Hugging Face's logo

d190305
/
speecht5_finetuned_voxpopuli_pl

Text-to-Audio
Transformers
TensorBoard
Safetensors
speecht5
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
d190305 commited on
Commit
ef4deea
·
verified ·
1 Parent(s): 4576d6b

Upload 53 files

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. checkpoint-1000/added_tokens.json +4 -0
  2. checkpoint-1000/config.json +92 -0
  3. checkpoint-1000/generation_config.json +9 -0
  4. checkpoint-1000/model.safetensors +3 -0
  5. checkpoint-1000/optimizer.pt +3 -0
  6. checkpoint-1000/preprocessor_config.json +19 -0
  7. checkpoint-1000/rng_state.pth +3 -0
  8. checkpoint-1000/scheduler.pt +3 -0
  9. checkpoint-1000/special_tokens_map.json +13 -0
  10. checkpoint-1000/spm_char.model +3 -0
  11. checkpoint-1000/tokenizer_config.json +63 -0
  12. checkpoint-1000/trainer_state.json +321 -0
  13. checkpoint-1000/training_args.bin +3 -0
  14. checkpoint-2000/added_tokens.json +4 -0
  15. checkpoint-2000/config.json +92 -0
  16. checkpoint-2000/generation_config.json +9 -0
  17. checkpoint-2000/model.safetensors +3 -0
  18. checkpoint-2000/optimizer.pt +3 -0
  19. checkpoint-2000/preprocessor_config.json +19 -0
  20. checkpoint-2000/rng_state.pth +3 -0
  21. checkpoint-2000/scheduler.pt +3 -0
  22. checkpoint-2000/special_tokens_map.json +13 -0
  23. checkpoint-2000/spm_char.model +3 -0
  24. checkpoint-2000/tokenizer_config.json +63 -0
  25. checkpoint-2000/trainer_state.json +609 -0
  26. checkpoint-2000/training_args.bin +3 -0
  27. checkpoint-3000/added_tokens.json +4 -0
  28. checkpoint-3000/config.json +92 -0
  29. checkpoint-3000/generation_config.json +9 -0
  30. checkpoint-3000/model.safetensors +3 -0
  31. checkpoint-3000/optimizer.pt +3 -0
  32. checkpoint-3000/preprocessor_config.json +19 -0
  33. checkpoint-3000/rng_state.pth +3 -0
  34. checkpoint-3000/scheduler.pt +3 -0
  35. checkpoint-3000/special_tokens_map.json +13 -0
  36. checkpoint-3000/spm_char.model +3 -0
  37. checkpoint-3000/tokenizer_config.json +63 -0
  38. checkpoint-3000/trainer_state.json +897 -0
  39. checkpoint-3000/training_args.bin +3 -0
  40. checkpoint-4000/added_tokens.json +4 -0
  41. checkpoint-4000/config.json +92 -0
  42. checkpoint-4000/generation_config.json +9 -0
  43. checkpoint-4000/model.safetensors +3 -0
  44. checkpoint-4000/optimizer.pt +3 -0
  45. checkpoint-4000/preprocessor_config.json +19 -0
  46. checkpoint-4000/rng_state.pth +3 -0
  47. checkpoint-4000/scheduler.pt +3 -0
  48. checkpoint-4000/special_tokens_map.json +13 -0
  49. checkpoint-4000/spm_char.model +3 -0
  50. checkpoint-4000/tokenizer_config.json +63 -0
checkpoint-1000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-1000/config.json ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/speecht5_tts",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "SpeechT5ForTextToSpeech"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 0,
10
+ "conv_bias": false,
11
+ "conv_dim": [
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "conv_kernel": [
21
+ 10,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 2,
27
+ 2
28
+ ],
29
+ "conv_stride": [
30
+ 5,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2
37
+ ],
38
+ "decoder_attention_heads": 12,
39
+ "decoder_ffn_dim": 3072,
40
+ "decoder_layerdrop": 0.1,
41
+ "decoder_layers": 6,
42
+ "decoder_start_token_id": 2,
43
+ "encoder_attention_heads": 12,
44
+ "encoder_ffn_dim": 3072,
45
+ "encoder_layerdrop": 0.1,
46
+ "encoder_layers": 12,
47
+ "encoder_max_relative_position": 160,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_norm": "group",
51
+ "feat_proj_dropout": 0.0,
52
+ "guided_attention_loss_num_heads": 2,
53
+ "guided_attention_loss_scale": 10.0,
54
+ "guided_attention_loss_sigma": 0.4,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 768,
58
+ "initializer_range": 0.02,
59
+ "is_encoder_decoder": true,
60
+ "layer_norm_eps": 1e-05,
61
+ "mask_feature_length": 10,
62
+ "mask_feature_min_masks": 0,
63
+ "mask_feature_prob": 0.0,
64
+ "mask_time_length": 10,
65
+ "mask_time_min_masks": 2,
66
+ "mask_time_prob": 0.05,
67
+ "max_length": null,
68
+ "max_speech_positions": 1876,
69
+ "max_text_positions": 600,
70
+ "model_type": "speecht5",
71
+ "num_conv_pos_embedding_groups": 16,
72
+ "num_conv_pos_embeddings": 128,
73
+ "num_feat_extract_layers": 7,
74
+ "num_mel_bins": 80,
75
+ "pad_token_id": 1,
76
+ "positional_dropout": 0.1,
77
+ "reduction_factor": 2,
78
+ "scale_embedding": false,
79
+ "speaker_embedding_dim": 512,
80
+ "speech_decoder_postnet_dropout": 0.5,
81
+ "speech_decoder_postnet_kernel": 5,
82
+ "speech_decoder_postnet_layers": 5,
83
+ "speech_decoder_postnet_units": 256,
84
+ "speech_decoder_prenet_dropout": 0.5,
85
+ "speech_decoder_prenet_layers": 2,
86
+ "speech_decoder_prenet_units": 256,
87
+ "torch_dtype": "float32",
88
+ "transformers_version": "4.46.1",
89
+ "use_cache": false,
90
+ "use_guided_attention_loss": true,
91
+ "vocab_size": 81
92
+ }
checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "decoder_start_token_id": 2,
5
+ "eos_token_id": 2,
6
+ "max_length": 1876,
7
+ "pad_token_id": 1,
8
+ "transformers_version": "4.46.1"
9
+ }
checkpoint-1000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c5523d209511a8128afd892203133c6ff25481d87762c85914ce2373695e897
3
+ size 577789320
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd0235c19281a1a11ce8bc188dc413063516990b69e039acfa1d4d10d9c91760
3
+ size 1155772233
checkpoint-1000/preprocessor_config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": false,
3
+ "feature_extractor_type": "SpeechT5FeatureExtractor",
4
+ "feature_size": 1,
5
+ "fmax": 7600,
6
+ "fmin": 80,
7
+ "frame_signal_scale": 1.0,
8
+ "hop_length": 16,
9
+ "mel_floor": 1e-10,
10
+ "num_mel_bins": 80,
11
+ "padding_side": "right",
12
+ "padding_value": 0.0,
13
+ "processor_class": "SpeechT5Processor",
14
+ "reduction_factor": 2,
15
+ "return_attention_mask": true,
16
+ "sampling_rate": 16000,
17
+ "win_function": "hann_window",
18
+ "win_length": 64
19
+ }
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2391edd5840a83f113162293c889574d0eab13eb0642f28aae70d752ea782ed
3
+ size 14244
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5101d8c1f86d6f48167e50b1164b9ba363ab76694ff2d5c1e326e3d5f94ecaef
3
+ size 1064
checkpoint-1000/special_tokens_map.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "mask_token": {
5
+ "content": "<mask>",
6
+ "lstrip": true,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false
10
+ },
11
+ "pad_token": "<pad>",
12
+ "unk_token": "<unk>"
13
+ }
checkpoint-1000/spm_char.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
3
+ size 238473
checkpoint-1000/tokenizer_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "79": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": true,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "80": {
44
+ "content": "<ctc_blank>",
45
+ "lstrip": false,
46
+ "normalized": true,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": false
50
+ }
51
+ },
52
+ "bos_token": "<s>",
53
+ "clean_up_tokenization_spaces": false,
54
+ "eos_token": "</s>",
55
+ "mask_token": "<mask>",
56
+ "model_max_length": 600,
57
+ "normalize": false,
58
+ "pad_token": "<pad>",
59
+ "processor_class": "SpeechT5Processor",
60
+ "sp_model_kwargs": {},
61
+ "tokenizer_class": "SpeechT5Tokenizer",
62
+ "unk_token": "<unk>"
63
+ }
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,321 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.46117448806762695,
3
+ "best_model_checkpoint": "speecht5_finetuned_voxpopuli_pl/checkpoint-1000",
4
+ "epoch": 5.025125628140704,
5
+ "eval_steps": 1000,
6
+ "global_step": 1000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.12562814070351758,
13
+ "grad_norm": 34.14754104614258,
14
+ "learning_rate": 4.6000000000000004e-07,
15
+ "loss": 6.2754,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.25125628140703515,
20
+ "grad_norm": 25.906333923339844,
21
+ "learning_rate": 9.400000000000001e-07,
22
+ "loss": 6.1497,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.3768844221105528,
27
+ "grad_norm": 29.097463607788086,
28
+ "learning_rate": 1.44e-06,
29
+ "loss": 6.0335,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.5025125628140703,
34
+ "grad_norm": 18.543378829956055,
35
+ "learning_rate": 1.94e-06,
36
+ "loss": 5.8815,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.628140703517588,
41
+ "grad_norm": 24.147188186645508,
42
+ "learning_rate": 2.4400000000000004e-06,
43
+ "loss": 5.5659,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.7537688442211056,
48
+ "grad_norm": 15.634539604187012,
49
+ "learning_rate": 2.9400000000000002e-06,
50
+ "loss": 5.3573,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.8793969849246231,
55
+ "grad_norm": 14.31460189819336,
56
+ "learning_rate": 3.44e-06,
57
+ "loss": 5.2616,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 1.0050251256281406,
62
+ "grad_norm": 19.75978660583496,
63
+ "learning_rate": 3.94e-06,
64
+ "loss": 5.1787,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 1.1306532663316582,
69
+ "grad_norm": 12.639854431152344,
70
+ "learning_rate": 4.440000000000001e-06,
71
+ "loss": 5.1214,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 1.2562814070351758,
76
+ "grad_norm": 17.792396545410156,
77
+ "learning_rate": 4.94e-06,
78
+ "loss": 5.0703,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 1.3819095477386933,
83
+ "grad_norm": 22.859201431274414,
84
+ "learning_rate": 5.4400000000000004e-06,
85
+ "loss": 4.9714,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 1.507537688442211,
90
+ "grad_norm": 16.62458610534668,
91
+ "learning_rate": 5.94e-06,
92
+ "loss": 4.8681,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 1.6331658291457285,
97
+ "grad_norm": 17.881174087524414,
98
+ "learning_rate": 6.440000000000001e-06,
99
+ "loss": 4.8252,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 1.758793969849246,
104
+ "grad_norm": 14.765595436096191,
105
+ "learning_rate": 6.9400000000000005e-06,
106
+ "loss": 4.5588,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 1.8844221105527639,
111
+ "grad_norm": 13.808978080749512,
112
+ "learning_rate": 7.440000000000001e-06,
113
+ "loss": 4.5211,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 2.0100502512562812,
118
+ "grad_norm": 10.674385070800781,
119
+ "learning_rate": 7.94e-06,
120
+ "loss": 4.4888,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 2.135678391959799,
125
+ "grad_norm": 15.555588722229004,
126
+ "learning_rate": 8.44e-06,
127
+ "loss": 4.4513,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 2.2613065326633164,
132
+ "grad_norm": 13.092530250549316,
133
+ "learning_rate": 8.94e-06,
134
+ "loss": 4.3646,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 2.3869346733668344,
139
+ "grad_norm": 12.971352577209473,
140
+ "learning_rate": 9.440000000000001e-06,
141
+ "loss": 4.3459,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 2.5125628140703515,
146
+ "grad_norm": 10.348517417907715,
147
+ "learning_rate": 9.940000000000001e-06,
148
+ "loss": 4.2915,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 2.6381909547738696,
153
+ "grad_norm": 13.625853538513184,
154
+ "learning_rate": 9.937142857142858e-06,
155
+ "loss": 4.3095,
156
+ "step": 525
157
+ },
158
+ {
159
+ "epoch": 2.7638190954773867,
160
+ "grad_norm": 11.426923751831055,
161
+ "learning_rate": 9.865714285714285e-06,
162
+ "loss": 4.1899,
163
+ "step": 550
164
+ },
165
+ {
166
+ "epoch": 2.8894472361809047,
167
+ "grad_norm": 11.516312599182129,
168
+ "learning_rate": 9.794285714285714e-06,
169
+ "loss": 4.2233,
170
+ "step": 575
171
+ },
172
+ {
173
+ "epoch": 3.0150753768844223,
174
+ "grad_norm": 16.141698837280273,
175
+ "learning_rate": 9.722857142857143e-06,
176
+ "loss": 4.1831,
177
+ "step": 600
178
+ },
179
+ {
180
+ "epoch": 3.14070351758794,
181
+ "grad_norm": 14.935202598571777,
182
+ "learning_rate": 9.651428571428572e-06,
183
+ "loss": 4.1739,
184
+ "step": 625
185
+ },
186
+ {
187
+ "epoch": 3.2663316582914574,
188
+ "grad_norm": 10.119762420654297,
189
+ "learning_rate": 9.58e-06,
190
+ "loss": 4.1649,
191
+ "step": 650
192
+ },
193
+ {
194
+ "epoch": 3.391959798994975,
195
+ "grad_norm": 16.50948143005371,
196
+ "learning_rate": 9.508571428571429e-06,
197
+ "loss": 4.1627,
198
+ "step": 675
199
+ },
200
+ {
201
+ "epoch": 3.5175879396984926,
202
+ "grad_norm": 10.08300495147705,
203
+ "learning_rate": 9.437142857142858e-06,
204
+ "loss": 4.1915,
205
+ "step": 700
206
+ },
207
+ {
208
+ "epoch": 3.64321608040201,
209
+ "grad_norm": 9.375288963317871,
210
+ "learning_rate": 9.365714285714287e-06,
211
+ "loss": 4.1279,
212
+ "step": 725
213
+ },
214
+ {
215
+ "epoch": 3.7688442211055277,
216
+ "grad_norm": 13.562234878540039,
217
+ "learning_rate": 9.294285714285714e-06,
218
+ "loss": 4.1083,
219
+ "step": 750
220
+ },
221
+ {
222
+ "epoch": 3.8944723618090453,
223
+ "grad_norm": 15.115653991699219,
224
+ "learning_rate": 9.222857142857143e-06,
225
+ "loss": 4.1236,
226
+ "step": 775
227
+ },
228
+ {
229
+ "epoch": 4.0201005025125625,
230
+ "grad_norm": 9.049237251281738,
231
+ "learning_rate": 9.151428571428572e-06,
232
+ "loss": 4.0964,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 4.1457286432160805,
237
+ "grad_norm": 9.250771522521973,
238
+ "learning_rate": 9.080000000000001e-06,
239
+ "loss": 4.0669,
240
+ "step": 825
241
+ },
242
+ {
243
+ "epoch": 4.271356783919598,
244
+ "grad_norm": 12.07302474975586,
245
+ "learning_rate": 9.00857142857143e-06,
246
+ "loss": 4.0587,
247
+ "step": 850
248
+ },
249
+ {
250
+ "epoch": 4.396984924623116,
251
+ "grad_norm": 9.538933753967285,
252
+ "learning_rate": 8.937142857142857e-06,
253
+ "loss": 4.0837,
254
+ "step": 875
255
+ },
256
+ {
257
+ "epoch": 4.522613065326633,
258
+ "grad_norm": 14.347626686096191,
259
+ "learning_rate": 8.865714285714287e-06,
260
+ "loss": 4.0086,
261
+ "step": 900
262
+ },
263
+ {
264
+ "epoch": 4.648241206030151,
265
+ "grad_norm": 10.831369400024414,
266
+ "learning_rate": 8.794285714285716e-06,
267
+ "loss": 4.0563,
268
+ "step": 925
269
+ },
270
+ {
271
+ "epoch": 4.773869346733669,
272
+ "grad_norm": 9.891343116760254,
273
+ "learning_rate": 8.722857142857145e-06,
274
+ "loss": 4.0057,
275
+ "step": 950
276
+ },
277
+ {
278
+ "epoch": 4.899497487437186,
279
+ "grad_norm": 9.572471618652344,
280
+ "learning_rate": 8.651428571428572e-06,
281
+ "loss": 4.0557,
282
+ "step": 975
283
+ },
284
+ {
285
+ "epoch": 5.025125628140704,
286
+ "grad_norm": 9.745298385620117,
287
+ "learning_rate": 8.580000000000001e-06,
288
+ "loss": 3.9902,
289
+ "step": 1000
290
+ },
291
+ {
292
+ "epoch": 5.025125628140704,
293
+ "eval_loss": 0.46117448806762695,
294
+ "eval_runtime": 75.4681,
295
+ "eval_samples_per_second": 9.381,
296
+ "eval_steps_per_second": 4.691,
297
+ "step": 1000
298
+ }
299
+ ],
300
+ "logging_steps": 25,
301
+ "max_steps": 4000,
302
+ "num_input_tokens_seen": 0,
303
+ "num_train_epochs": 21,
304
+ "save_steps": 1000,
305
+ "stateful_callbacks": {
306
+ "TrainerControl": {
307
+ "args": {
308
+ "should_epoch_stop": false,
309
+ "should_evaluate": false,
310
+ "should_log": false,
311
+ "should_save": true,
312
+ "should_training_stop": false
313
+ },
314
+ "attributes": {}
315
+ }
316
+ },
317
+ "total_flos": 4502414650961304.0,
318
+ "train_batch_size": 4,
319
+ "trial_name": null,
320
+ "trial_params": null
321
+ }
checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c6ae756af4eb40168228ea5b672dedd9146f11deae4f8d64668e0b2d5a69781
3
+ size 5432
checkpoint-2000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-2000/config.json ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/speecht5_tts",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "SpeechT5ForTextToSpeech"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 0,
10
+ "conv_bias": false,
11
+ "conv_dim": [
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "conv_kernel": [
21
+ 10,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 2,
27
+ 2
28
+ ],
29
+ "conv_stride": [
30
+ 5,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2
37
+ ],
38
+ "decoder_attention_heads": 12,
39
+ "decoder_ffn_dim": 3072,
40
+ "decoder_layerdrop": 0.1,
41
+ "decoder_layers": 6,
42
+ "decoder_start_token_id": 2,
43
+ "encoder_attention_heads": 12,
44
+ "encoder_ffn_dim": 3072,
45
+ "encoder_layerdrop": 0.1,
46
+ "encoder_layers": 12,
47
+ "encoder_max_relative_position": 160,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_norm": "group",
51
+ "feat_proj_dropout": 0.0,
52
+ "guided_attention_loss_num_heads": 2,
53
+ "guided_attention_loss_scale": 10.0,
54
+ "guided_attention_loss_sigma": 0.4,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 768,
58
+ "initializer_range": 0.02,
59
+ "is_encoder_decoder": true,
60
+ "layer_norm_eps": 1e-05,
61
+ "mask_feature_length": 10,
62
+ "mask_feature_min_masks": 0,
63
+ "mask_feature_prob": 0.0,
64
+ "mask_time_length": 10,
65
+ "mask_time_min_masks": 2,
66
+ "mask_time_prob": 0.05,
67
+ "max_length": null,
68
+ "max_speech_positions": 1876,
69
+ "max_text_positions": 600,
70
+ "model_type": "speecht5",
71
+ "num_conv_pos_embedding_groups": 16,
72
+ "num_conv_pos_embeddings": 128,
73
+ "num_feat_extract_layers": 7,
74
+ "num_mel_bins": 80,
75
+ "pad_token_id": 1,
76
+ "positional_dropout": 0.1,
77
+ "reduction_factor": 2,
78
+ "scale_embedding": false,
79
+ "speaker_embedding_dim": 512,
80
+ "speech_decoder_postnet_dropout": 0.5,
81
+ "speech_decoder_postnet_kernel": 5,
82
+ "speech_decoder_postnet_layers": 5,
83
+ "speech_decoder_postnet_units": 256,
84
+ "speech_decoder_prenet_dropout": 0.5,
85
+ "speech_decoder_prenet_layers": 2,
86
+ "speech_decoder_prenet_units": 256,
87
+ "torch_dtype": "float32",
88
+ "transformers_version": "4.46.1",
89
+ "use_cache": false,
90
+ "use_guided_attention_loss": true,
91
+ "vocab_size": 81
92
+ }
checkpoint-2000/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "decoder_start_token_id": 2,
5
+ "eos_token_id": 2,
6
+ "max_length": 1876,
7
+ "pad_token_id": 1,
8
+ "transformers_version": "4.46.1"
9
+ }
checkpoint-2000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df394cff2c937d31a29d6243ea8e3ebcc79a50d23be775a4fba7dd7734dd1582
3
+ size 577789320
checkpoint-2000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12c2f15b423ce9e08c24c5f007046f57b502016d95208dbd3b0962042f7351fb
3
+ size 1155772233
checkpoint-2000/preprocessor_config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": false,
3
+ "feature_extractor_type": "SpeechT5FeatureExtractor",
4
+ "feature_size": 1,
5
+ "fmax": 7600,
6
+ "fmin": 80,
7
+ "frame_signal_scale": 1.0,
8
+ "hop_length": 16,
9
+ "mel_floor": 1e-10,
10
+ "num_mel_bins": 80,
11
+ "padding_side": "right",
12
+ "padding_value": 0.0,
13
+ "processor_class": "SpeechT5Processor",
14
+ "reduction_factor": 2,
15
+ "return_attention_mask": true,
16
+ "sampling_rate": 16000,
17
+ "win_function": "hann_window",
18
+ "win_length": 64
19
+ }
checkpoint-2000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3866dfdf0b13997fd63f79fbe98278f2ebe8e2dceb446d3d72314597efc00327
3
+ size 14244
checkpoint-2000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b80a609c64a12b4db2f38941ea479b9a30f9351b7aac74f4956e8686dc338317
3
+ size 1064
checkpoint-2000/special_tokens_map.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "mask_token": {
5
+ "content": "<mask>",
6
+ "lstrip": true,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false
10
+ },
11
+ "pad_token": "<pad>",
12
+ "unk_token": "<unk>"
13
+ }
checkpoint-2000/spm_char.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
3
+ size 238473
checkpoint-2000/tokenizer_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "79": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": true,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "80": {
44
+ "content": "<ctc_blank>",
45
+ "lstrip": false,
46
+ "normalized": true,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": false
50
+ }
51
+ },
52
+ "bos_token": "<s>",
53
+ "clean_up_tokenization_spaces": false,
54
+ "eos_token": "</s>",
55
+ "mask_token": "<mask>",
56
+ "model_max_length": 600,
57
+ "normalize": false,
58
+ "pad_token": "<pad>",
59
+ "processor_class": "SpeechT5Processor",
60
+ "sp_model_kwargs": {},
61
+ "tokenizer_class": "SpeechT5Tokenizer",
62
+ "unk_token": "<unk>"
63
+ }
checkpoint-2000/trainer_state.json ADDED
@@ -0,0 +1,609 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.44917792081832886,
3
+ "best_model_checkpoint": "speecht5_finetuned_voxpopuli_pl/checkpoint-2000",
4
+ "epoch": 10.050251256281408,
5
+ "eval_steps": 1000,
6
+ "global_step": 2000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.12562814070351758,
13
+ "grad_norm": 34.14754104614258,
14
+ "learning_rate": 4.6000000000000004e-07,
15
+ "loss": 6.2754,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.25125628140703515,
20
+ "grad_norm": 25.906333923339844,
21
+ "learning_rate": 9.400000000000001e-07,
22
+ "loss": 6.1497,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.3768844221105528,
27
+ "grad_norm": 29.097463607788086,
28
+ "learning_rate": 1.44e-06,
29
+ "loss": 6.0335,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.5025125628140703,
34
+ "grad_norm": 18.543378829956055,
35
+ "learning_rate": 1.94e-06,
36
+ "loss": 5.8815,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.628140703517588,
41
+ "grad_norm": 24.147188186645508,
42
+ "learning_rate": 2.4400000000000004e-06,
43
+ "loss": 5.5659,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.7537688442211056,
48
+ "grad_norm": 15.634539604187012,
49
+ "learning_rate": 2.9400000000000002e-06,
50
+ "loss": 5.3573,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.8793969849246231,
55
+ "grad_norm": 14.31460189819336,
56
+ "learning_rate": 3.44e-06,
57
+ "loss": 5.2616,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 1.0050251256281406,
62
+ "grad_norm": 19.75978660583496,
63
+ "learning_rate": 3.94e-06,
64
+ "loss": 5.1787,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 1.1306532663316582,
69
+ "grad_norm": 12.639854431152344,
70
+ "learning_rate": 4.440000000000001e-06,
71
+ "loss": 5.1214,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 1.2562814070351758,
76
+ "grad_norm": 17.792396545410156,
77
+ "learning_rate": 4.94e-06,
78
+ "loss": 5.0703,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 1.3819095477386933,
83
+ "grad_norm": 22.859201431274414,
84
+ "learning_rate": 5.4400000000000004e-06,
85
+ "loss": 4.9714,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 1.507537688442211,
90
+ "grad_norm": 16.62458610534668,
91
+ "learning_rate": 5.94e-06,
92
+ "loss": 4.8681,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 1.6331658291457285,
97
+ "grad_norm": 17.881174087524414,
98
+ "learning_rate": 6.440000000000001e-06,
99
+ "loss": 4.8252,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 1.758793969849246,
104
+ "grad_norm": 14.765595436096191,
105
+ "learning_rate": 6.9400000000000005e-06,
106
+ "loss": 4.5588,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 1.8844221105527639,
111
+ "grad_norm": 13.808978080749512,
112
+ "learning_rate": 7.440000000000001e-06,
113
+ "loss": 4.5211,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 2.0100502512562812,
118
+ "grad_norm": 10.674385070800781,
119
+ "learning_rate": 7.94e-06,
120
+ "loss": 4.4888,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 2.135678391959799,
125
+ "grad_norm": 15.555588722229004,
126
+ "learning_rate": 8.44e-06,
127
+ "loss": 4.4513,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 2.2613065326633164,
132
+ "grad_norm": 13.092530250549316,
133
+ "learning_rate": 8.94e-06,
134
+ "loss": 4.3646,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 2.3869346733668344,
139
+ "grad_norm": 12.971352577209473,
140
+ "learning_rate": 9.440000000000001e-06,
141
+ "loss": 4.3459,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 2.5125628140703515,
146
+ "grad_norm": 10.348517417907715,
147
+ "learning_rate": 9.940000000000001e-06,
148
+ "loss": 4.2915,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 2.6381909547738696,
153
+ "grad_norm": 13.625853538513184,
154
+ "learning_rate": 9.937142857142858e-06,
155
+ "loss": 4.3095,
156
+ "step": 525
157
+ },
158
+ {
159
+ "epoch": 2.7638190954773867,
160
+ "grad_norm": 11.426923751831055,
161
+ "learning_rate": 9.865714285714285e-06,
162
+ "loss": 4.1899,
163
+ "step": 550
164
+ },
165
+ {
166
+ "epoch": 2.8894472361809047,
167
+ "grad_norm": 11.516312599182129,
168
+ "learning_rate": 9.794285714285714e-06,
169
+ "loss": 4.2233,
170
+ "step": 575
171
+ },
172
+ {
173
+ "epoch": 3.0150753768844223,
174
+ "grad_norm": 16.141698837280273,
175
+ "learning_rate": 9.722857142857143e-06,
176
+ "loss": 4.1831,
177
+ "step": 600
178
+ },
179
+ {
180
+ "epoch": 3.14070351758794,
181
+ "grad_norm": 14.935202598571777,
182
+ "learning_rate": 9.651428571428572e-06,
183
+ "loss": 4.1739,
184
+ "step": 625
185
+ },
186
+ {
187
+ "epoch": 3.2663316582914574,
188
+ "grad_norm": 10.119762420654297,
189
+ "learning_rate": 9.58e-06,
190
+ "loss": 4.1649,
191
+ "step": 650
192
+ },
193
+ {
194
+ "epoch": 3.391959798994975,
195
+ "grad_norm": 16.50948143005371,
196
+ "learning_rate": 9.508571428571429e-06,
197
+ "loss": 4.1627,
198
+ "step": 675
199
+ },
200
+ {
201
+ "epoch": 3.5175879396984926,
202
+ "grad_norm": 10.08300495147705,
203
+ "learning_rate": 9.437142857142858e-06,
204
+ "loss": 4.1915,
205
+ "step": 700
206
+ },
207
+ {
208
+ "epoch": 3.64321608040201,
209
+ "grad_norm": 9.375288963317871,
210
+ "learning_rate": 9.365714285714287e-06,
211
+ "loss": 4.1279,
212
+ "step": 725
213
+ },
214
+ {
215
+ "epoch": 3.7688442211055277,
216
+ "grad_norm": 13.562234878540039,
217
+ "learning_rate": 9.294285714285714e-06,
218
+ "loss": 4.1083,
219
+ "step": 750
220
+ },
221
+ {
222
+ "epoch": 3.8944723618090453,
223
+ "grad_norm": 15.115653991699219,
224
+ "learning_rate": 9.222857142857143e-06,
225
+ "loss": 4.1236,
226
+ "step": 775
227
+ },
228
+ {
229
+ "epoch": 4.0201005025125625,
230
+ "grad_norm": 9.049237251281738,
231
+ "learning_rate": 9.151428571428572e-06,
232
+ "loss": 4.0964,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 4.1457286432160805,
237
+ "grad_norm": 9.250771522521973,
238
+ "learning_rate": 9.080000000000001e-06,
239
+ "loss": 4.0669,
240
+ "step": 825
241
+ },
242
+ {
243
+ "epoch": 4.271356783919598,
244
+ "grad_norm": 12.07302474975586,
245
+ "learning_rate": 9.00857142857143e-06,
246
+ "loss": 4.0587,
247
+ "step": 850
248
+ },
249
+ {
250
+ "epoch": 4.396984924623116,
251
+ "grad_norm": 9.538933753967285,
252
+ "learning_rate": 8.937142857142857e-06,
253
+ "loss": 4.0837,
254
+ "step": 875
255
+ },
256
+ {
257
+ "epoch": 4.522613065326633,
258
+ "grad_norm": 14.347626686096191,
259
+ "learning_rate": 8.865714285714287e-06,
260
+ "loss": 4.0086,
261
+ "step": 900
262
+ },
263
+ {
264
+ "epoch": 4.648241206030151,
265
+ "grad_norm": 10.831369400024414,
266
+ "learning_rate": 8.794285714285716e-06,
267
+ "loss": 4.0563,
268
+ "step": 925
269
+ },
270
+ {
271
+ "epoch": 4.773869346733669,
272
+ "grad_norm": 9.891343116760254,
273
+ "learning_rate": 8.722857142857145e-06,
274
+ "loss": 4.0057,
275
+ "step": 950
276
+ },
277
+ {
278
+ "epoch": 4.899497487437186,
279
+ "grad_norm": 9.572471618652344,
280
+ "learning_rate": 8.651428571428572e-06,
281
+ "loss": 4.0557,
282
+ "step": 975
283
+ },
284
+ {
285
+ "epoch": 5.025125628140704,
286
+ "grad_norm": 9.745298385620117,
287
+ "learning_rate": 8.580000000000001e-06,
288
+ "loss": 3.9902,
289
+ "step": 1000
290
+ },
291
+ {
292
+ "epoch": 5.025125628140704,
293
+ "eval_loss": 0.46117448806762695,
294
+ "eval_runtime": 75.4681,
295
+ "eval_samples_per_second": 9.381,
296
+ "eval_steps_per_second": 4.691,
297
+ "step": 1000
298
+ },
299
+ {
300
+ "epoch": 5.150753768844221,
301
+ "grad_norm": 10.219923973083496,
302
+ "learning_rate": 8.50857142857143e-06,
303
+ "loss": 3.9991,
304
+ "step": 1025
305
+ },
306
+ {
307
+ "epoch": 5.276381909547739,
308
+ "grad_norm": 9.258833885192871,
309
+ "learning_rate": 8.437142857142859e-06,
310
+ "loss": 4.0032,
311
+ "step": 1050
312
+ },
313
+ {
314
+ "epoch": 5.402010050251256,
315
+ "grad_norm": 12.21104907989502,
316
+ "learning_rate": 8.365714285714286e-06,
317
+ "loss": 4.0337,
318
+ "step": 1075
319
+ },
320
+ {
321
+ "epoch": 5.527638190954773,
322
+ "grad_norm": 10.777599334716797,
323
+ "learning_rate": 8.294285714285715e-06,
324
+ "loss": 4.0036,
325
+ "step": 1100
326
+ },
327
+ {
328
+ "epoch": 5.653266331658291,
329
+ "grad_norm": 10.322349548339844,
330
+ "learning_rate": 8.222857142857144e-06,
331
+ "loss": 3.9811,
332
+ "step": 1125
333
+ },
334
+ {
335
+ "epoch": 5.778894472361809,
336
+ "grad_norm": 9.940053939819336,
337
+ "learning_rate": 8.151428571428572e-06,
338
+ "loss": 4.0264,
339
+ "step": 1150
340
+ },
341
+ {
342
+ "epoch": 5.9045226130653266,
343
+ "grad_norm": 12.996850967407227,
344
+ "learning_rate": 8.08e-06,
345
+ "loss": 3.9783,
346
+ "step": 1175
347
+ },
348
+ {
349
+ "epoch": 6.030150753768845,
350
+ "grad_norm": 10.784270286560059,
351
+ "learning_rate": 8.00857142857143e-06,
352
+ "loss": 3.936,
353
+ "step": 1200
354
+ },
355
+ {
356
+ "epoch": 6.155778894472362,
357
+ "grad_norm": 10.86549186706543,
358
+ "learning_rate": 7.937142857142857e-06,
359
+ "loss": 3.9803,
360
+ "step": 1225
361
+ },
362
+ {
363
+ "epoch": 6.28140703517588,
364
+ "grad_norm": 10.926647186279297,
365
+ "learning_rate": 7.865714285714286e-06,
366
+ "loss": 3.9587,
367
+ "step": 1250
368
+ },
369
+ {
370
+ "epoch": 6.407035175879397,
371
+ "grad_norm": 12.103693008422852,
372
+ "learning_rate": 7.794285714285715e-06,
373
+ "loss": 3.9441,
374
+ "step": 1275
375
+ },
376
+ {
377
+ "epoch": 6.532663316582915,
378
+ "grad_norm": 14.731566429138184,
379
+ "learning_rate": 7.722857142857142e-06,
380
+ "loss": 3.9541,
381
+ "step": 1300
382
+ },
383
+ {
384
+ "epoch": 6.658291457286432,
385
+ "grad_norm": 13.810087203979492,
386
+ "learning_rate": 7.651428571428571e-06,
387
+ "loss": 3.962,
388
+ "step": 1325
389
+ },
390
+ {
391
+ "epoch": 6.78391959798995,
392
+ "grad_norm": 10.733597755432129,
393
+ "learning_rate": 7.58e-06,
394
+ "loss": 3.9696,
395
+ "step": 1350
396
+ },
397
+ {
398
+ "epoch": 6.909547738693467,
399
+ "grad_norm": 12.271906852722168,
400
+ "learning_rate": 7.508571428571429e-06,
401
+ "loss": 3.9319,
402
+ "step": 1375
403
+ },
404
+ {
405
+ "epoch": 7.035175879396985,
406
+ "grad_norm": 17.454214096069336,
407
+ "learning_rate": 7.4371428571428575e-06,
408
+ "loss": 3.9096,
409
+ "step": 1400
410
+ },
411
+ {
412
+ "epoch": 7.160804020100502,
413
+ "grad_norm": 9.891731262207031,
414
+ "learning_rate": 7.365714285714286e-06,
415
+ "loss": 3.963,
416
+ "step": 1425
417
+ },
418
+ {
419
+ "epoch": 7.28643216080402,
420
+ "grad_norm": 10.400883674621582,
421
+ "learning_rate": 7.294285714285715e-06,
422
+ "loss": 3.9529,
423
+ "step": 1450
424
+ },
425
+ {
426
+ "epoch": 7.4120603015075375,
427
+ "grad_norm": 8.886434555053711,
428
+ "learning_rate": 7.222857142857144e-06,
429
+ "loss": 3.9048,
430
+ "step": 1475
431
+ },
432
+ {
433
+ "epoch": 7.5376884422110555,
434
+ "grad_norm": 13.72711181640625,
435
+ "learning_rate": 7.151428571428573e-06,
436
+ "loss": 3.9097,
437
+ "step": 1500
438
+ },
439
+ {
440
+ "epoch": 7.663316582914573,
441
+ "grad_norm": 10.161803245544434,
442
+ "learning_rate": 7.08e-06,
443
+ "loss": 3.9029,
444
+ "step": 1525
445
+ },
446
+ {
447
+ "epoch": 7.788944723618091,
448
+ "grad_norm": 9.717086791992188,
449
+ "learning_rate": 7.008571428571429e-06,
450
+ "loss": 3.9028,
451
+ "step": 1550
452
+ },
453
+ {
454
+ "epoch": 7.914572864321608,
455
+ "grad_norm": 9.483698844909668,
456
+ "learning_rate": 6.937142857142858e-06,
457
+ "loss": 3.9157,
458
+ "step": 1575
459
+ },
460
+ {
461
+ "epoch": 8.040201005025125,
462
+ "grad_norm": 8.951908111572266,
463
+ "learning_rate": 6.865714285714287e-06,
464
+ "loss": 3.9106,
465
+ "step": 1600
466
+ },
467
+ {
468
+ "epoch": 8.165829145728644,
469
+ "grad_norm": 8.099654197692871,
470
+ "learning_rate": 6.794285714285714e-06,
471
+ "loss": 3.9062,
472
+ "step": 1625
473
+ },
474
+ {
475
+ "epoch": 8.291457286432161,
476
+ "grad_norm": 10.005452156066895,
477
+ "learning_rate": 6.722857142857143e-06,
478
+ "loss": 3.9139,
479
+ "step": 1650
480
+ },
481
+ {
482
+ "epoch": 8.417085427135678,
483
+ "grad_norm": 7.832030296325684,
484
+ "learning_rate": 6.651428571428572e-06,
485
+ "loss": 3.8731,
486
+ "step": 1675
487
+ },
488
+ {
489
+ "epoch": 8.542713567839195,
490
+ "grad_norm": 9.519272804260254,
491
+ "learning_rate": 6.5800000000000005e-06,
492
+ "loss": 3.9234,
493
+ "step": 1700
494
+ },
495
+ {
496
+ "epoch": 8.668341708542714,
497
+ "grad_norm": 7.6004252433776855,
498
+ "learning_rate": 6.5085714285714295e-06,
499
+ "loss": 3.8889,
500
+ "step": 1725
501
+ },
502
+ {
503
+ "epoch": 8.793969849246231,
504
+ "grad_norm": 8.43916130065918,
505
+ "learning_rate": 6.437142857142858e-06,
506
+ "loss": 3.8813,
507
+ "step": 1750
508
+ },
509
+ {
510
+ "epoch": 8.919597989949748,
511
+ "grad_norm": 9.573629379272461,
512
+ "learning_rate": 6.365714285714286e-06,
513
+ "loss": 3.9234,
514
+ "step": 1775
515
+ },
516
+ {
517
+ "epoch": 9.045226130653266,
518
+ "grad_norm": 11.034466743469238,
519
+ "learning_rate": 6.294285714285715e-06,
520
+ "loss": 3.9227,
521
+ "step": 1800
522
+ },
523
+ {
524
+ "epoch": 9.170854271356784,
525
+ "grad_norm": 9.65125560760498,
526
+ "learning_rate": 6.222857142857144e-06,
527
+ "loss": 3.8806,
528
+ "step": 1825
529
+ },
530
+ {
531
+ "epoch": 9.296482412060302,
532
+ "grad_norm": 8.907113075256348,
533
+ "learning_rate": 6.151428571428571e-06,
534
+ "loss": 3.8878,
535
+ "step": 1850
536
+ },
537
+ {
538
+ "epoch": 9.422110552763819,
539
+ "grad_norm": 9.000490188598633,
540
+ "learning_rate": 6.08e-06,
541
+ "loss": 3.858,
542
+ "step": 1875
543
+ },
544
+ {
545
+ "epoch": 9.547738693467338,
546
+ "grad_norm": 10.825636863708496,
547
+ "learning_rate": 6.008571428571429e-06,
548
+ "loss": 3.8546,
549
+ "step": 1900
550
+ },
551
+ {
552
+ "epoch": 9.673366834170855,
553
+ "grad_norm": 8.722935676574707,
554
+ "learning_rate": 5.937142857142858e-06,
555
+ "loss": 3.871,
556
+ "step": 1925
557
+ },
558
+ {
559
+ "epoch": 9.798994974874372,
560
+ "grad_norm": 9.243980407714844,
561
+ "learning_rate": 5.865714285714286e-06,
562
+ "loss": 3.8486,
563
+ "step": 1950
564
+ },
565
+ {
566
+ "epoch": 9.924623115577889,
567
+ "grad_norm": 10.595390319824219,
568
+ "learning_rate": 5.794285714285715e-06,
569
+ "loss": 3.8973,
570
+ "step": 1975
571
+ },
572
+ {
573
+ "epoch": 10.050251256281408,
574
+ "grad_norm": 11.8334321975708,
575
+ "learning_rate": 5.722857142857144e-06,
576
+ "loss": 3.8124,
577
+ "step": 2000
578
+ },
579
+ {
580
+ "epoch": 10.050251256281408,
581
+ "eval_loss": 0.44917792081832886,
582
+ "eval_runtime": 76.9989,
583
+ "eval_samples_per_second": 9.195,
584
+ "eval_steps_per_second": 4.597,
585
+ "step": 2000
586
+ }
587
+ ],
588
+ "logging_steps": 25,
589
+ "max_steps": 4000,
590
+ "num_input_tokens_seen": 0,
591
+ "num_train_epochs": 21,
592
+ "save_steps": 1000,
593
+ "stateful_callbacks": {
594
+ "TrainerControl": {
595
+ "args": {
596
+ "should_epoch_stop": false,
597
+ "should_evaluate": false,
598
+ "should_log": false,
599
+ "should_save": true,
600
+ "should_training_stop": false
601
+ },
602
+ "attributes": {}
603
+ }
604
+ },
605
+ "total_flos": 9000871252452288.0,
606
+ "train_batch_size": 4,
607
+ "trial_name": null,
608
+ "trial_params": null
609
+ }
checkpoint-2000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c6ae756af4eb40168228ea5b672dedd9146f11deae4f8d64668e0b2d5a69781
3
+ size 5432
checkpoint-3000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-3000/config.json ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/speecht5_tts",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "SpeechT5ForTextToSpeech"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 0,
10
+ "conv_bias": false,
11
+ "conv_dim": [
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "conv_kernel": [
21
+ 10,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 2,
27
+ 2
28
+ ],
29
+ "conv_stride": [
30
+ 5,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2
37
+ ],
38
+ "decoder_attention_heads": 12,
39
+ "decoder_ffn_dim": 3072,
40
+ "decoder_layerdrop": 0.1,
41
+ "decoder_layers": 6,
42
+ "decoder_start_token_id": 2,
43
+ "encoder_attention_heads": 12,
44
+ "encoder_ffn_dim": 3072,
45
+ "encoder_layerdrop": 0.1,
46
+ "encoder_layers": 12,
47
+ "encoder_max_relative_position": 160,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_norm": "group",
51
+ "feat_proj_dropout": 0.0,
52
+ "guided_attention_loss_num_heads": 2,
53
+ "guided_attention_loss_scale": 10.0,
54
+ "guided_attention_loss_sigma": 0.4,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 768,
58
+ "initializer_range": 0.02,
59
+ "is_encoder_decoder": true,
60
+ "layer_norm_eps": 1e-05,
61
+ "mask_feature_length": 10,
62
+ "mask_feature_min_masks": 0,
63
+ "mask_feature_prob": 0.0,
64
+ "mask_time_length": 10,
65
+ "mask_time_min_masks": 2,
66
+ "mask_time_prob": 0.05,
67
+ "max_length": null,
68
+ "max_speech_positions": 1876,
69
+ "max_text_positions": 600,
70
+ "model_type": "speecht5",
71
+ "num_conv_pos_embedding_groups": 16,
72
+ "num_conv_pos_embeddings": 128,
73
+ "num_feat_extract_layers": 7,
74
+ "num_mel_bins": 80,
75
+ "pad_token_id": 1,
76
+ "positional_dropout": 0.1,
77
+ "reduction_factor": 2,
78
+ "scale_embedding": false,
79
+ "speaker_embedding_dim": 512,
80
+ "speech_decoder_postnet_dropout": 0.5,
81
+ "speech_decoder_postnet_kernel": 5,
82
+ "speech_decoder_postnet_layers": 5,
83
+ "speech_decoder_postnet_units": 256,
84
+ "speech_decoder_prenet_dropout": 0.5,
85
+ "speech_decoder_prenet_layers": 2,
86
+ "speech_decoder_prenet_units": 256,
87
+ "torch_dtype": "float32",
88
+ "transformers_version": "4.46.1",
89
+ "use_cache": false,
90
+ "use_guided_attention_loss": true,
91
+ "vocab_size": 81
92
+ }
checkpoint-3000/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "decoder_start_token_id": 2,
5
+ "eos_token_id": 2,
6
+ "max_length": 1876,
7
+ "pad_token_id": 1,
8
+ "transformers_version": "4.46.1"
9
+ }
checkpoint-3000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3d626a5faf08be1905193f75162a4ec7b6f770ff4e9fa828e8b9ae33d56ed90
3
+ size 577789320
checkpoint-3000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f763cc61d97cb2027a0997575ca0b715bc34ab7c43dbbf5e513f229b92fe081
3
+ size 1155772233
checkpoint-3000/preprocessor_config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": false,
3
+ "feature_extractor_type": "SpeechT5FeatureExtractor",
4
+ "feature_size": 1,
5
+ "fmax": 7600,
6
+ "fmin": 80,
7
+ "frame_signal_scale": 1.0,
8
+ "hop_length": 16,
9
+ "mel_floor": 1e-10,
10
+ "num_mel_bins": 80,
11
+ "padding_side": "right",
12
+ "padding_value": 0.0,
13
+ "processor_class": "SpeechT5Processor",
14
+ "reduction_factor": 2,
15
+ "return_attention_mask": true,
16
+ "sampling_rate": 16000,
17
+ "win_function": "hann_window",
18
+ "win_length": 64
19
+ }
checkpoint-3000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28e16300e84ad755819fc9d8ce7454fb75d39320dda51990693e5315c30d74b2
3
+ size 14244
checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ff56efc76c16a3b9a712527179ae61c8d6dfccc7e3a53f8c421d6329adacfbb
3
+ size 1064
checkpoint-3000/special_tokens_map.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "mask_token": {
5
+ "content": "<mask>",
6
+ "lstrip": true,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false
10
+ },
11
+ "pad_token": "<pad>",
12
+ "unk_token": "<unk>"
13
+ }
checkpoint-3000/spm_char.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
3
+ size 238473
checkpoint-3000/tokenizer_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "79": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": true,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "80": {
44
+ "content": "<ctc_blank>",
45
+ "lstrip": false,
46
+ "normalized": true,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": false
50
+ }
51
+ },
52
+ "bos_token": "<s>",
53
+ "clean_up_tokenization_spaces": false,
54
+ "eos_token": "</s>",
55
+ "mask_token": "<mask>",
56
+ "model_max_length": 600,
57
+ "normalize": false,
58
+ "pad_token": "<pad>",
59
+ "processor_class": "SpeechT5Processor",
60
+ "sp_model_kwargs": {},
61
+ "tokenizer_class": "SpeechT5Tokenizer",
62
+ "unk_token": "<unk>"
63
+ }
checkpoint-3000/trainer_state.json ADDED
@@ -0,0 +1,897 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.44469836354255676,
3
+ "best_model_checkpoint": "speecht5_finetuned_voxpopuli_pl/checkpoint-3000",
4
+ "epoch": 15.075376884422111,
5
+ "eval_steps": 1000,
6
+ "global_step": 3000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.12562814070351758,
13
+ "grad_norm": 34.14754104614258,
14
+ "learning_rate": 4.6000000000000004e-07,
15
+ "loss": 6.2754,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.25125628140703515,
20
+ "grad_norm": 25.906333923339844,
21
+ "learning_rate": 9.400000000000001e-07,
22
+ "loss": 6.1497,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.3768844221105528,
27
+ "grad_norm": 29.097463607788086,
28
+ "learning_rate": 1.44e-06,
29
+ "loss": 6.0335,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.5025125628140703,
34
+ "grad_norm": 18.543378829956055,
35
+ "learning_rate": 1.94e-06,
36
+ "loss": 5.8815,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.628140703517588,
41
+ "grad_norm": 24.147188186645508,
42
+ "learning_rate": 2.4400000000000004e-06,
43
+ "loss": 5.5659,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.7537688442211056,
48
+ "grad_norm": 15.634539604187012,
49
+ "learning_rate": 2.9400000000000002e-06,
50
+ "loss": 5.3573,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.8793969849246231,
55
+ "grad_norm": 14.31460189819336,
56
+ "learning_rate": 3.44e-06,
57
+ "loss": 5.2616,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 1.0050251256281406,
62
+ "grad_norm": 19.75978660583496,
63
+ "learning_rate": 3.94e-06,
64
+ "loss": 5.1787,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 1.1306532663316582,
69
+ "grad_norm": 12.639854431152344,
70
+ "learning_rate": 4.440000000000001e-06,
71
+ "loss": 5.1214,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 1.2562814070351758,
76
+ "grad_norm": 17.792396545410156,
77
+ "learning_rate": 4.94e-06,
78
+ "loss": 5.0703,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 1.3819095477386933,
83
+ "grad_norm": 22.859201431274414,
84
+ "learning_rate": 5.4400000000000004e-06,
85
+ "loss": 4.9714,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 1.507537688442211,
90
+ "grad_norm": 16.62458610534668,
91
+ "learning_rate": 5.94e-06,
92
+ "loss": 4.8681,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 1.6331658291457285,
97
+ "grad_norm": 17.881174087524414,
98
+ "learning_rate": 6.440000000000001e-06,
99
+ "loss": 4.8252,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 1.758793969849246,
104
+ "grad_norm": 14.765595436096191,
105
+ "learning_rate": 6.9400000000000005e-06,
106
+ "loss": 4.5588,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 1.8844221105527639,
111
+ "grad_norm": 13.808978080749512,
112
+ "learning_rate": 7.440000000000001e-06,
113
+ "loss": 4.5211,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 2.0100502512562812,
118
+ "grad_norm": 10.674385070800781,
119
+ "learning_rate": 7.94e-06,
120
+ "loss": 4.4888,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 2.135678391959799,
125
+ "grad_norm": 15.555588722229004,
126
+ "learning_rate": 8.44e-06,
127
+ "loss": 4.4513,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 2.2613065326633164,
132
+ "grad_norm": 13.092530250549316,
133
+ "learning_rate": 8.94e-06,
134
+ "loss": 4.3646,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 2.3869346733668344,
139
+ "grad_norm": 12.971352577209473,
140
+ "learning_rate": 9.440000000000001e-06,
141
+ "loss": 4.3459,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 2.5125628140703515,
146
+ "grad_norm": 10.348517417907715,
147
+ "learning_rate": 9.940000000000001e-06,
148
+ "loss": 4.2915,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 2.6381909547738696,
153
+ "grad_norm": 13.625853538513184,
154
+ "learning_rate": 9.937142857142858e-06,
155
+ "loss": 4.3095,
156
+ "step": 525
157
+ },
158
+ {
159
+ "epoch": 2.7638190954773867,
160
+ "grad_norm": 11.426923751831055,
161
+ "learning_rate": 9.865714285714285e-06,
162
+ "loss": 4.1899,
163
+ "step": 550
164
+ },
165
+ {
166
+ "epoch": 2.8894472361809047,
167
+ "grad_norm": 11.516312599182129,
168
+ "learning_rate": 9.794285714285714e-06,
169
+ "loss": 4.2233,
170
+ "step": 575
171
+ },
172
+ {
173
+ "epoch": 3.0150753768844223,
174
+ "grad_norm": 16.141698837280273,
175
+ "learning_rate": 9.722857142857143e-06,
176
+ "loss": 4.1831,
177
+ "step": 600
178
+ },
179
+ {
180
+ "epoch": 3.14070351758794,
181
+ "grad_norm": 14.935202598571777,
182
+ "learning_rate": 9.651428571428572e-06,
183
+ "loss": 4.1739,
184
+ "step": 625
185
+ },
186
+ {
187
+ "epoch": 3.2663316582914574,
188
+ "grad_norm": 10.119762420654297,
189
+ "learning_rate": 9.58e-06,
190
+ "loss": 4.1649,
191
+ "step": 650
192
+ },
193
+ {
194
+ "epoch": 3.391959798994975,
195
+ "grad_norm": 16.50948143005371,
196
+ "learning_rate": 9.508571428571429e-06,
197
+ "loss": 4.1627,
198
+ "step": 675
199
+ },
200
+ {
201
+ "epoch": 3.5175879396984926,
202
+ "grad_norm": 10.08300495147705,
203
+ "learning_rate": 9.437142857142858e-06,
204
+ "loss": 4.1915,
205
+ "step": 700
206
+ },
207
+ {
208
+ "epoch": 3.64321608040201,
209
+ "grad_norm": 9.375288963317871,
210
+ "learning_rate": 9.365714285714287e-06,
211
+ "loss": 4.1279,
212
+ "step": 725
213
+ },
214
+ {
215
+ "epoch": 3.7688442211055277,
216
+ "grad_norm": 13.562234878540039,
217
+ "learning_rate": 9.294285714285714e-06,
218
+ "loss": 4.1083,
219
+ "step": 750
220
+ },
221
+ {
222
+ "epoch": 3.8944723618090453,
223
+ "grad_norm": 15.115653991699219,
224
+ "learning_rate": 9.222857142857143e-06,
225
+ "loss": 4.1236,
226
+ "step": 775
227
+ },
228
+ {
229
+ "epoch": 4.0201005025125625,
230
+ "grad_norm": 9.049237251281738,
231
+ "learning_rate": 9.151428571428572e-06,
232
+ "loss": 4.0964,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 4.1457286432160805,
237
+ "grad_norm": 9.250771522521973,
238
+ "learning_rate": 9.080000000000001e-06,
239
+ "loss": 4.0669,
240
+ "step": 825
241
+ },
242
+ {
243
+ "epoch": 4.271356783919598,
244
+ "grad_norm": 12.07302474975586,
245
+ "learning_rate": 9.00857142857143e-06,
246
+ "loss": 4.0587,
247
+ "step": 850
248
+ },
249
+ {
250
+ "epoch": 4.396984924623116,
251
+ "grad_norm": 9.538933753967285,
252
+ "learning_rate": 8.937142857142857e-06,
253
+ "loss": 4.0837,
254
+ "step": 875
255
+ },
256
+ {
257
+ "epoch": 4.522613065326633,
258
+ "grad_norm": 14.347626686096191,
259
+ "learning_rate": 8.865714285714287e-06,
260
+ "loss": 4.0086,
261
+ "step": 900
262
+ },
263
+ {
264
+ "epoch": 4.648241206030151,
265
+ "grad_norm": 10.831369400024414,
266
+ "learning_rate": 8.794285714285716e-06,
267
+ "loss": 4.0563,
268
+ "step": 925
269
+ },
270
+ {
271
+ "epoch": 4.773869346733669,
272
+ "grad_norm": 9.891343116760254,
273
+ "learning_rate": 8.722857142857145e-06,
274
+ "loss": 4.0057,
275
+ "step": 950
276
+ },
277
+ {
278
+ "epoch": 4.899497487437186,
279
+ "grad_norm": 9.572471618652344,
280
+ "learning_rate": 8.651428571428572e-06,
281
+ "loss": 4.0557,
282
+ "step": 975
283
+ },
284
+ {
285
+ "epoch": 5.025125628140704,
286
+ "grad_norm": 9.745298385620117,
287
+ "learning_rate": 8.580000000000001e-06,
288
+ "loss": 3.9902,
289
+ "step": 1000
290
+ },
291
+ {
292
+ "epoch": 5.025125628140704,
293
+ "eval_loss": 0.46117448806762695,
294
+ "eval_runtime": 75.4681,
295
+ "eval_samples_per_second": 9.381,
296
+ "eval_steps_per_second": 4.691,
297
+ "step": 1000
298
+ },
299
+ {
300
+ "epoch": 5.150753768844221,
301
+ "grad_norm": 10.219923973083496,
302
+ "learning_rate": 8.50857142857143e-06,
303
+ "loss": 3.9991,
304
+ "step": 1025
305
+ },
306
+ {
307
+ "epoch": 5.276381909547739,
308
+ "grad_norm": 9.258833885192871,
309
+ "learning_rate": 8.437142857142859e-06,
310
+ "loss": 4.0032,
311
+ "step": 1050
312
+ },
313
+ {
314
+ "epoch": 5.402010050251256,
315
+ "grad_norm": 12.21104907989502,
316
+ "learning_rate": 8.365714285714286e-06,
317
+ "loss": 4.0337,
318
+ "step": 1075
319
+ },
320
+ {
321
+ "epoch": 5.527638190954773,
322
+ "grad_norm": 10.777599334716797,
323
+ "learning_rate": 8.294285714285715e-06,
324
+ "loss": 4.0036,
325
+ "step": 1100
326
+ },
327
+ {
328
+ "epoch": 5.653266331658291,
329
+ "grad_norm": 10.322349548339844,
330
+ "learning_rate": 8.222857142857144e-06,
331
+ "loss": 3.9811,
332
+ "step": 1125
333
+ },
334
+ {
335
+ "epoch": 5.778894472361809,
336
+ "grad_norm": 9.940053939819336,
337
+ "learning_rate": 8.151428571428572e-06,
338
+ "loss": 4.0264,
339
+ "step": 1150
340
+ },
341
+ {
342
+ "epoch": 5.9045226130653266,
343
+ "grad_norm": 12.996850967407227,
344
+ "learning_rate": 8.08e-06,
345
+ "loss": 3.9783,
346
+ "step": 1175
347
+ },
348
+ {
349
+ "epoch": 6.030150753768845,
350
+ "grad_norm": 10.784270286560059,
351
+ "learning_rate": 8.00857142857143e-06,
352
+ "loss": 3.936,
353
+ "step": 1200
354
+ },
355
+ {
356
+ "epoch": 6.155778894472362,
357
+ "grad_norm": 10.86549186706543,
358
+ "learning_rate": 7.937142857142857e-06,
359
+ "loss": 3.9803,
360
+ "step": 1225
361
+ },
362
+ {
363
+ "epoch": 6.28140703517588,
364
+ "grad_norm": 10.926647186279297,
365
+ "learning_rate": 7.865714285714286e-06,
366
+ "loss": 3.9587,
367
+ "step": 1250
368
+ },
369
+ {
370
+ "epoch": 6.407035175879397,
371
+ "grad_norm": 12.103693008422852,
372
+ "learning_rate": 7.794285714285715e-06,
373
+ "loss": 3.9441,
374
+ "step": 1275
375
+ },
376
+ {
377
+ "epoch": 6.532663316582915,
378
+ "grad_norm": 14.731566429138184,
379
+ "learning_rate": 7.722857142857142e-06,
380
+ "loss": 3.9541,
381
+ "step": 1300
382
+ },
383
+ {
384
+ "epoch": 6.658291457286432,
385
+ "grad_norm": 13.810087203979492,
386
+ "learning_rate": 7.651428571428571e-06,
387
+ "loss": 3.962,
388
+ "step": 1325
389
+ },
390
+ {
391
+ "epoch": 6.78391959798995,
392
+ "grad_norm": 10.733597755432129,
393
+ "learning_rate": 7.58e-06,
394
+ "loss": 3.9696,
395
+ "step": 1350
396
+ },
397
+ {
398
+ "epoch": 6.909547738693467,
399
+ "grad_norm": 12.271906852722168,
400
+ "learning_rate": 7.508571428571429e-06,
401
+ "loss": 3.9319,
402
+ "step": 1375
403
+ },
404
+ {
405
+ "epoch": 7.035175879396985,
406
+ "grad_norm": 17.454214096069336,
407
+ "learning_rate": 7.4371428571428575e-06,
408
+ "loss": 3.9096,
409
+ "step": 1400
410
+ },
411
+ {
412
+ "epoch": 7.160804020100502,
413
+ "grad_norm": 9.891731262207031,
414
+ "learning_rate": 7.365714285714286e-06,
415
+ "loss": 3.963,
416
+ "step": 1425
417
+ },
418
+ {
419
+ "epoch": 7.28643216080402,
420
+ "grad_norm": 10.400883674621582,
421
+ "learning_rate": 7.294285714285715e-06,
422
+ "loss": 3.9529,
423
+ "step": 1450
424
+ },
425
+ {
426
+ "epoch": 7.4120603015075375,
427
+ "grad_norm": 8.886434555053711,
428
+ "learning_rate": 7.222857142857144e-06,
429
+ "loss": 3.9048,
430
+ "step": 1475
431
+ },
432
+ {
433
+ "epoch": 7.5376884422110555,
434
+ "grad_norm": 13.72711181640625,
435
+ "learning_rate": 7.151428571428573e-06,
436
+ "loss": 3.9097,
437
+ "step": 1500
438
+ },
439
+ {
440
+ "epoch": 7.663316582914573,
441
+ "grad_norm": 10.161803245544434,
442
+ "learning_rate": 7.08e-06,
443
+ "loss": 3.9029,
444
+ "step": 1525
445
+ },
446
+ {
447
+ "epoch": 7.788944723618091,
448
+ "grad_norm": 9.717086791992188,
449
+ "learning_rate": 7.008571428571429e-06,
450
+ "loss": 3.9028,
451
+ "step": 1550
452
+ },
453
+ {
454
+ "epoch": 7.914572864321608,
455
+ "grad_norm": 9.483698844909668,
456
+ "learning_rate": 6.937142857142858e-06,
457
+ "loss": 3.9157,
458
+ "step": 1575
459
+ },
460
+ {
461
+ "epoch": 8.040201005025125,
462
+ "grad_norm": 8.951908111572266,
463
+ "learning_rate": 6.865714285714287e-06,
464
+ "loss": 3.9106,
465
+ "step": 1600
466
+ },
467
+ {
468
+ "epoch": 8.165829145728644,
469
+ "grad_norm": 8.099654197692871,
470
+ "learning_rate": 6.794285714285714e-06,
471
+ "loss": 3.9062,
472
+ "step": 1625
473
+ },
474
+ {
475
+ "epoch": 8.291457286432161,
476
+ "grad_norm": 10.005452156066895,
477
+ "learning_rate": 6.722857142857143e-06,
478
+ "loss": 3.9139,
479
+ "step": 1650
480
+ },
481
+ {
482
+ "epoch": 8.417085427135678,
483
+ "grad_norm": 7.832030296325684,
484
+ "learning_rate": 6.651428571428572e-06,
485
+ "loss": 3.8731,
486
+ "step": 1675
487
+ },
488
+ {
489
+ "epoch": 8.542713567839195,
490
+ "grad_norm": 9.519272804260254,
491
+ "learning_rate": 6.5800000000000005e-06,
492
+ "loss": 3.9234,
493
+ "step": 1700
494
+ },
495
+ {
496
+ "epoch": 8.668341708542714,
497
+ "grad_norm": 7.6004252433776855,
498
+ "learning_rate": 6.5085714285714295e-06,
499
+ "loss": 3.8889,
500
+ "step": 1725
501
+ },
502
+ {
503
+ "epoch": 8.793969849246231,
504
+ "grad_norm": 8.43916130065918,
505
+ "learning_rate": 6.437142857142858e-06,
506
+ "loss": 3.8813,
507
+ "step": 1750
508
+ },
509
+ {
510
+ "epoch": 8.919597989949748,
511
+ "grad_norm": 9.573629379272461,
512
+ "learning_rate": 6.365714285714286e-06,
513
+ "loss": 3.9234,
514
+ "step": 1775
515
+ },
516
+ {
517
+ "epoch": 9.045226130653266,
518
+ "grad_norm": 11.034466743469238,
519
+ "learning_rate": 6.294285714285715e-06,
520
+ "loss": 3.9227,
521
+ "step": 1800
522
+ },
523
+ {
524
+ "epoch": 9.170854271356784,
525
+ "grad_norm": 9.65125560760498,
526
+ "learning_rate": 6.222857142857144e-06,
527
+ "loss": 3.8806,
528
+ "step": 1825
529
+ },
530
+ {
531
+ "epoch": 9.296482412060302,
532
+ "grad_norm": 8.907113075256348,
533
+ "learning_rate": 6.151428571428571e-06,
534
+ "loss": 3.8878,
535
+ "step": 1850
536
+ },
537
+ {
538
+ "epoch": 9.422110552763819,
539
+ "grad_norm": 9.000490188598633,
540
+ "learning_rate": 6.08e-06,
541
+ "loss": 3.858,
542
+ "step": 1875
543
+ },
544
+ {
545
+ "epoch": 9.547738693467338,
546
+ "grad_norm": 10.825636863708496,
547
+ "learning_rate": 6.008571428571429e-06,
548
+ "loss": 3.8546,
549
+ "step": 1900
550
+ },
551
+ {
552
+ "epoch": 9.673366834170855,
553
+ "grad_norm": 8.722935676574707,
554
+ "learning_rate": 5.937142857142858e-06,
555
+ "loss": 3.871,
556
+ "step": 1925
557
+ },
558
+ {
559
+ "epoch": 9.798994974874372,
560
+ "grad_norm": 9.243980407714844,
561
+ "learning_rate": 5.865714285714286e-06,
562
+ "loss": 3.8486,
563
+ "step": 1950
564
+ },
565
+ {
566
+ "epoch": 9.924623115577889,
567
+ "grad_norm": 10.595390319824219,
568
+ "learning_rate": 5.794285714285715e-06,
569
+ "loss": 3.8973,
570
+ "step": 1975
571
+ },
572
+ {
573
+ "epoch": 10.050251256281408,
574
+ "grad_norm": 11.8334321975708,
575
+ "learning_rate": 5.722857142857144e-06,
576
+ "loss": 3.8124,
577
+ "step": 2000
578
+ },
579
+ {
580
+ "epoch": 10.050251256281408,
581
+ "eval_loss": 0.44917792081832886,
582
+ "eval_runtime": 76.9989,
583
+ "eval_samples_per_second": 9.195,
584
+ "eval_steps_per_second": 4.597,
585
+ "step": 2000
586
+ },
587
+ {
588
+ "epoch": 10.175879396984925,
589
+ "grad_norm": 12.580653190612793,
590
+ "learning_rate": 5.651428571428572e-06,
591
+ "loss": 3.8754,
592
+ "step": 2025
593
+ },
594
+ {
595
+ "epoch": 10.301507537688442,
596
+ "grad_norm": 9.411772727966309,
597
+ "learning_rate": 5.582857142857143e-06,
598
+ "loss": 3.839,
599
+ "step": 2050
600
+ },
601
+ {
602
+ "epoch": 10.42713567839196,
603
+ "grad_norm": 8.65829849243164,
604
+ "learning_rate": 5.511428571428572e-06,
605
+ "loss": 3.8123,
606
+ "step": 2075
607
+ },
608
+ {
609
+ "epoch": 10.552763819095478,
610
+ "grad_norm": 9.79350757598877,
611
+ "learning_rate": 5.4400000000000004e-06,
612
+ "loss": 3.8609,
613
+ "step": 2100
614
+ },
615
+ {
616
+ "epoch": 10.678391959798995,
617
+ "grad_norm": 9.445945739746094,
618
+ "learning_rate": 5.368571428571429e-06,
619
+ "loss": 3.8359,
620
+ "step": 2125
621
+ },
622
+ {
623
+ "epoch": 10.804020100502512,
624
+ "grad_norm": 11.811120986938477,
625
+ "learning_rate": 5.297142857142858e-06,
626
+ "loss": 3.8632,
627
+ "step": 2150
628
+ },
629
+ {
630
+ "epoch": 10.92964824120603,
631
+ "grad_norm": 10.508607864379883,
632
+ "learning_rate": 5.225714285714286e-06,
633
+ "loss": 3.8324,
634
+ "step": 2175
635
+ },
636
+ {
637
+ "epoch": 11.055276381909549,
638
+ "grad_norm": 8.808614730834961,
639
+ "learning_rate": 5.154285714285715e-06,
640
+ "loss": 3.8314,
641
+ "step": 2200
642
+ },
643
+ {
644
+ "epoch": 11.180904522613066,
645
+ "grad_norm": 11.709993362426758,
646
+ "learning_rate": 5.082857142857144e-06,
647
+ "loss": 3.8904,
648
+ "step": 2225
649
+ },
650
+ {
651
+ "epoch": 11.306532663316583,
652
+ "grad_norm": 8.451616287231445,
653
+ "learning_rate": 5.011428571428571e-06,
654
+ "loss": 3.8471,
655
+ "step": 2250
656
+ },
657
+ {
658
+ "epoch": 11.4321608040201,
659
+ "grad_norm": 10.084294319152832,
660
+ "learning_rate": 4.94e-06,
661
+ "loss": 3.8881,
662
+ "step": 2275
663
+ },
664
+ {
665
+ "epoch": 11.557788944723619,
666
+ "grad_norm": 17.51856803894043,
667
+ "learning_rate": 4.868571428571429e-06,
668
+ "loss": 3.8931,
669
+ "step": 2300
670
+ },
671
+ {
672
+ "epoch": 11.683417085427136,
673
+ "grad_norm": 11.671639442443848,
674
+ "learning_rate": 4.797142857142857e-06,
675
+ "loss": 3.8578,
676
+ "step": 2325
677
+ },
678
+ {
679
+ "epoch": 11.809045226130653,
680
+ "grad_norm": 8.713754653930664,
681
+ "learning_rate": 4.725714285714286e-06,
682
+ "loss": 3.832,
683
+ "step": 2350
684
+ },
685
+ {
686
+ "epoch": 11.93467336683417,
687
+ "grad_norm": 11.429880142211914,
688
+ "learning_rate": 4.6542857142857145e-06,
689
+ "loss": 3.8556,
690
+ "step": 2375
691
+ },
692
+ {
693
+ "epoch": 12.06030150753769,
694
+ "grad_norm": 9.178875923156738,
695
+ "learning_rate": 4.5828571428571435e-06,
696
+ "loss": 3.8361,
697
+ "step": 2400
698
+ },
699
+ {
700
+ "epoch": 12.185929648241206,
701
+ "grad_norm": 7.655440330505371,
702
+ "learning_rate": 4.511428571428572e-06,
703
+ "loss": 3.8337,
704
+ "step": 2425
705
+ },
706
+ {
707
+ "epoch": 12.311557788944723,
708
+ "grad_norm": 10.44965648651123,
709
+ "learning_rate": 4.440000000000001e-06,
710
+ "loss": 3.877,
711
+ "step": 2450
712
+ },
713
+ {
714
+ "epoch": 12.43718592964824,
715
+ "grad_norm": 9.04977035522461,
716
+ "learning_rate": 4.368571428571429e-06,
717
+ "loss": 3.8504,
718
+ "step": 2475
719
+ },
720
+ {
721
+ "epoch": 12.56281407035176,
722
+ "grad_norm": 8.15368366241455,
723
+ "learning_rate": 4.297142857142858e-06,
724
+ "loss": 3.8297,
725
+ "step": 2500
726
+ },
727
+ {
728
+ "epoch": 12.688442211055277,
729
+ "grad_norm": 9.811213493347168,
730
+ "learning_rate": 4.225714285714286e-06,
731
+ "loss": 3.8289,
732
+ "step": 2525
733
+ },
734
+ {
735
+ "epoch": 12.814070351758794,
736
+ "grad_norm": 8.397767066955566,
737
+ "learning_rate": 4.154285714285714e-06,
738
+ "loss": 3.8352,
739
+ "step": 2550
740
+ },
741
+ {
742
+ "epoch": 12.93969849246231,
743
+ "grad_norm": 10.471597671508789,
744
+ "learning_rate": 4.082857142857143e-06,
745
+ "loss": 3.8438,
746
+ "step": 2575
747
+ },
748
+ {
749
+ "epoch": 13.06532663316583,
750
+ "grad_norm": 8.042658805847168,
751
+ "learning_rate": 4.011428571428571e-06,
752
+ "loss": 3.8308,
753
+ "step": 2600
754
+ },
755
+ {
756
+ "epoch": 13.190954773869347,
757
+ "grad_norm": 9.240007400512695,
758
+ "learning_rate": 3.94e-06,
759
+ "loss": 3.811,
760
+ "step": 2625
761
+ },
762
+ {
763
+ "epoch": 13.316582914572864,
764
+ "grad_norm": 8.514374732971191,
765
+ "learning_rate": 3.8685714285714286e-06,
766
+ "loss": 3.8253,
767
+ "step": 2650
768
+ },
769
+ {
770
+ "epoch": 13.442211055276381,
771
+ "grad_norm": 8.689332008361816,
772
+ "learning_rate": 3.7971428571428576e-06,
773
+ "loss": 3.8387,
774
+ "step": 2675
775
+ },
776
+ {
777
+ "epoch": 13.5678391959799,
778
+ "grad_norm": 11.250374794006348,
779
+ "learning_rate": 3.7257142857142857e-06,
780
+ "loss": 3.8216,
781
+ "step": 2700
782
+ },
783
+ {
784
+ "epoch": 13.693467336683417,
785
+ "grad_norm": 8.129416465759277,
786
+ "learning_rate": 3.6542857142857148e-06,
787
+ "loss": 3.8273,
788
+ "step": 2725
789
+ },
790
+ {
791
+ "epoch": 13.819095477386934,
792
+ "grad_norm": 9.398397445678711,
793
+ "learning_rate": 3.582857142857143e-06,
794
+ "loss": 3.8116,
795
+ "step": 2750
796
+ },
797
+ {
798
+ "epoch": 13.944723618090451,
799
+ "grad_norm": 9.966887474060059,
800
+ "learning_rate": 3.511428571428572e-06,
801
+ "loss": 3.8401,
802
+ "step": 2775
803
+ },
804
+ {
805
+ "epoch": 14.07035175879397,
806
+ "grad_norm": 7.829891204833984,
807
+ "learning_rate": 3.44e-06,
808
+ "loss": 3.7607,
809
+ "step": 2800
810
+ },
811
+ {
812
+ "epoch": 14.195979899497488,
813
+ "grad_norm": 10.33727741241455,
814
+ "learning_rate": 3.3685714285714287e-06,
815
+ "loss": 3.8406,
816
+ "step": 2825
817
+ },
818
+ {
819
+ "epoch": 14.321608040201005,
820
+ "grad_norm": 8.579668998718262,
821
+ "learning_rate": 3.2971428571428577e-06,
822
+ "loss": 3.7893,
823
+ "step": 2850
824
+ },
825
+ {
826
+ "epoch": 14.447236180904522,
827
+ "grad_norm": 9.909623146057129,
828
+ "learning_rate": 3.225714285714286e-06,
829
+ "loss": 3.8661,
830
+ "step": 2875
831
+ },
832
+ {
833
+ "epoch": 14.57286432160804,
834
+ "grad_norm": 9.630623817443848,
835
+ "learning_rate": 3.154285714285715e-06,
836
+ "loss": 3.8248,
837
+ "step": 2900
838
+ },
839
+ {
840
+ "epoch": 14.698492462311558,
841
+ "grad_norm": 13.711705207824707,
842
+ "learning_rate": 3.082857142857143e-06,
843
+ "loss": 3.8268,
844
+ "step": 2925
845
+ },
846
+ {
847
+ "epoch": 14.824120603015075,
848
+ "grad_norm": 8.66063117980957,
849
+ "learning_rate": 3.0114285714285716e-06,
850
+ "loss": 3.8154,
851
+ "step": 2950
852
+ },
853
+ {
854
+ "epoch": 14.949748743718594,
855
+ "grad_norm": 10.465475082397461,
856
+ "learning_rate": 2.9400000000000002e-06,
857
+ "loss": 3.826,
858
+ "step": 2975
859
+ },
860
+ {
861
+ "epoch": 15.075376884422111,
862
+ "grad_norm": 9.025050163269043,
863
+ "learning_rate": 2.868571428571429e-06,
864
+ "loss": 3.8132,
865
+ "step": 3000
866
+ },
867
+ {
868
+ "epoch": 15.075376884422111,
869
+ "eval_loss": 0.44469836354255676,
870
+ "eval_runtime": 78.3312,
871
+ "eval_samples_per_second": 9.039,
872
+ "eval_steps_per_second": 4.519,
873
+ "step": 3000
874
+ }
875
+ ],
876
+ "logging_steps": 25,
877
+ "max_steps": 4000,
878
+ "num_input_tokens_seen": 0,
879
+ "num_train_epochs": 21,
880
+ "save_steps": 1000,
881
+ "stateful_callbacks": {
882
+ "TrainerControl": {
883
+ "args": {
884
+ "should_epoch_stop": false,
885
+ "should_evaluate": false,
886
+ "should_log": false,
887
+ "should_save": true,
888
+ "should_training_stop": false
889
+ },
890
+ "attributes": {}
891
+ }
892
+ },
893
+ "total_flos": 1.3502296824093e+16,
894
+ "train_batch_size": 4,
895
+ "trial_name": null,
896
+ "trial_params": null
897
+ }
checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c6ae756af4eb40168228ea5b672dedd9146f11deae4f8d64668e0b2d5a69781
3
+ size 5432
checkpoint-4000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-4000/config.json ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/speecht5_tts",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "SpeechT5ForTextToSpeech"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 0,
10
+ "conv_bias": false,
11
+ "conv_dim": [
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "conv_kernel": [
21
+ 10,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 2,
27
+ 2
28
+ ],
29
+ "conv_stride": [
30
+ 5,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2
37
+ ],
38
+ "decoder_attention_heads": 12,
39
+ "decoder_ffn_dim": 3072,
40
+ "decoder_layerdrop": 0.1,
41
+ "decoder_layers": 6,
42
+ "decoder_start_token_id": 2,
43
+ "encoder_attention_heads": 12,
44
+ "encoder_ffn_dim": 3072,
45
+ "encoder_layerdrop": 0.1,
46
+ "encoder_layers": 12,
47
+ "encoder_max_relative_position": 160,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_norm": "group",
51
+ "feat_proj_dropout": 0.0,
52
+ "guided_attention_loss_num_heads": 2,
53
+ "guided_attention_loss_scale": 10.0,
54
+ "guided_attention_loss_sigma": 0.4,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 768,
58
+ "initializer_range": 0.02,
59
+ "is_encoder_decoder": true,
60
+ "layer_norm_eps": 1e-05,
61
+ "mask_feature_length": 10,
62
+ "mask_feature_min_masks": 0,
63
+ "mask_feature_prob": 0.0,
64
+ "mask_time_length": 10,
65
+ "mask_time_min_masks": 2,
66
+ "mask_time_prob": 0.05,
67
+ "max_length": null,
68
+ "max_speech_positions": 1876,
69
+ "max_text_positions": 600,
70
+ "model_type": "speecht5",
71
+ "num_conv_pos_embedding_groups": 16,
72
+ "num_conv_pos_embeddings": 128,
73
+ "num_feat_extract_layers": 7,
74
+ "num_mel_bins": 80,
75
+ "pad_token_id": 1,
76
+ "positional_dropout": 0.1,
77
+ "reduction_factor": 2,
78
+ "scale_embedding": false,
79
+ "speaker_embedding_dim": 512,
80
+ "speech_decoder_postnet_dropout": 0.5,
81
+ "speech_decoder_postnet_kernel": 5,
82
+ "speech_decoder_postnet_layers": 5,
83
+ "speech_decoder_postnet_units": 256,
84
+ "speech_decoder_prenet_dropout": 0.5,
85
+ "speech_decoder_prenet_layers": 2,
86
+ "speech_decoder_prenet_units": 256,
87
+ "torch_dtype": "float32",
88
+ "transformers_version": "4.46.1",
89
+ "use_cache": false,
90
+ "use_guided_attention_loss": true,
91
+ "vocab_size": 81
92
+ }
checkpoint-4000/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "decoder_start_token_id": 2,
5
+ "eos_token_id": 2,
6
+ "max_length": 1876,
7
+ "pad_token_id": 1,
8
+ "transformers_version": "4.46.1"
9
+ }
checkpoint-4000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eef602e8e24bc81f46de9f25767181905eb11eef70a202f911f339f87f352347
3
+ size 577789320
checkpoint-4000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8420507944acab9b5096619d2ae70286620e694ab3f8a5679789b8e6eca4260
3
+ size 1155772233
checkpoint-4000/preprocessor_config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": false,
3
+ "feature_extractor_type": "SpeechT5FeatureExtractor",
4
+ "feature_size": 1,
5
+ "fmax": 7600,
6
+ "fmin": 80,
7
+ "frame_signal_scale": 1.0,
8
+ "hop_length": 16,
9
+ "mel_floor": 1e-10,
10
+ "num_mel_bins": 80,
11
+ "padding_side": "right",
12
+ "padding_value": 0.0,
13
+ "processor_class": "SpeechT5Processor",
14
+ "reduction_factor": 2,
15
+ "return_attention_mask": true,
16
+ "sampling_rate": 16000,
17
+ "win_function": "hann_window",
18
+ "win_length": 64
19
+ }
checkpoint-4000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f46389eb0872e6838f448d5803d25f4f567ce04f0cad9805ddd1e11b37f2317
3
+ size 14244
checkpoint-4000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7efbf80617c96c78286826ce59d9a12c86da62d7631874b3d6364a8e993ada60
3
+ size 1064
checkpoint-4000/special_tokens_map.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "mask_token": {
5
+ "content": "<mask>",
6
+ "lstrip": true,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false
10
+ },
11
+ "pad_token": "<pad>",
12
+ "unk_token": "<unk>"
13
+ }
checkpoint-4000/spm_char.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7fcc48f3e225f627b1641db410ceb0c8649bd2b0c982e150b03f8be3728ab560
3
+ size 238473
checkpoint-4000/tokenizer_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "79": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": true,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "80": {
44
+ "content": "<ctc_blank>",
45
+ "lstrip": false,
46
+ "normalized": true,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": false
50
+ }
51
+ },
52
+ "bos_token": "<s>",
53
+ "clean_up_tokenization_spaces": false,
54
+ "eos_token": "</s>",
55
+ "mask_token": "<mask>",
56
+ "model_max_length": 600,
57
+ "normalize": false,
58
+ "pad_token": "<pad>",
59
+ "processor_class": "SpeechT5Processor",
60
+ "sp_model_kwargs": {},
61
+ "tokenizer_class": "SpeechT5Tokenizer",
62
+ "unk_token": "<unk>"
63
+ }