mikhail-panzo commited on
Commit
b5f5fa9
1 Parent(s): 84013ed

Training in progress, step 3500, checkpoint

Browse files
checkpoint-3500/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-3500/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": 1876,
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": "4.41.0.dev0",
89
+ "use_cache": false,
90
+ "use_guided_attention_loss": true,
91
+ "vocab_size": 81
92
+ }
checkpoint-3500/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": "4.41.0.dev0"
9
+ }
checkpoint-3500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b18c1afcf47ac44e97af90a4bec5f168e23ce8af7cc4bc18569db9bbda6fcf71
3
+ size 577789320
checkpoint-3500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ffa28b556bfbbe36bfa0dad17473965b27519741fca272893352a72b048c9541
3
+ size 1155772233
checkpoint-3500/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-3500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ac6e3558db29f7fa3b309fd2919b1e8c40f88b39183e6963403ddc7222dbfa2
3
+ size 14244
checkpoint-3500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa333443f5267d7c085c81e62ab3f19a14d2e18c44af1adf5a25e7c8f744eab9
3
+ size 1064
checkpoint-3500/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-3500/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-3500/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": true,
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-3500/trainer_state.json ADDED
@@ -0,0 +1,579 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.4463141858577728,
3
+ "best_model_checkpoint": "mikhail-panzo/fil_b32_le3_s4000/checkpoint-500",
4
+ "epoch": 77.77777777777777,
5
+ "eval_steps": 500,
6
+ "global_step": 3500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.1111111111111112,
13
+ "grad_norm": 3.548590898513794,
14
+ "learning_rate": 2.5e-05,
15
+ "loss": 0.7354,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 2.2222222222222223,
20
+ "grad_norm": 5.517350196838379,
21
+ "learning_rate": 4.9500000000000004e-05,
22
+ "loss": 0.5892,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 3.3333333333333335,
27
+ "grad_norm": 4.143002033233643,
28
+ "learning_rate": 7.45e-05,
29
+ "loss": 0.532,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 4.444444444444445,
34
+ "grad_norm": 6.414477825164795,
35
+ "learning_rate": 9.95e-05,
36
+ "loss": 0.5284,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 5.555555555555555,
41
+ "grad_norm": 2.265143394470215,
42
+ "learning_rate": 0.0001245,
43
+ "loss": 0.517,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 6.666666666666667,
48
+ "grad_norm": 4.954311847686768,
49
+ "learning_rate": 0.0001495,
50
+ "loss": 0.5295,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 7.777777777777778,
55
+ "grad_norm": 2.5993571281433105,
56
+ "learning_rate": 0.00017449999999999999,
57
+ "loss": 0.517,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 8.88888888888889,
62
+ "grad_norm": 3.27908992767334,
63
+ "learning_rate": 0.00019950000000000002,
64
+ "loss": 0.4939,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 10.0,
69
+ "grad_norm": 4.22811222076416,
70
+ "learning_rate": 0.0002245,
71
+ "loss": 0.5043,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 11.11111111111111,
76
+ "grad_norm": 2.7051336765289307,
77
+ "learning_rate": 0.0002495,
78
+ "loss": 0.5028,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 11.11111111111111,
83
+ "eval_loss": 0.4463141858577728,
84
+ "eval_runtime": 16.8864,
85
+ "eval_samples_per_second": 9.416,
86
+ "eval_steps_per_second": 1.184,
87
+ "step": 500
88
+ },
89
+ {
90
+ "epoch": 12.222222222222221,
91
+ "grad_norm": 3.0811643600463867,
92
+ "learning_rate": 0.0002745,
93
+ "loss": 0.545,
94
+ "step": 550
95
+ },
96
+ {
97
+ "epoch": 13.333333333333334,
98
+ "grad_norm": 7.707492828369141,
99
+ "learning_rate": 0.0002995,
100
+ "loss": 0.4979,
101
+ "step": 600
102
+ },
103
+ {
104
+ "epoch": 14.444444444444445,
105
+ "grad_norm": 5.85659122467041,
106
+ "learning_rate": 0.00032450000000000003,
107
+ "loss": 0.4916,
108
+ "step": 650
109
+ },
110
+ {
111
+ "epoch": 15.555555555555555,
112
+ "grad_norm": 8.300439834594727,
113
+ "learning_rate": 0.0003495,
114
+ "loss": 0.5112,
115
+ "step": 700
116
+ },
117
+ {
118
+ "epoch": 16.666666666666668,
119
+ "grad_norm": 10.664773941040039,
120
+ "learning_rate": 0.0003745,
121
+ "loss": 0.5312,
122
+ "step": 750
123
+ },
124
+ {
125
+ "epoch": 17.77777777777778,
126
+ "grad_norm": 6.4097161293029785,
127
+ "learning_rate": 0.0003995,
128
+ "loss": 0.6087,
129
+ "step": 800
130
+ },
131
+ {
132
+ "epoch": 18.88888888888889,
133
+ "grad_norm": 5.435739994049072,
134
+ "learning_rate": 0.0004245,
135
+ "loss": 0.6127,
136
+ "step": 850
137
+ },
138
+ {
139
+ "epoch": 20.0,
140
+ "grad_norm": 7.4649834632873535,
141
+ "learning_rate": 0.00044950000000000003,
142
+ "loss": 0.6904,
143
+ "step": 900
144
+ },
145
+ {
146
+ "epoch": 21.11111111111111,
147
+ "grad_norm": 10.81910228729248,
148
+ "learning_rate": 0.0004745,
149
+ "loss": 0.6948,
150
+ "step": 950
151
+ },
152
+ {
153
+ "epoch": 22.22222222222222,
154
+ "grad_norm": 4.4550886154174805,
155
+ "learning_rate": 0.0004995,
156
+ "loss": 0.6348,
157
+ "step": 1000
158
+ },
159
+ {
160
+ "epoch": 22.22222222222222,
161
+ "eval_loss": 0.5990382432937622,
162
+ "eval_runtime": 9.0191,
163
+ "eval_samples_per_second": 17.629,
164
+ "eval_steps_per_second": 2.218,
165
+ "step": 1000
166
+ },
167
+ {
168
+ "epoch": 23.333333333333332,
169
+ "grad_norm": 3.101961374282837,
170
+ "learning_rate": 0.0005245,
171
+ "loss": 0.6609,
172
+ "step": 1050
173
+ },
174
+ {
175
+ "epoch": 24.444444444444443,
176
+ "grad_norm": 5.606886386871338,
177
+ "learning_rate": 0.0005495,
178
+ "loss": 0.6128,
179
+ "step": 1100
180
+ },
181
+ {
182
+ "epoch": 25.555555555555557,
183
+ "grad_norm": 3.7152516841888428,
184
+ "learning_rate": 0.0005745,
185
+ "loss": 0.6197,
186
+ "step": 1150
187
+ },
188
+ {
189
+ "epoch": 26.666666666666668,
190
+ "grad_norm": 2.9999840259552,
191
+ "learning_rate": 0.0005995000000000001,
192
+ "loss": 0.6346,
193
+ "step": 1200
194
+ },
195
+ {
196
+ "epoch": 27.77777777777778,
197
+ "grad_norm": 3.5235953330993652,
198
+ "learning_rate": 0.0006245000000000001,
199
+ "loss": 0.6246,
200
+ "step": 1250
201
+ },
202
+ {
203
+ "epoch": 28.88888888888889,
204
+ "grad_norm": 3.9665956497192383,
205
+ "learning_rate": 0.0006495,
206
+ "loss": 0.6143,
207
+ "step": 1300
208
+ },
209
+ {
210
+ "epoch": 30.0,
211
+ "grad_norm": 2.920466661453247,
212
+ "learning_rate": 0.000674,
213
+ "loss": 0.6245,
214
+ "step": 1350
215
+ },
216
+ {
217
+ "epoch": 31.11111111111111,
218
+ "grad_norm": 16.839616775512695,
219
+ "learning_rate": 0.000699,
220
+ "loss": 0.8179,
221
+ "step": 1400
222
+ },
223
+ {
224
+ "epoch": 32.22222222222222,
225
+ "grad_norm": 0.23306678235530853,
226
+ "learning_rate": 0.0007235000000000001,
227
+ "loss": 1.8511,
228
+ "step": 1450
229
+ },
230
+ {
231
+ "epoch": 33.333333333333336,
232
+ "grad_norm": 0.6476005911827087,
233
+ "learning_rate": 0.000748,
234
+ "loss": 1.7172,
235
+ "step": 1500
236
+ },
237
+ {
238
+ "epoch": 33.333333333333336,
239
+ "eval_loss": 1.566503882408142,
240
+ "eval_runtime": 8.82,
241
+ "eval_samples_per_second": 18.027,
242
+ "eval_steps_per_second": 2.268,
243
+ "step": 1500
244
+ },
245
+ {
246
+ "epoch": 34.44444444444444,
247
+ "grad_norm": 0.22506840527057648,
248
+ "learning_rate": 0.000773,
249
+ "loss": 1.6751,
250
+ "step": 1550
251
+ },
252
+ {
253
+ "epoch": 35.55555555555556,
254
+ "grad_norm": 0.2123740017414093,
255
+ "learning_rate": 0.0007980000000000001,
256
+ "loss": 1.585,
257
+ "step": 1600
258
+ },
259
+ {
260
+ "epoch": 36.666666666666664,
261
+ "grad_norm": 0.2724120318889618,
262
+ "learning_rate": 0.000823,
263
+ "loss": 1.6075,
264
+ "step": 1650
265
+ },
266
+ {
267
+ "epoch": 37.77777777777778,
268
+ "grad_norm": 0.1595434546470642,
269
+ "learning_rate": 0.000848,
270
+ "loss": 1.5939,
271
+ "step": 1700
272
+ },
273
+ {
274
+ "epoch": 38.888888888888886,
275
+ "grad_norm": 0.329246461391449,
276
+ "learning_rate": 0.000873,
277
+ "loss": 1.6055,
278
+ "step": 1750
279
+ },
280
+ {
281
+ "epoch": 40.0,
282
+ "grad_norm": 0.5109756588935852,
283
+ "learning_rate": 0.000898,
284
+ "loss": 1.5925,
285
+ "step": 1800
286
+ },
287
+ {
288
+ "epoch": 41.111111111111114,
289
+ "grad_norm": 0.13621735572814941,
290
+ "learning_rate": 0.0009230000000000001,
291
+ "loss": 1.5823,
292
+ "step": 1850
293
+ },
294
+ {
295
+ "epoch": 42.22222222222222,
296
+ "grad_norm": 0.13570882380008698,
297
+ "learning_rate": 0.000948,
298
+ "loss": 1.5806,
299
+ "step": 1900
300
+ },
301
+ {
302
+ "epoch": 43.333333333333336,
303
+ "grad_norm": 0.10358592867851257,
304
+ "learning_rate": 0.000973,
305
+ "loss": 1.5762,
306
+ "step": 1950
307
+ },
308
+ {
309
+ "epoch": 44.44444444444444,
310
+ "grad_norm": 0.2059299349784851,
311
+ "learning_rate": 0.000998,
312
+ "loss": 1.5863,
313
+ "step": 2000
314
+ },
315
+ {
316
+ "epoch": 44.44444444444444,
317
+ "eval_loss": 1.5431957244873047,
318
+ "eval_runtime": 8.5158,
319
+ "eval_samples_per_second": 18.671,
320
+ "eval_steps_per_second": 2.349,
321
+ "step": 2000
322
+ },
323
+ {
324
+ "epoch": 45.55555555555556,
325
+ "grad_norm": 0.1365118771791458,
326
+ "learning_rate": 0.000977,
327
+ "loss": 1.578,
328
+ "step": 2050
329
+ },
330
+ {
331
+ "epoch": 46.666666666666664,
332
+ "grad_norm": 0.482020765542984,
333
+ "learning_rate": 0.0009519999999999999,
334
+ "loss": 1.5794,
335
+ "step": 2100
336
+ },
337
+ {
338
+ "epoch": 47.77777777777778,
339
+ "grad_norm": 0.0648043230175972,
340
+ "learning_rate": 0.0009270000000000001,
341
+ "loss": 1.6363,
342
+ "step": 2150
343
+ },
344
+ {
345
+ "epoch": 48.888888888888886,
346
+ "grad_norm": 0.11886011809110641,
347
+ "learning_rate": 0.000902,
348
+ "loss": 1.5797,
349
+ "step": 2200
350
+ },
351
+ {
352
+ "epoch": 50.0,
353
+ "grad_norm": 0.35273927450180054,
354
+ "learning_rate": 0.0008770000000000001,
355
+ "loss": 1.5799,
356
+ "step": 2250
357
+ },
358
+ {
359
+ "epoch": 51.111111111111114,
360
+ "grad_norm": 0.3486323356628418,
361
+ "learning_rate": 0.000852,
362
+ "loss": 1.5786,
363
+ "step": 2300
364
+ },
365
+ {
366
+ "epoch": 52.22222222222222,
367
+ "grad_norm": 0.10198960453271866,
368
+ "learning_rate": 0.0008269999999999999,
369
+ "loss": 1.5792,
370
+ "step": 2350
371
+ },
372
+ {
373
+ "epoch": 53.333333333333336,
374
+ "grad_norm": 0.45619985461235046,
375
+ "learning_rate": 0.0008020000000000001,
376
+ "loss": 1.5791,
377
+ "step": 2400
378
+ },
379
+ {
380
+ "epoch": 54.44444444444444,
381
+ "grad_norm": 0.18695257604122162,
382
+ "learning_rate": 0.000777,
383
+ "loss": 1.589,
384
+ "step": 2450
385
+ },
386
+ {
387
+ "epoch": 55.55555555555556,
388
+ "grad_norm": 0.16741132736206055,
389
+ "learning_rate": 0.0007520000000000001,
390
+ "loss": 1.5854,
391
+ "step": 2500
392
+ },
393
+ {
394
+ "epoch": 55.55555555555556,
395
+ "eval_loss": 1.5435634851455688,
396
+ "eval_runtime": 8.0681,
397
+ "eval_samples_per_second": 19.707,
398
+ "eval_steps_per_second": 2.479,
399
+ "step": 2500
400
+ },
401
+ {
402
+ "epoch": 56.666666666666664,
403
+ "grad_norm": 0.12719061970710754,
404
+ "learning_rate": 0.000727,
405
+ "loss": 1.5736,
406
+ "step": 2550
407
+ },
408
+ {
409
+ "epoch": 57.77777777777778,
410
+ "grad_norm": 0.2572345435619354,
411
+ "learning_rate": 0.0007019999999999999,
412
+ "loss": 1.5808,
413
+ "step": 2600
414
+ },
415
+ {
416
+ "epoch": 58.888888888888886,
417
+ "grad_norm": 0.13093958795070648,
418
+ "learning_rate": 0.0006770000000000001,
419
+ "loss": 1.5701,
420
+ "step": 2650
421
+ },
422
+ {
423
+ "epoch": 60.0,
424
+ "grad_norm": 0.1683294177055359,
425
+ "learning_rate": 0.000652,
426
+ "loss": 1.6017,
427
+ "step": 2700
428
+ },
429
+ {
430
+ "epoch": 61.111111111111114,
431
+ "grad_norm": 0.15231318771839142,
432
+ "learning_rate": 0.0006270000000000001,
433
+ "loss": 1.5809,
434
+ "step": 2750
435
+ },
436
+ {
437
+ "epoch": 62.22222222222222,
438
+ "grad_norm": 0.1944093108177185,
439
+ "learning_rate": 0.000602,
440
+ "loss": 1.5782,
441
+ "step": 2800
442
+ },
443
+ {
444
+ "epoch": 63.333333333333336,
445
+ "grad_norm": 0.3554971218109131,
446
+ "learning_rate": 0.0005769999999999999,
447
+ "loss": 1.5842,
448
+ "step": 2850
449
+ },
450
+ {
451
+ "epoch": 64.44444444444444,
452
+ "grad_norm": 0.3123963475227356,
453
+ "learning_rate": 0.0005520000000000001,
454
+ "loss": 1.5645,
455
+ "step": 2900
456
+ },
457
+ {
458
+ "epoch": 65.55555555555556,
459
+ "grad_norm": 0.14703120291233063,
460
+ "learning_rate": 0.000527,
461
+ "loss": 1.5861,
462
+ "step": 2950
463
+ },
464
+ {
465
+ "epoch": 66.66666666666667,
466
+ "grad_norm": 0.363651305437088,
467
+ "learning_rate": 0.0005020000000000001,
468
+ "loss": 1.5647,
469
+ "step": 3000
470
+ },
471
+ {
472
+ "epoch": 66.66666666666667,
473
+ "eval_loss": 1.5431376695632935,
474
+ "eval_runtime": 7.9782,
475
+ "eval_samples_per_second": 19.929,
476
+ "eval_steps_per_second": 2.507,
477
+ "step": 3000
478
+ },
479
+ {
480
+ "epoch": 67.77777777777777,
481
+ "grad_norm": 0.08024784922599792,
482
+ "learning_rate": 0.000477,
483
+ "loss": 1.5842,
484
+ "step": 3050
485
+ },
486
+ {
487
+ "epoch": 68.88888888888889,
488
+ "grad_norm": 0.18207858502864838,
489
+ "learning_rate": 0.00045200000000000004,
490
+ "loss": 1.5793,
491
+ "step": 3100
492
+ },
493
+ {
494
+ "epoch": 70.0,
495
+ "grad_norm": 0.17078548669815063,
496
+ "learning_rate": 0.000427,
497
+ "loss": 1.5718,
498
+ "step": 3150
499
+ },
500
+ {
501
+ "epoch": 71.11111111111111,
502
+ "grad_norm": 0.3073755204677582,
503
+ "learning_rate": 0.000402,
504
+ "loss": 1.5741,
505
+ "step": 3200
506
+ },
507
+ {
508
+ "epoch": 72.22222222222223,
509
+ "grad_norm": 0.06820886582136154,
510
+ "learning_rate": 0.000377,
511
+ "loss": 1.5748,
512
+ "step": 3250
513
+ },
514
+ {
515
+ "epoch": 73.33333333333333,
516
+ "grad_norm": 0.034480493515729904,
517
+ "learning_rate": 0.000352,
518
+ "loss": 1.5807,
519
+ "step": 3300
520
+ },
521
+ {
522
+ "epoch": 74.44444444444444,
523
+ "grad_norm": 0.06758826971054077,
524
+ "learning_rate": 0.00032700000000000003,
525
+ "loss": 1.5719,
526
+ "step": 3350
527
+ },
528
+ {
529
+ "epoch": 75.55555555555556,
530
+ "grad_norm": 0.11566317081451416,
531
+ "learning_rate": 0.000302,
532
+ "loss": 1.5747,
533
+ "step": 3400
534
+ },
535
+ {
536
+ "epoch": 76.66666666666667,
537
+ "grad_norm": 0.15969228744506836,
538
+ "learning_rate": 0.000277,
539
+ "loss": 1.5846,
540
+ "step": 3450
541
+ },
542
+ {
543
+ "epoch": 77.77777777777777,
544
+ "grad_norm": 0.07962628453969955,
545
+ "learning_rate": 0.000252,
546
+ "loss": 1.5731,
547
+ "step": 3500
548
+ },
549
+ {
550
+ "epoch": 77.77777777777777,
551
+ "eval_loss": 1.5472484827041626,
552
+ "eval_runtime": 8.7234,
553
+ "eval_samples_per_second": 18.227,
554
+ "eval_steps_per_second": 2.293,
555
+ "step": 3500
556
+ }
557
+ ],
558
+ "logging_steps": 50,
559
+ "max_steps": 4000,
560
+ "num_input_tokens_seen": 0,
561
+ "num_train_epochs": 89,
562
+ "save_steps": 500,
563
+ "stateful_callbacks": {
564
+ "TrainerControl": {
565
+ "args": {
566
+ "should_epoch_stop": false,
567
+ "should_evaluate": false,
568
+ "should_log": false,
569
+ "should_save": true,
570
+ "should_training_stop": false
571
+ },
572
+ "attributes": {}
573
+ }
574
+ },
575
+ "total_flos": 2.408072794054013e+16,
576
+ "train_batch_size": 16,
577
+ "trial_name": null,
578
+ "trial_params": null
579
+ }
checkpoint-3500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86205dd0f5e4c109b754d1a2e96d6be52f2878290f435671b3cc191b44de8f2c
3
+ size 5304