Hugging Face's logo

froabera
/
tts_processed

Safetensors
speecht5
Model card Files
xet
 Community
froabera commited on
Commit
44efa2c
·
verified ·
1 Parent(s): ca23884

Training in progress, step 1000, checkpoint

Browse files
Files changed (9) hide show
  1. checkpoint-1000/config.json +91 -0
  2. checkpoint-1000/generation_config.json +11 -0
  3. checkpoint-1000/model.safetensors +3 -0
  4. checkpoint-1000/optimizer.pt +3 -0
  5. checkpoint-1000/rng_state.pth +3 -0
  6. checkpoint-1000/scaler.pt +3 -0
  7. checkpoint-1000/scheduler.pt +3 -0
  8. checkpoint-1000/trainer_state.json +339 -0
  9. checkpoint-1000/training_args.bin +3 -0
checkpoint-1000/config.json ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.1,
3
+ "apply_spec_augment": true,
4
+ "architectures": [
5
+ "SpeechT5ForTextToSpeech"
6
+ ],
7
+ "attention_dropout": 0.1,
8
+ "bos_token_id": 0,
9
+ "conv_bias": false,
10
+ "conv_dim": [
11
+ 512,
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512
18
+ ],
19
+ "conv_kernel": [
20
+ 10,
21
+ 3,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 2,
26
+ 2
27
+ ],
28
+ "conv_stride": [
29
+ 5,
30
+ 2,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2
36
+ ],
37
+ "decoder_attention_heads": 12,
38
+ "decoder_ffn_dim": 3072,
39
+ "decoder_layerdrop": 0.1,
40
+ "decoder_layers": 6,
41
+ "decoder_start_token_id": 2,
42
+ "dtype": "float32",
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_speech_positions": 1876,
68
+ "max_text_positions": 600,
69
+ "model_type": "speecht5",
70
+ "num_conv_pos_embedding_groups": 16,
71
+ "num_conv_pos_embeddings": 128,
72
+ "num_feat_extract_layers": 7,
73
+ "num_mel_bins": 80,
74
+ "pad_token_id": 1,
75
+ "positional_dropout": 0.1,
76
+ "reduction_factor": 2,
77
+ "scale_embedding": false,
78
+ "speaker_embedding_dim": 512,
79
+ "speech_decoder_postnet_dropout": 0.5,
80
+ "speech_decoder_postnet_kernel": 5,
81
+ "speech_decoder_postnet_layers": 5,
82
+ "speech_decoder_postnet_units": 256,
83
+ "speech_decoder_prenet_dropout": 0.5,
84
+ "speech_decoder_prenet_layers": 2,
85
+ "speech_decoder_prenet_units": 256,
86
+ "tie_word_embeddings": true,
87
+ "transformers_version": "5.19.0.dev0",
88
+ "use_cache": false,
89
+ "use_guided_attention_loss": false,
90
+ "vocab_size": 81
91
+ }
checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "decoder_start_token_id": 2,
5
+ "eos_token_id": 2,
6
+ "output_attentions": false,
7
+ "output_hidden_states": false,
8
+ "pad_token_id": 1,
9
+ "transformers_version": "5.19.0.dev0",
10
+ "use_cache": true
11
+ }
checkpoint-1000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61b9399cd0c6f80046b8b3c1e822d099a1959b002e8abd26abd8a7344677340e
3
+ size 577789320
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:333fdcff07242722fde6dbd5cb1bca85e35126b329ae689db6e33a776eb9ce07
3
+ size 1155772634
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c63cef3dd020eb7615b466004946593076b764ad5bfa0f7e30c6614328be0b0
3
+ size 14645
checkpoint-1000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14ae2a2128444abab378aa06c09a61a84665f758fcc19fc46f5789b0bc1b5665
3
+ size 1383
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd9282238785ccb0a449bb8b75998b539c60bece70aa84e176934685cf98fe15
3
+ size 1465
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,339 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1000,
3
+ "best_metric": 1.01394784450531,
4
+ "best_model_checkpoint": "/kaggle/working/tts_processed/checkpoint-1000",
5
+ "epoch": 4.951083591331269,
6
+ "eval_steps": 500,
7
+ "global_step": 1000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.1238390092879257,
14
+ "grad_norm": 829.4996337890625,
15
+ "learning_rate": 4.8e-08,
16
+ "loss": 21.73699951171875,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.2476780185758514,
21
+ "grad_norm": 701.2330932617188,
22
+ "learning_rate": 9.8e-08,
23
+ "loss": 21.00220947265625,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.3715170278637771,
28
+ "grad_norm": 631.0809326171875,
29
+ "learning_rate": 1.4799999999999998e-07,
30
+ "loss": 19.772623291015623,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.4953560371517028,
35
+ "grad_norm": 556.35791015625,
36
+ "learning_rate": 1.98e-07,
37
+ "loss": 18.46118896484375,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.6191950464396285,
42
+ "grad_norm": 536.2819213867188,
43
+ "learning_rate": 2.48e-07,
44
+ "loss": 17.51901611328125,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.7430340557275542,
49
+ "grad_norm": 459.290283203125,
50
+ "learning_rate": 2.98e-07,
51
+ "loss": 16.43676513671875,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.8668730650154799,
56
+ "grad_norm": 230.56405639648438,
57
+ "learning_rate": 3.4799999999999994e-07,
58
+ "loss": 15.75008056640625,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 0.9907120743034056,
63
+ "grad_norm": 233.90635681152344,
64
+ "learning_rate": 3.98e-07,
65
+ "loss": 14.56845947265625,
66
+ "step": 200
67
+ },
68
+ {
69
+ "epoch": 1.1139318885448917,
70
+ "grad_norm": 107.61080169677734,
71
+ "learning_rate": 4.48e-07,
72
+ "loss": 13.99685546875,
73
+ "step": 225
74
+ },
75
+ {
76
+ "epoch": 1.2377708978328172,
77
+ "grad_norm": 115.48291015625,
78
+ "learning_rate": 4.979999999999999e-07,
79
+ "loss": 13.390751953125,
80
+ "step": 250
81
+ },
82
+ {
83
+ "epoch": 1.361609907120743,
84
+ "grad_norm": 150.88845825195312,
85
+ "learning_rate": 5.48e-07,
86
+ "loss": 12.885958251953125,
87
+ "step": 275
88
+ },
89
+ {
90
+ "epoch": 1.4854489164086688,
91
+ "grad_norm": 102.25688934326172,
92
+ "learning_rate": 5.979999999999999e-07,
93
+ "loss": 12.281358642578125,
94
+ "step": 300
95
+ },
96
+ {
97
+ "epoch": 1.6092879256965944,
98
+ "grad_norm": 91.96704864501953,
99
+ "learning_rate": 6.48e-07,
100
+ "loss": 12.51485107421875,
101
+ "step": 325
102
+ },
103
+ {
104
+ "epoch": 1.73312693498452,
105
+ "grad_norm": 89.19316864013672,
106
+ "learning_rate": 6.979999999999999e-07,
107
+ "loss": 11.758416748046875,
108
+ "step": 350
109
+ },
110
+ {
111
+ "epoch": 1.8569659442724458,
112
+ "grad_norm": 43.06995391845703,
113
+ "learning_rate": 7.48e-07,
114
+ "loss": 11.7668603515625,
115
+ "step": 375
116
+ },
117
+ {
118
+ "epoch": 1.9808049535603716,
119
+ "grad_norm": 37.787078857421875,
120
+ "learning_rate": 7.98e-07,
121
+ "loss": 11.696016845703125,
122
+ "step": 400
123
+ },
124
+ {
125
+ "epoch": 2.1040247678018575,
126
+ "grad_norm": 52.63007354736328,
127
+ "learning_rate": 8.48e-07,
128
+ "loss": 11.60255859375,
129
+ "step": 425
130
+ },
131
+ {
132
+ "epoch": 2.2278637770897833,
133
+ "grad_norm": 37.73246383666992,
134
+ "learning_rate": 8.98e-07,
135
+ "loss": 11.121099853515625,
136
+ "step": 450
137
+ },
138
+ {
139
+ "epoch": 2.351702786377709,
140
+ "grad_norm": 50.654396057128906,
141
+ "learning_rate": 9.479999999999999e-07,
142
+ "loss": 10.981490478515624,
143
+ "step": 475
144
+ },
145
+ {
146
+ "epoch": 2.4755417956656345,
147
+ "grad_norm": 34.40877151489258,
148
+ "learning_rate": 9.98e-07,
149
+ "loss": 10.89916015625,
150
+ "step": 500
151
+ },
152
+ {
153
+ "epoch": 2.4755417956656345,
154
+ "eval_loss": 1.2403377294540405,
155
+ "eval_runtime": 78.2214,
156
+ "eval_samples_per_second": 20.659,
157
+ "eval_steps_per_second": 5.165,
158
+ "step": 500
159
+ },
160
+ {
161
+ "epoch": 2.5993808049535603,
162
+ "grad_norm": 32.53867721557617,
163
+ "learning_rate": 9.986285714285714e-07,
164
+ "loss": 10.70981201171875,
165
+ "step": 525
166
+ },
167
+ {
168
+ "epoch": 2.723219814241486,
169
+ "grad_norm": 37.28243637084961,
170
+ "learning_rate": 9.972e-07,
171
+ "loss": 10.68586669921875,
172
+ "step": 550
173
+ },
174
+ {
175
+ "epoch": 2.847058823529412,
176
+ "grad_norm": 33.349178314208984,
177
+ "learning_rate": 9.957714285714285e-07,
178
+ "loss": 10.5280859375,
179
+ "step": 575
180
+ },
181
+ {
182
+ "epoch": 2.9708978328173377,
183
+ "grad_norm": 40.658817291259766,
184
+ "learning_rate": 9.94342857142857e-07,
185
+ "loss": 10.160640258789062,
186
+ "step": 600
187
+ },
188
+ {
189
+ "epoch": 3.0941176470588236,
190
+ "grad_norm": 35.320701599121094,
191
+ "learning_rate": 9.929142857142856e-07,
192
+ "loss": 10.215652465820312,
193
+ "step": 625
194
+ },
195
+ {
196
+ "epoch": 3.2179566563467494,
197
+ "grad_norm": 31.319137573242188,
198
+ "learning_rate": 9.914857142857142e-07,
199
+ "loss": 10.158941650390625,
200
+ "step": 650
201
+ },
202
+ {
203
+ "epoch": 3.341795665634675,
204
+ "grad_norm": 28.438013076782227,
205
+ "learning_rate": 9.900571428571427e-07,
206
+ "loss": 9.933291625976562,
207
+ "step": 675
208
+ },
209
+ {
210
+ "epoch": 3.4656346749226006,
211
+ "grad_norm": 41.65836715698242,
212
+ "learning_rate": 9.886285714285713e-07,
213
+ "loss": 9.679817504882813,
214
+ "step": 700
215
+ },
216
+ {
217
+ "epoch": 3.5894736842105264,
218
+ "grad_norm": 49.5134162902832,
219
+ "learning_rate": 9.871999999999998e-07,
220
+ "loss": 9.66293212890625,
221
+ "step": 725
222
+ },
223
+ {
224
+ "epoch": 3.713312693498452,
225
+ "grad_norm": 37.431331634521484,
226
+ "learning_rate": 9.857714285714284e-07,
227
+ "loss": 9.642633056640625,
228
+ "step": 750
229
+ },
230
+ {
231
+ "epoch": 3.837151702786378,
232
+ "grad_norm": 39.974544525146484,
233
+ "learning_rate": 9.843428571428572e-07,
234
+ "loss": 9.296731567382812,
235
+ "step": 775
236
+ },
237
+ {
238
+ "epoch": 3.9609907120743033,
239
+ "grad_norm": 39.033599853515625,
240
+ "learning_rate": 9.829142857142857e-07,
241
+ "loss": 9.472877807617188,
242
+ "step": 800
243
+ },
244
+ {
245
+ "epoch": 4.08421052631579,
246
+ "grad_norm": 29.367778778076172,
247
+ "learning_rate": 9.814857142857143e-07,
248
+ "loss": 9.286629028320313,
249
+ "step": 825
250
+ },
251
+ {
252
+ "epoch": 4.208049535603715,
253
+ "grad_norm": 30.283451080322266,
254
+ "learning_rate": 9.800571428571428e-07,
255
+ "loss": 9.21810546875,
256
+ "step": 850
257
+ },
258
+ {
259
+ "epoch": 4.33188854489164,
260
+ "grad_norm": 43.98243713378906,
261
+ "learning_rate": 9.786285714285714e-07,
262
+ "loss": 9.4026904296875,
263
+ "step": 875
264
+ },
265
+ {
266
+ "epoch": 4.455727554179567,
267
+ "grad_norm": 35.94754409790039,
268
+ "learning_rate": 9.772e-07,
269
+ "loss": 9.243385009765625,
270
+ "step": 900
271
+ },
272
+ {
273
+ "epoch": 4.579566563467492,
274
+ "grad_norm": 97.34254455566406,
275
+ "learning_rate": 9.757714285714285e-07,
276
+ "loss": 9.081926879882813,
277
+ "step": 925
278
+ },
279
+ {
280
+ "epoch": 4.703405572755418,
281
+ "grad_norm": 35.22529602050781,
282
+ "learning_rate": 9.74342857142857e-07,
283
+ "loss": 9.177518310546875,
284
+ "step": 950
285
+ },
286
+ {
287
+ "epoch": 4.827244582043344,
288
+ "grad_norm": 31.761302947998047,
289
+ "learning_rate": 9.729142857142856e-07,
290
+ "loss": 9.266145629882812,
291
+ "step": 975
292
+ },
293
+ {
294
+ "epoch": 4.951083591331269,
295
+ "grad_norm": 34.29395294189453,
296
+ "learning_rate": 9.714857142857142e-07,
297
+ "loss": 9.146566162109375,
298
+ "step": 1000
299
+ },
300
+ {
301
+ "epoch": 4.951083591331269,
302
+ "eval_loss": 1.01394784450531,
303
+ "eval_runtime": 78.015,
304
+ "eval_samples_per_second": 20.714,
305
+ "eval_steps_per_second": 5.178,
306
+ "step": 1000
307
+ }
308
+ ],
309
+ "logging_steps": 25,
310
+ "max_steps": 18000,
311
+ "num_input_tokens_seen": 0,
312
+ "num_train_epochs": 90,
313
+ "save_steps": 500,
314
+ "stateful_callbacks": {
315
+ "EarlyStoppingCallback": {
316
+ "args": {
317
+ "early_stopping_patience": 5,
318
+ "early_stopping_threshold": 0.0
319
+ },
320
+ "attributes": {
321
+ "early_stopping_patience_counter": 0
322
+ }
323
+ },
324
+ "TrainerControl": {
325
+ "args": {
326
+ "should_epoch_stop": false,
327
+ "should_evaluate": false,
328
+ "should_log": false,
329
+ "should_save": true,
330
+ "should_training_stop": false
331
+ },
332
+ "attributes": {}
333
+ }
334
+ },
335
+ "total_flos": 3883083840133344.0,
336
+ "train_batch_size": 4,
337
+ "trial_name": null,
338
+ "trial_params": null
339
+ }
checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc3ae1f3188fc78f14b1e24c35f0f902ff13c8475d4503f269d542bb33a09e53
3
+ size 5393