2837,
16
+ "step": 269
17
+ },
18
+ {
19
+ "epoch": 1.0,
20
+ "eval_accuracy": 0.43661972880363464,
21
+ "eval_loss": 1.0678160190582275,
22
+ "eval_runtime": 31.5966,
23
+ "eval_samples_per_second": 2.247,
24
+ "eval_steps_per_second": 2.247,
25
+ "step": 269
26
+ },
27
+ {
28
+ "epoch": 2.0,
29
+ "grad_norm": 4.506842136383057,
30
+ "learning_rate": 8.671623296158614e-06,
31
+ "loss": 1.036,
32
+ "step": 538
33
+ },
34
+ {
35
+ "epoch": 2.0,
36
+ "eval_accuracy": 0.3802816867828369,
37
+ "eval_loss": 1.1444693803787231,
38
+ "eval_runtime": 31.9179,
39
+ "eval_samples_per_second": 2.224,
40
+ "eval_steps_per_second": 2.224,
41
+ "step": 538
42
+ },
43
+ {
44
+ "epoch": 3.0,
45
+ "grad_norm": 3.0099592208862305,
46
+ "learning_rate": 8.007434944237918e-06,
47
+ "loss": 0.9878,
48
+ "step": 807
49
+ },
50
+ {
51
+ "epoch": 3.0,
52
+ "eval_accuracy": 0.39436620473861694,
53
+ "eval_loss": 1.0700511932373047,
54
+ "eval_runtime": 31.7503,
55
+ "eval_samples_per_second": 2.236,
56
+ "eval_steps_per_second": 2.236,
57
+ "step": 807
58
+ },
59
+ {
60
+ "epoch": 4.0,
61
+ "grad_norm": 8.774005889892578,
62
+ "learning_rate": 7.343246592317225e-06,
63
+ "loss": 0.9453,
64
+ "step": 1076
65
+ },
66
+ {
67
+ "epoch": 4.0,
68
+ "eval_accuracy": 0.4225352108478546,
69
+ "eval_loss": 1.1832224130630493,
70
+ "eval_runtime": 31.6924,
71
+ "eval_samples_per_second": 2.24,
72
+ "eval_steps_per_second": 2.24,
73
+ "step": 1076
74
+ },
75
+ {
76
+ "epoch": 5.0,
77
+ "grad_norm": 3.969104290008545,
78
+ "learning_rate": 6.676579925650558e-06,
79
+ "loss": 0.9144,
80
+ "step": 1345
81
+ },
82
+ {
83
+ "epoch": 5.0,
84
+ "eval_accuracy": 0.39436620473861694,
85
+ "eval_loss": 1.2068214416503906,
86
+ "eval_runtime": 31.6897,
87
+ "eval_samples_per_second": 2.24,
88
+ "eval_steps_per_second": 2.24,
89
+ "step": 1345
90
+ },
91
+ {
92
+ "epoch": 6.0,
93
+ "grad_norm": 3.4965126514434814,
94
+ "learning_rate": 6.009913258983892e-06,
95
+ "loss": 0.8847,
96
+ "step": 1614
97
+ },
98
+ {
99
+ "epoch": 6.0,
100
+ "eval_accuracy": 0.4647887349128723,
101
+ "eval_loss": 1.0656119585037231,
102
+ "eval_runtime": 31.6534,
103
+ "eval_samples_per_second": 2.243,
104
+ "eval_steps_per_second": 2.243,
105
+ "step": 1614
106
+ },
107
+ {
108
+ "epoch": 7.0,
109
+ "grad_norm": 38.132389068603516,
110
+ "learning_rate": 5.343246592317225e-06,
111
+ "loss": 0.8561,
112
+ "step": 1883
113
+ },
114
+ {
115
+ "epoch": 7.0,
116
+ "eval_accuracy": 0.49295774102211,
117
+ "eval_loss": 1.1534854173660278,
118
+ "eval_runtime": 31.8557,
119
+ "eval_samples_per_second": 2.229,
120
+ "eval_steps_per_second": 2.229,
121
+ "step": 1883
122
+ },
123
+ {
124
+ "epoch": 8.0,
125
+ "grad_norm": 75.70648956298828,
126
+ "learning_rate": 4.676579925650558e-06,
127
+ "loss": 0.8303,
128
+ "step": 2152
129
+ },
130
+ {
131
+ "epoch": 8.0,
132
+ "eval_accuracy": 0.5633803009986877,
133
+ "eval_loss": 1.0276618003845215,
134
+ "eval_runtime": 31.9505,
135
+ "eval_samples_per_second": 2.222,
136
+ "eval_steps_per_second": 2.222,
137
+ "step": 2152
138
+ },
139
+ {
140
+ "epoch": 9.0,
141
+ "grad_norm": 14.245861053466797,
142
+ "learning_rate": 4.009913258983891e-06,
143
+ "loss": 0.8153,
144
+ "step": 2421
145
+ },
146
+ {
147
+ "epoch": 9.0,
148
+ "eval_accuracy": 0.6056337952613831,
149
+ "eval_loss": 1.0781611204147339,
150
+ "eval_runtime": 31.8449,
151
+ "eval_samples_per_second": 2.23,
152
+ "eval_steps_per_second": 2.23,
153
+ "step": 2421
154
+ },
155
+ {
156
+ "epoch": 10.0,
157
+ "grad_norm": 12.339863777160645,
158
+ "learning_rate": 3.343246592317225e-06,
159
+ "loss": 0.7697,
160
+ "step": 2690
161
+ },
162
+ {
163
+ "epoch": 10.0,
164
+ "eval_accuracy": 0.7042253613471985,
165
+ "eval_loss": 0.8602883815765381,
166
+ "eval_runtime": 31.84,
167
+ "eval_samples_per_second": 2.23,
168
+ "eval_steps_per_second": 2.23,
169
+ "step": 2690
170
+ },
171
+ {
172
+ "epoch": 11.0,
173
+ "grad_norm": 37.19743728637695,
174
+ "learning_rate": 2.6765799256505576e-06,
175
+ "loss": 0.7142,
176
+ "step": 2959
177
+ },
178
+ {
179
+ "epoch": 11.0,
180
+ "eval_accuracy": 0.6760563254356384,
181
+ "eval_loss": 0.9535621404647827,
182
+ "eval_runtime": 32.0342,
183
+ "eval_samples_per_second": 2.216,
184
+ "eval_steps_per_second": 2.216,
185
+ "step": 2959
186
+ },
187
+ {
188
+ "epoch": 12.0,
189
+ "grad_norm": 11.80981731414795,
190
+ "learning_rate": 2.009913258983891e-06,
191
+ "loss": 0.7283,
192
+ "step": 3228
193
+ },
194
+ {
195
+ "epoch": 12.0,
196
+ "eval_accuracy": 0.7183098793029785,
197
+ "eval_loss": 0.8900589346885681,
198
+ "eval_runtime": 31.9143,
199
+ "eval_samples_per_second": 2.225,
200
+ "eval_steps_per_second": 2.225,
201
+ "step": 3228
202
+ },
203
+ {
204
+ "epoch": 13.0,
205
+ "grad_norm": 13.319975852966309,
206
+ "learning_rate": 1.3432465923172244e-06,
207
+ "loss": 0.7062,
208
+ "step": 3497
209
+ },
210
+ {
211
+ "epoch": 13.0,
212
+ "eval_accuracy": 0.7183098793029785,
213
+ "eval_loss": 0.9478991627693176,
214
+ "eval_runtime": 31.814,
215
+ "eval_samples_per_second": 2.232,
216
+ "eval_steps_per_second": 2.232,
217
+ "step": 3497
218
+ },
219
+ {
220
+ "epoch": 14.0,
221
+ "grad_norm": 25.11409568786621,
222
+ "learning_rate": 6.765799256505578e-07,
223
+ "loss": 0.655,
224
+ "step": 3766
225
+ },
226
+ {
227
+ "epoch": 14.0,
228
+ "eval_accuracy": 0.7323943376541138,
229
+ "eval_loss": 0.9101635813713074,
230
+ "eval_runtime": 32.0529,
231
+ "eval_samples_per_second": 2.215,
232
+ "eval_steps_per_second": 2.215,
233
+ "step": 3766
234
+ },
235
+ {
236
+ "epoch": 15.0,
237
+ "grad_norm": 63.15331268310547,
238
+ "learning_rate": 1.2391573729863694e-08,
239
+ "loss": 0.6567,
240
+ "step": 4035
241
+ },
242
+ {
243
+ "epoch": 15.0,
244
+ "eval_accuracy": 0.7183098793029785,
245
+ "eval_loss": 0.9294992089271545,
246
+ "eval_runtime": 32.0448,
247
+ "eval_samples_per_second": 2.216,
248
+ "eval_steps_per_second": 2.216,
249
+ "step": 4035
250
+ },
251
+ {
252
+ "epoch": 15.0,
253
+ "step": 4035,
254
+ "total_flos": 7.1272809219168e+18,
255
+ "train_loss": 0.8522506052085753,
256
+ "train_runtime": 9001.0226,
257
+ "train_samples_per_second": 1.345,
258
+ "train_steps_per_second": 0.448
259
+ }
260
+ ],
261
+ "logging_steps": 8,
262
+ "max_steps": 4035,
263
+ "num_input_tokens_seen": 0,
264
+ "num_train_epochs": 15,
265
+ "save_steps": 500,
266
+ "stateful_callbacks": {
267
+ "TrainerControl": {
268
+ "args": {
269
+ "should_epoch_stop": false,
270
+ "should_evaluate": false,
271
+ "should_log": false,
272
+ "should_save": false,
273
+ "should_training_stop": false
274
+ },
275
+ "attributes": {}
276
+ }
277
+ },
278
+ "total_flos": 7.1272809219168e+18,
279
+ "train_batch_size": 1,
280
+ "trial_name": null,
281
+ "trial_params": null
282
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:65cb90b65180f45b15fc7618832b929076d5156094ce81a25f4b74306463160d
3
+ size 5176