mikhail-panzo commited on
Commit
5f6c04a
1 Parent(s): 6012249

Training in progress, step 4000, checkpoint

Browse files
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": 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-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": "4.41.0.dev0"
9
+ }
checkpoint-4000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5735dbb2304f699b0469fd78dea73939fc3406f7b38f38486169adb346bb1202
3
+ size 577789320
checkpoint-4000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28393028951306a2352a14ff085e0712c3e78fb15f3458cd43714bbdef8f23c9
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 sha256:784f1db4fed30ab61de3c5dbe336b223f5c8543909844b95259c1e0c6bb85fb8
3
+ size 14244
checkpoint-4000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ce8989bf898af955ced2a0e50e75ff17c2343dbf0730bd2aa80a1bedcf51a7c9
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": 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-4000/trainer_state.json ADDED
@@ -0,0 +1,657 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.389825701713562,
3
+ "best_model_checkpoint": "mikhail_panzo/zlm_b64_le5_s12000/checkpoint-4000",
4
+ "epoch": 3.350785340314136,
5
+ "eval_steps": 500,
6
+ "global_step": 4000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.041884816753926704,
13
+ "grad_norm": 12.821951866149902,
14
+ "learning_rate": 2.4000000000000003e-07,
15
+ "loss": 1.1152,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.08376963350785341,
20
+ "grad_norm": 6.186221122741699,
21
+ "learning_rate": 4.900000000000001e-07,
22
+ "loss": 1.0584,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.1256544502617801,
27
+ "grad_norm": 13.08674430847168,
28
+ "learning_rate": 7.4e-07,
29
+ "loss": 0.9996,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.16753926701570682,
34
+ "grad_norm": 4.10632848739624,
35
+ "learning_rate": 9.9e-07,
36
+ "loss": 0.9355,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.2094240837696335,
41
+ "grad_norm": 3.2428064346313477,
42
+ "learning_rate": 1.2400000000000002e-06,
43
+ "loss": 0.852,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.2513089005235602,
48
+ "grad_norm": 2.0861401557922363,
49
+ "learning_rate": 1.4900000000000001e-06,
50
+ "loss": 0.8241,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.2931937172774869,
55
+ "grad_norm": 8.602317810058594,
56
+ "learning_rate": 1.74e-06,
57
+ "loss": 0.7629,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.33507853403141363,
62
+ "grad_norm": 2.7594990730285645,
63
+ "learning_rate": 1.9900000000000004e-06,
64
+ "loss": 0.7418,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.3769633507853403,
69
+ "grad_norm": 2.7047348022460938,
70
+ "learning_rate": 2.24e-06,
71
+ "loss": 0.7166,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.418848167539267,
76
+ "grad_norm": 2.0666158199310303,
77
+ "learning_rate": 2.4900000000000003e-06,
78
+ "loss": 0.6876,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.418848167539267,
83
+ "eval_loss": 0.5937426090240479,
84
+ "eval_runtime": 263.0955,
85
+ "eval_samples_per_second": 32.266,
86
+ "eval_steps_per_second": 4.037,
87
+ "step": 500
88
+ },
89
+ {
90
+ "epoch": 0.4607329842931937,
91
+ "grad_norm": 2.955193519592285,
92
+ "learning_rate": 2.7400000000000004e-06,
93
+ "loss": 0.673,
94
+ "step": 550
95
+ },
96
+ {
97
+ "epoch": 0.5026178010471204,
98
+ "grad_norm": 4.856322765350342,
99
+ "learning_rate": 2.99e-06,
100
+ "loss": 0.6539,
101
+ "step": 600
102
+ },
103
+ {
104
+ "epoch": 0.5445026178010471,
105
+ "grad_norm": 7.631103992462158,
106
+ "learning_rate": 3.2400000000000003e-06,
107
+ "loss": 0.6588,
108
+ "step": 650
109
+ },
110
+ {
111
+ "epoch": 0.5863874345549738,
112
+ "grad_norm": 2.715771436691284,
113
+ "learning_rate": 3.49e-06,
114
+ "loss": 0.624,
115
+ "step": 700
116
+ },
117
+ {
118
+ "epoch": 0.6282722513089005,
119
+ "grad_norm": 2.7684273719787598,
120
+ "learning_rate": 3.74e-06,
121
+ "loss": 0.6116,
122
+ "step": 750
123
+ },
124
+ {
125
+ "epoch": 0.6701570680628273,
126
+ "grad_norm": 2.219599723815918,
127
+ "learning_rate": 3.990000000000001e-06,
128
+ "loss": 0.6097,
129
+ "step": 800
130
+ },
131
+ {
132
+ "epoch": 0.7120418848167539,
133
+ "grad_norm": 3.749188184738159,
134
+ "learning_rate": 4.24e-06,
135
+ "loss": 0.614,
136
+ "step": 850
137
+ },
138
+ {
139
+ "epoch": 0.7539267015706806,
140
+ "grad_norm": 2.120619773864746,
141
+ "learning_rate": 4.49e-06,
142
+ "loss": 0.5814,
143
+ "step": 900
144
+ },
145
+ {
146
+ "epoch": 0.7958115183246073,
147
+ "grad_norm": 2.7725327014923096,
148
+ "learning_rate": 4.74e-06,
149
+ "loss": 0.5895,
150
+ "step": 950
151
+ },
152
+ {
153
+ "epoch": 0.837696335078534,
154
+ "grad_norm": 9.851961135864258,
155
+ "learning_rate": 4.9900000000000005e-06,
156
+ "loss": 0.5623,
157
+ "step": 1000
158
+ },
159
+ {
160
+ "epoch": 0.837696335078534,
161
+ "eval_loss": 0.4933945834636688,
162
+ "eval_runtime": 258.739,
163
+ "eval_samples_per_second": 32.809,
164
+ "eval_steps_per_second": 4.105,
165
+ "step": 1000
166
+ },
167
+ {
168
+ "epoch": 0.8795811518324608,
169
+ "grad_norm": 5.074089527130127,
170
+ "learning_rate": 5.240000000000001e-06,
171
+ "loss": 0.5748,
172
+ "step": 1050
173
+ },
174
+ {
175
+ "epoch": 0.9214659685863874,
176
+ "grad_norm": 5.884639739990234,
177
+ "learning_rate": 5.485e-06,
178
+ "loss": 0.5395,
179
+ "step": 1100
180
+ },
181
+ {
182
+ "epoch": 0.9633507853403142,
183
+ "grad_norm": 3.1588447093963623,
184
+ "learning_rate": 5.735e-06,
185
+ "loss": 0.5478,
186
+ "step": 1150
187
+ },
188
+ {
189
+ "epoch": 1.0052356020942408,
190
+ "grad_norm": 2.4452970027923584,
191
+ "learning_rate": 5.985000000000001e-06,
192
+ "loss": 0.538,
193
+ "step": 1200
194
+ },
195
+ {
196
+ "epoch": 1.0471204188481675,
197
+ "grad_norm": 8.290769577026367,
198
+ "learning_rate": 6.235000000000001e-06,
199
+ "loss": 0.5411,
200
+ "step": 1250
201
+ },
202
+ {
203
+ "epoch": 1.0890052356020943,
204
+ "grad_norm": 4.080046653747559,
205
+ "learning_rate": 6.485000000000001e-06,
206
+ "loss": 0.5387,
207
+ "step": 1300
208
+ },
209
+ {
210
+ "epoch": 1.130890052356021,
211
+ "grad_norm": 2.7520201206207275,
212
+ "learning_rate": 6.735000000000001e-06,
213
+ "loss": 0.5426,
214
+ "step": 1350
215
+ },
216
+ {
217
+ "epoch": 1.1727748691099475,
218
+ "grad_norm": 3.8192107677459717,
219
+ "learning_rate": 6.985000000000001e-06,
220
+ "loss": 0.5315,
221
+ "step": 1400
222
+ },
223
+ {
224
+ "epoch": 1.2146596858638743,
225
+ "grad_norm": 9.535676956176758,
226
+ "learning_rate": 7.235000000000001e-06,
227
+ "loss": 0.5159,
228
+ "step": 1450
229
+ },
230
+ {
231
+ "epoch": 1.256544502617801,
232
+ "grad_norm": 2.5731022357940674,
233
+ "learning_rate": 7.485000000000001e-06,
234
+ "loss": 0.5165,
235
+ "step": 1500
236
+ },
237
+ {
238
+ "epoch": 1.256544502617801,
239
+ "eval_loss": 0.4628298878669739,
240
+ "eval_runtime": 257.0265,
241
+ "eval_samples_per_second": 33.028,
242
+ "eval_steps_per_second": 4.132,
243
+ "step": 1500
244
+ },
245
+ {
246
+ "epoch": 1.2984293193717278,
247
+ "grad_norm": 3.662090539932251,
248
+ "learning_rate": 7.735e-06,
249
+ "loss": 0.5188,
250
+ "step": 1550
251
+ },
252
+ {
253
+ "epoch": 1.3403141361256545,
254
+ "grad_norm": 3.1680572032928467,
255
+ "learning_rate": 7.985e-06,
256
+ "loss": 0.5148,
257
+ "step": 1600
258
+ },
259
+ {
260
+ "epoch": 1.3821989528795813,
261
+ "grad_norm": 3.601471185684204,
262
+ "learning_rate": 8.235e-06,
263
+ "loss": 0.5127,
264
+ "step": 1650
265
+ },
266
+ {
267
+ "epoch": 1.4240837696335078,
268
+ "grad_norm": 4.497046947479248,
269
+ "learning_rate": 8.485000000000001e-06,
270
+ "loss": 0.5039,
271
+ "step": 1700
272
+ },
273
+ {
274
+ "epoch": 1.4659685863874345,
275
+ "grad_norm": 3.6496424674987793,
276
+ "learning_rate": 8.735000000000002e-06,
277
+ "loss": 0.5114,
278
+ "step": 1750
279
+ },
280
+ {
281
+ "epoch": 1.5078534031413613,
282
+ "grad_norm": 2.7080061435699463,
283
+ "learning_rate": 8.985000000000001e-06,
284
+ "loss": 0.4953,
285
+ "step": 1800
286
+ },
287
+ {
288
+ "epoch": 1.5497382198952878,
289
+ "grad_norm": 2.5439274311065674,
290
+ "learning_rate": 9.235e-06,
291
+ "loss": 0.497,
292
+ "step": 1850
293
+ },
294
+ {
295
+ "epoch": 1.5916230366492146,
296
+ "grad_norm": 2.004573106765747,
297
+ "learning_rate": 9.485000000000002e-06,
298
+ "loss": 0.4889,
299
+ "step": 1900
300
+ },
301
+ {
302
+ "epoch": 1.6335078534031413,
303
+ "grad_norm": 4.699184894561768,
304
+ "learning_rate": 9.735e-06,
305
+ "loss": 0.4925,
306
+ "step": 1950
307
+ },
308
+ {
309
+ "epoch": 1.675392670157068,
310
+ "grad_norm": 3.2080323696136475,
311
+ "learning_rate": 9.985000000000002e-06,
312
+ "loss": 0.4851,
313
+ "step": 2000
314
+ },
315
+ {
316
+ "epoch": 1.675392670157068,
317
+ "eval_loss": 0.43240800499916077,
318
+ "eval_runtime": 256.4745,
319
+ "eval_samples_per_second": 33.099,
320
+ "eval_steps_per_second": 4.141,
321
+ "step": 2000
322
+ },
323
+ {
324
+ "epoch": 1.7172774869109948,
325
+ "grad_norm": 2.5463037490844727,
326
+ "learning_rate": 9.953000000000001e-06,
327
+ "loss": 0.4802,
328
+ "step": 2050
329
+ },
330
+ {
331
+ "epoch": 1.7591623036649215,
332
+ "grad_norm": 3.6697189807891846,
333
+ "learning_rate": 9.903e-06,
334
+ "loss": 0.485,
335
+ "step": 2100
336
+ },
337
+ {
338
+ "epoch": 1.8010471204188483,
339
+ "grad_norm": 2.6455466747283936,
340
+ "learning_rate": 9.853e-06,
341
+ "loss": 0.4863,
342
+ "step": 2150
343
+ },
344
+ {
345
+ "epoch": 1.8429319371727748,
346
+ "grad_norm": 2.849780321121216,
347
+ "learning_rate": 9.803e-06,
348
+ "loss": 0.4824,
349
+ "step": 2200
350
+ },
351
+ {
352
+ "epoch": 1.8848167539267016,
353
+ "grad_norm": 2.037757158279419,
354
+ "learning_rate": 9.753e-06,
355
+ "loss": 0.4762,
356
+ "step": 2250
357
+ },
358
+ {
359
+ "epoch": 1.9267015706806283,
360
+ "grad_norm": 2.060755491256714,
361
+ "learning_rate": 9.703000000000002e-06,
362
+ "loss": 0.4793,
363
+ "step": 2300
364
+ },
365
+ {
366
+ "epoch": 1.9685863874345548,
367
+ "grad_norm": 2.3721907138824463,
368
+ "learning_rate": 9.653e-06,
369
+ "loss": 0.4744,
370
+ "step": 2350
371
+ },
372
+ {
373
+ "epoch": 2.0104712041884816,
374
+ "grad_norm": 2.605724573135376,
375
+ "learning_rate": 9.603000000000001e-06,
376
+ "loss": 0.4665,
377
+ "step": 2400
378
+ },
379
+ {
380
+ "epoch": 2.0523560209424083,
381
+ "grad_norm": 1.9419960975646973,
382
+ "learning_rate": 9.553000000000002e-06,
383
+ "loss": 0.4701,
384
+ "step": 2450
385
+ },
386
+ {
387
+ "epoch": 2.094240837696335,
388
+ "grad_norm": 2.3074657917022705,
389
+ "learning_rate": 9.503e-06,
390
+ "loss": 0.4672,
391
+ "step": 2500
392
+ },
393
+ {
394
+ "epoch": 2.094240837696335,
395
+ "eval_loss": 0.4174647927284241,
396
+ "eval_runtime": 259.1813,
397
+ "eval_samples_per_second": 32.753,
398
+ "eval_steps_per_second": 4.098,
399
+ "step": 2500
400
+ },
401
+ {
402
+ "epoch": 2.136125654450262,
403
+ "grad_norm": 20.03424644470215,
404
+ "learning_rate": 9.453e-06,
405
+ "loss": 0.4676,
406
+ "step": 2550
407
+ },
408
+ {
409
+ "epoch": 2.1780104712041886,
410
+ "grad_norm": 3.410041093826294,
411
+ "learning_rate": 9.403000000000001e-06,
412
+ "loss": 0.471,
413
+ "step": 2600
414
+ },
415
+ {
416
+ "epoch": 2.2198952879581153,
417
+ "grad_norm": 2.1835362911224365,
418
+ "learning_rate": 9.353000000000002e-06,
419
+ "loss": 0.4639,
420
+ "step": 2650
421
+ },
422
+ {
423
+ "epoch": 2.261780104712042,
424
+ "grad_norm": 2.245004177093506,
425
+ "learning_rate": 9.303e-06,
426
+ "loss": 0.4642,
427
+ "step": 2700
428
+ },
429
+ {
430
+ "epoch": 2.303664921465969,
431
+ "grad_norm": 2.2452590465545654,
432
+ "learning_rate": 9.253000000000001e-06,
433
+ "loss": 0.4638,
434
+ "step": 2750
435
+ },
436
+ {
437
+ "epoch": 2.345549738219895,
438
+ "grad_norm": 3.5843396186828613,
439
+ "learning_rate": 9.203000000000002e-06,
440
+ "loss": 0.4546,
441
+ "step": 2800
442
+ },
443
+ {
444
+ "epoch": 2.387434554973822,
445
+ "grad_norm": 1.689903974533081,
446
+ "learning_rate": 9.153e-06,
447
+ "loss": 0.4607,
448
+ "step": 2850
449
+ },
450
+ {
451
+ "epoch": 2.4293193717277486,
452
+ "grad_norm": 2.26529860496521,
453
+ "learning_rate": 9.103e-06,
454
+ "loss": 0.4732,
455
+ "step": 2900
456
+ },
457
+ {
458
+ "epoch": 2.4712041884816753,
459
+ "grad_norm": 1.669374942779541,
460
+ "learning_rate": 9.053000000000001e-06,
461
+ "loss": 0.4578,
462
+ "step": 2950
463
+ },
464
+ {
465
+ "epoch": 2.513089005235602,
466
+ "grad_norm": 2.6614444255828857,
467
+ "learning_rate": 9.003e-06,
468
+ "loss": 0.4615,
469
+ "step": 3000
470
+ },
471
+ {
472
+ "epoch": 2.513089005235602,
473
+ "eval_loss": 0.40531307458877563,
474
+ "eval_runtime": 260.7294,
475
+ "eval_samples_per_second": 32.559,
476
+ "eval_steps_per_second": 4.073,
477
+ "step": 3000
478
+ },
479
+ {
480
+ "epoch": 2.554973821989529,
481
+ "grad_norm": 2.6586270332336426,
482
+ "learning_rate": 8.953e-06,
483
+ "loss": 0.4486,
484
+ "step": 3050
485
+ },
486
+ {
487
+ "epoch": 2.5968586387434556,
488
+ "grad_norm": 1.5337822437286377,
489
+ "learning_rate": 8.903000000000001e-06,
490
+ "loss": 0.4474,
491
+ "step": 3100
492
+ },
493
+ {
494
+ "epoch": 2.6387434554973823,
495
+ "grad_norm": 5.568012714385986,
496
+ "learning_rate": 8.853e-06,
497
+ "loss": 0.4466,
498
+ "step": 3150
499
+ },
500
+ {
501
+ "epoch": 2.680628272251309,
502
+ "grad_norm": 1.8297160863876343,
503
+ "learning_rate": 8.803e-06,
504
+ "loss": 0.454,
505
+ "step": 3200
506
+ },
507
+ {
508
+ "epoch": 2.7225130890052354,
509
+ "grad_norm": 3.0298142433166504,
510
+ "learning_rate": 8.753e-06,
511
+ "loss": 0.4513,
512
+ "step": 3250
513
+ },
514
+ {
515
+ "epoch": 2.7643979057591626,
516
+ "grad_norm": 2.6482794284820557,
517
+ "learning_rate": 8.703e-06,
518
+ "loss": 0.4461,
519
+ "step": 3300
520
+ },
521
+ {
522
+ "epoch": 2.806282722513089,
523
+ "grad_norm": 1.9304325580596924,
524
+ "learning_rate": 8.653e-06,
525
+ "loss": 0.4464,
526
+ "step": 3350
527
+ },
528
+ {
529
+ "epoch": 2.8481675392670156,
530
+ "grad_norm": 2.3229286670684814,
531
+ "learning_rate": 8.603e-06,
532
+ "loss": 0.4475,
533
+ "step": 3400
534
+ },
535
+ {
536
+ "epoch": 2.8900523560209423,
537
+ "grad_norm": 1.8658236265182495,
538
+ "learning_rate": 8.553000000000001e-06,
539
+ "loss": 0.4476,
540
+ "step": 3450
541
+ },
542
+ {
543
+ "epoch": 2.931937172774869,
544
+ "grad_norm": 1.574404239654541,
545
+ "learning_rate": 8.503e-06,
546
+ "loss": 0.4426,
547
+ "step": 3500
548
+ },
549
+ {
550
+ "epoch": 2.931937172774869,
551
+ "eval_loss": 0.39671415090560913,
552
+ "eval_runtime": 259.1033,
553
+ "eval_samples_per_second": 32.763,
554
+ "eval_steps_per_second": 4.099,
555
+ "step": 3500
556
+ },
557
+ {
558
+ "epoch": 2.973821989528796,
559
+ "grad_norm": 2.0523784160614014,
560
+ "learning_rate": 8.453000000000002e-06,
561
+ "loss": 0.4545,
562
+ "step": 3550
563
+ },
564
+ {
565
+ "epoch": 3.0157068062827226,
566
+ "grad_norm": 2.5431597232818604,
567
+ "learning_rate": 8.403e-06,
568
+ "loss": 0.4416,
569
+ "step": 3600
570
+ },
571
+ {
572
+ "epoch": 3.0575916230366493,
573
+ "grad_norm": 1.7480727434158325,
574
+ "learning_rate": 8.353000000000001e-06,
575
+ "loss": 0.4396,
576
+ "step": 3650
577
+ },
578
+ {
579
+ "epoch": 3.099476439790576,
580
+ "grad_norm": 2.19262433052063,
581
+ "learning_rate": 8.303000000000002e-06,
582
+ "loss": 0.4403,
583
+ "step": 3700
584
+ },
585
+ {
586
+ "epoch": 3.141361256544503,
587
+ "grad_norm": 1.6949037313461304,
588
+ "learning_rate": 8.253e-06,
589
+ "loss": 0.4345,
590
+ "step": 3750
591
+ },
592
+ {
593
+ "epoch": 3.183246073298429,
594
+ "grad_norm": 3.506704807281494,
595
+ "learning_rate": 8.203000000000001e-06,
596
+ "loss": 0.4364,
597
+ "step": 3800
598
+ },
599
+ {
600
+ "epoch": 3.225130890052356,
601
+ "grad_norm": 2.155595064163208,
602
+ "learning_rate": 8.153000000000001e-06,
603
+ "loss": 0.4361,
604
+ "step": 3850
605
+ },
606
+ {
607
+ "epoch": 3.2670157068062826,
608
+ "grad_norm": 1.7618342638015747,
609
+ "learning_rate": 8.103e-06,
610
+ "loss": 0.4348,
611
+ "step": 3900
612
+ },
613
+ {
614
+ "epoch": 3.3089005235602094,
615
+ "grad_norm": 2.2834861278533936,
616
+ "learning_rate": 8.053e-06,
617
+ "loss": 0.4435,
618
+ "step": 3950
619
+ },
620
+ {
621
+ "epoch": 3.350785340314136,
622
+ "grad_norm": 1.7081154584884644,
623
+ "learning_rate": 8.003000000000001e-06,
624
+ "loss": 0.4295,
625
+ "step": 4000
626
+ },
627
+ {
628
+ "epoch": 3.350785340314136,
629
+ "eval_loss": 0.389825701713562,
630
+ "eval_runtime": 254.7511,
631
+ "eval_samples_per_second": 33.323,
632
+ "eval_steps_per_second": 4.169,
633
+ "step": 4000
634
+ }
635
+ ],
636
+ "logging_steps": 50,
637
+ "max_steps": 12000,
638
+ "num_input_tokens_seen": 0,
639
+ "num_train_epochs": 11,
640
+ "save_steps": 500,
641
+ "stateful_callbacks": {
642
+ "TrainerControl": {
643
+ "args": {
644
+ "should_epoch_stop": false,
645
+ "should_evaluate": false,
646
+ "should_log": false,
647
+ "should_save": true,
648
+ "should_training_stop": false
649
+ },
650
+ "attributes": {}
651
+ }
652
+ },
653
+ "total_flos": 3.597539584997952e+16,
654
+ "train_batch_size": 16,
655
+ "trial_name": null,
656
+ "trial_params": null
657
+ }
checkpoint-4000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b20f3b06aa67136878a19fd6f17a63e91db785ca54662fa7debf3d46f1fe910d
3
+ size 5304