mikhail-panzo commited on
Commit
63a9644
1 Parent(s): f2f16e8

Training in progress, step 6000, checkpoint

Browse files
checkpoint-6000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-6000/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-6000/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-6000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b1b129a08163a9aadf60133b723359c691d2d2666733a832d95069d611c59e1
3
+ size 577789320
checkpoint-6000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a1423f77507826ea403c1e7b9b78c1a50da035f5dd0fe8b955b1c55b04d3f77
3
+ size 1155772233
checkpoint-6000/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-6000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba3f58e9864c3a8fe3a4dbacb8f3dbe609eb9c01b9217973110b46351e9fcbe2
3
+ size 14244
checkpoint-6000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:708432cecb713f0d929a61d32593f88e089a379662ca4d36e58512c37db9aeb7
3
+ size 1064
checkpoint-6000/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-6000/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-6000/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-6000/trainer_state.json ADDED
@@ -0,0 +1,969 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.4078540802001953,
3
+ "best_model_checkpoint": "mikhail-panzo/fil_b128_le4_s8000/checkpoint-1000",
4
+ "epoch": 530.3867403314918,
5
+ "eval_steps": 500,
6
+ "global_step": 6000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 4.419889502762431,
13
+ "grad_norm": 2.1200406551361084,
14
+ "learning_rate": 2.5e-06,
15
+ "loss": 0.77,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 8.839779005524862,
20
+ "grad_norm": 0.8919622302055359,
21
+ "learning_rate": 5e-06,
22
+ "loss": 0.695,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 13.259668508287293,
27
+ "grad_norm": 0.9864558577537537,
28
+ "learning_rate": 7.5e-06,
29
+ "loss": 0.6278,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 17.679558011049725,
34
+ "grad_norm": 0.9195986986160278,
35
+ "learning_rate": 1e-05,
36
+ "loss": 0.5431,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 22.099447513812155,
41
+ "grad_norm": 1.0175281763076782,
42
+ "learning_rate": 1.25e-05,
43
+ "loss": 0.5155,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 26.519337016574585,
48
+ "grad_norm": 1.0303566455841064,
49
+ "learning_rate": 1.5e-05,
50
+ "loss": 0.502,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 30.939226519337016,
55
+ "grad_norm": 1.005897879600525,
56
+ "learning_rate": 1.75e-05,
57
+ "loss": 0.4889,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 35.35911602209945,
62
+ "grad_norm": 1.0381360054016113,
63
+ "learning_rate": 2e-05,
64
+ "loss": 0.477,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 39.77900552486188,
69
+ "grad_norm": 1.2148911952972412,
70
+ "learning_rate": 2.25e-05,
71
+ "loss": 0.4711,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 44.19889502762431,
76
+ "grad_norm": 1.5404785871505737,
77
+ "learning_rate": 2.5e-05,
78
+ "loss": 0.463,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 44.19889502762431,
83
+ "eval_loss": 0.42523157596588135,
84
+ "eval_runtime": 15.6706,
85
+ "eval_samples_per_second": 10.274,
86
+ "eval_steps_per_second": 1.34,
87
+ "step": 500
88
+ },
89
+ {
90
+ "epoch": 48.61878453038674,
91
+ "grad_norm": 1.1097612380981445,
92
+ "learning_rate": 2.7500000000000004e-05,
93
+ "loss": 0.4601,
94
+ "step": 550
95
+ },
96
+ {
97
+ "epoch": 53.03867403314917,
98
+ "grad_norm": 1.2796109914779663,
99
+ "learning_rate": 3e-05,
100
+ "loss": 0.454,
101
+ "step": 600
102
+ },
103
+ {
104
+ "epoch": 57.4585635359116,
105
+ "grad_norm": 1.0656301975250244,
106
+ "learning_rate": 3.2500000000000004e-05,
107
+ "loss": 0.4498,
108
+ "step": 650
109
+ },
110
+ {
111
+ "epoch": 61.87845303867403,
112
+ "grad_norm": 1.3971511125564575,
113
+ "learning_rate": 3.5e-05,
114
+ "loss": 0.4439,
115
+ "step": 700
116
+ },
117
+ {
118
+ "epoch": 66.29834254143647,
119
+ "grad_norm": 1.9344899654388428,
120
+ "learning_rate": 3.7500000000000003e-05,
121
+ "loss": 0.4418,
122
+ "step": 750
123
+ },
124
+ {
125
+ "epoch": 70.7182320441989,
126
+ "grad_norm": 1.080334186553955,
127
+ "learning_rate": 4e-05,
128
+ "loss": 0.4379,
129
+ "step": 800
130
+ },
131
+ {
132
+ "epoch": 75.13812154696133,
133
+ "grad_norm": 1.9891581535339355,
134
+ "learning_rate": 4.25e-05,
135
+ "loss": 0.4334,
136
+ "step": 850
137
+ },
138
+ {
139
+ "epoch": 79.55801104972376,
140
+ "grad_norm": 1.3854402303695679,
141
+ "learning_rate": 4.5e-05,
142
+ "loss": 0.4331,
143
+ "step": 900
144
+ },
145
+ {
146
+ "epoch": 83.97790055248619,
147
+ "grad_norm": 1.5818183422088623,
148
+ "learning_rate": 4.75e-05,
149
+ "loss": 0.428,
150
+ "step": 950
151
+ },
152
+ {
153
+ "epoch": 88.39779005524862,
154
+ "grad_norm": 1.8474106788635254,
155
+ "learning_rate": 5e-05,
156
+ "loss": 0.4238,
157
+ "step": 1000
158
+ },
159
+ {
160
+ "epoch": 88.39779005524862,
161
+ "eval_loss": 0.4078540802001953,
162
+ "eval_runtime": 14.8436,
163
+ "eval_samples_per_second": 10.846,
164
+ "eval_steps_per_second": 1.415,
165
+ "step": 1000
166
+ },
167
+ {
168
+ "epoch": 92.81767955801105,
169
+ "grad_norm": 3.1851751804351807,
170
+ "learning_rate": 5.25e-05,
171
+ "loss": 0.4234,
172
+ "step": 1050
173
+ },
174
+ {
175
+ "epoch": 97.23756906077348,
176
+ "grad_norm": 1.3754140138626099,
177
+ "learning_rate": 5.500000000000001e-05,
178
+ "loss": 0.4209,
179
+ "step": 1100
180
+ },
181
+ {
182
+ "epoch": 101.65745856353591,
183
+ "grad_norm": 1.6143475770950317,
184
+ "learning_rate": 5.7499999999999995e-05,
185
+ "loss": 0.4228,
186
+ "step": 1150
187
+ },
188
+ {
189
+ "epoch": 106.07734806629834,
190
+ "grad_norm": 1.9961720705032349,
191
+ "learning_rate": 6e-05,
192
+ "loss": 0.4162,
193
+ "step": 1200
194
+ },
195
+ {
196
+ "epoch": 110.49723756906077,
197
+ "grad_norm": 1.9125347137451172,
198
+ "learning_rate": 6.25e-05,
199
+ "loss": 0.4178,
200
+ "step": 1250
201
+ },
202
+ {
203
+ "epoch": 114.9171270718232,
204
+ "grad_norm": 3.1913959980010986,
205
+ "learning_rate": 6.500000000000001e-05,
206
+ "loss": 0.4143,
207
+ "step": 1300
208
+ },
209
+ {
210
+ "epoch": 119.33701657458563,
211
+ "grad_norm": 2.1010069847106934,
212
+ "learning_rate": 6.750000000000001e-05,
213
+ "loss": 0.4118,
214
+ "step": 1350
215
+ },
216
+ {
217
+ "epoch": 123.75690607734806,
218
+ "grad_norm": 1.3079431056976318,
219
+ "learning_rate": 7e-05,
220
+ "loss": 0.4091,
221
+ "step": 1400
222
+ },
223
+ {
224
+ "epoch": 128.1767955801105,
225
+ "grad_norm": 2.3663132190704346,
226
+ "learning_rate": 7.25e-05,
227
+ "loss": 0.4092,
228
+ "step": 1450
229
+ },
230
+ {
231
+ "epoch": 132.59668508287294,
232
+ "grad_norm": 1.031698226928711,
233
+ "learning_rate": 7.500000000000001e-05,
234
+ "loss": 0.4083,
235
+ "step": 1500
236
+ },
237
+ {
238
+ "epoch": 132.59668508287294,
239
+ "eval_loss": 0.4082767069339752,
240
+ "eval_runtime": 15.1068,
241
+ "eval_samples_per_second": 10.657,
242
+ "eval_steps_per_second": 1.39,
243
+ "step": 1500
244
+ },
245
+ {
246
+ "epoch": 137.01657458563537,
247
+ "grad_norm": 2.7781426906585693,
248
+ "learning_rate": 7.75e-05,
249
+ "loss": 0.4079,
250
+ "step": 1550
251
+ },
252
+ {
253
+ "epoch": 141.4364640883978,
254
+ "grad_norm": 1.136169672012329,
255
+ "learning_rate": 8e-05,
256
+ "loss": 0.4028,
257
+ "step": 1600
258
+ },
259
+ {
260
+ "epoch": 145.85635359116023,
261
+ "grad_norm": 1.7299158573150635,
262
+ "learning_rate": 8.25e-05,
263
+ "loss": 0.4034,
264
+ "step": 1650
265
+ },
266
+ {
267
+ "epoch": 150.27624309392266,
268
+ "grad_norm": 2.4281983375549316,
269
+ "learning_rate": 8.5e-05,
270
+ "loss": 0.3992,
271
+ "step": 1700
272
+ },
273
+ {
274
+ "epoch": 154.6961325966851,
275
+ "grad_norm": 3.6834373474121094,
276
+ "learning_rate": 8.75e-05,
277
+ "loss": 0.3997,
278
+ "step": 1750
279
+ },
280
+ {
281
+ "epoch": 159.11602209944752,
282
+ "grad_norm": 3.1484293937683105,
283
+ "learning_rate": 9e-05,
284
+ "loss": 0.396,
285
+ "step": 1800
286
+ },
287
+ {
288
+ "epoch": 163.53591160220995,
289
+ "grad_norm": 3.978259563446045,
290
+ "learning_rate": 9.250000000000001e-05,
291
+ "loss": 0.3969,
292
+ "step": 1850
293
+ },
294
+ {
295
+ "epoch": 167.95580110497238,
296
+ "grad_norm": 2.906266212463379,
297
+ "learning_rate": 9.5e-05,
298
+ "loss": 0.4017,
299
+ "step": 1900
300
+ },
301
+ {
302
+ "epoch": 172.3756906077348,
303
+ "grad_norm": 1.1569880247116089,
304
+ "learning_rate": 9.75e-05,
305
+ "loss": 0.396,
306
+ "step": 1950
307
+ },
308
+ {
309
+ "epoch": 176.79558011049724,
310
+ "grad_norm": 2.2872982025146484,
311
+ "learning_rate": 0.0001,
312
+ "loss": 0.3928,
313
+ "step": 2000
314
+ },
315
+ {
316
+ "epoch": 176.79558011049724,
317
+ "eval_loss": 0.414584219455719,
318
+ "eval_runtime": 15.2895,
319
+ "eval_samples_per_second": 10.53,
320
+ "eval_steps_per_second": 1.373,
321
+ "step": 2000
322
+ },
323
+ {
324
+ "epoch": 181.21546961325967,
325
+ "grad_norm": 1.1768990755081177,
326
+ "learning_rate": 9.916666666666667e-05,
327
+ "loss": 0.3928,
328
+ "step": 2050
329
+ },
330
+ {
331
+ "epoch": 185.6353591160221,
332
+ "grad_norm": 1.2087491750717163,
333
+ "learning_rate": 9.833333333333333e-05,
334
+ "loss": 0.3901,
335
+ "step": 2100
336
+ },
337
+ {
338
+ "epoch": 190.05524861878453,
339
+ "grad_norm": 1.17588210105896,
340
+ "learning_rate": 9.75e-05,
341
+ "loss": 0.3877,
342
+ "step": 2150
343
+ },
344
+ {
345
+ "epoch": 194.47513812154696,
346
+ "grad_norm": 1.5979018211364746,
347
+ "learning_rate": 9.666666666666667e-05,
348
+ "loss": 0.3879,
349
+ "step": 2200
350
+ },
351
+ {
352
+ "epoch": 198.8950276243094,
353
+ "grad_norm": 1.0489567518234253,
354
+ "learning_rate": 9.583333333333334e-05,
355
+ "loss": 0.3869,
356
+ "step": 2250
357
+ },
358
+ {
359
+ "epoch": 203.31491712707182,
360
+ "grad_norm": 1.833038091659546,
361
+ "learning_rate": 9.501666666666668e-05,
362
+ "loss": 0.3859,
363
+ "step": 2300
364
+ },
365
+ {
366
+ "epoch": 207.73480662983425,
367
+ "grad_norm": 2.075849771499634,
368
+ "learning_rate": 9.418333333333334e-05,
369
+ "loss": 0.3843,
370
+ "step": 2350
371
+ },
372
+ {
373
+ "epoch": 212.15469613259668,
374
+ "grad_norm": 2.8613317012786865,
375
+ "learning_rate": 9.335e-05,
376
+ "loss": 0.381,
377
+ "step": 2400
378
+ },
379
+ {
380
+ "epoch": 216.5745856353591,
381
+ "grad_norm": 1.6291433572769165,
382
+ "learning_rate": 9.251666666666667e-05,
383
+ "loss": 0.3802,
384
+ "step": 2450
385
+ },
386
+ {
387
+ "epoch": 220.99447513812154,
388
+ "grad_norm": 1.1548006534576416,
389
+ "learning_rate": 9.168333333333333e-05,
390
+ "loss": 0.381,
391
+ "step": 2500
392
+ },
393
+ {
394
+ "epoch": 220.99447513812154,
395
+ "eval_loss": 0.4127160906791687,
396
+ "eval_runtime": 15.1026,
397
+ "eval_samples_per_second": 10.66,
398
+ "eval_steps_per_second": 1.39,
399
+ "step": 2500
400
+ },
401
+ {
402
+ "epoch": 225.41436464088397,
403
+ "grad_norm": 1.3956973552703857,
404
+ "learning_rate": 9.085e-05,
405
+ "loss": 0.3795,
406
+ "step": 2550
407
+ },
408
+ {
409
+ "epoch": 229.8342541436464,
410
+ "grad_norm": 1.6613599061965942,
411
+ "learning_rate": 9.001666666666667e-05,
412
+ "loss": 0.3757,
413
+ "step": 2600
414
+ },
415
+ {
416
+ "epoch": 234.25414364640883,
417
+ "grad_norm": 1.4951831102371216,
418
+ "learning_rate": 8.918333333333334e-05,
419
+ "loss": 0.3763,
420
+ "step": 2650
421
+ },
422
+ {
423
+ "epoch": 238.67403314917127,
424
+ "grad_norm": 3.0197391510009766,
425
+ "learning_rate": 8.834999999999999e-05,
426
+ "loss": 0.3765,
427
+ "step": 2700
428
+ },
429
+ {
430
+ "epoch": 243.0939226519337,
431
+ "grad_norm": 0.9016386866569519,
432
+ "learning_rate": 8.751666666666668e-05,
433
+ "loss": 0.3724,
434
+ "step": 2750
435
+ },
436
+ {
437
+ "epoch": 247.51381215469613,
438
+ "grad_norm": 1.6974703073501587,
439
+ "learning_rate": 8.668333333333334e-05,
440
+ "loss": 0.3729,
441
+ "step": 2800
442
+ },
443
+ {
444
+ "epoch": 251.93370165745856,
445
+ "grad_norm": 2.8571081161499023,
446
+ "learning_rate": 8.585000000000001e-05,
447
+ "loss": 0.3715,
448
+ "step": 2850
449
+ },
450
+ {
451
+ "epoch": 256.353591160221,
452
+ "grad_norm": 2.8940412998199463,
453
+ "learning_rate": 8.501666666666667e-05,
454
+ "loss": 0.3716,
455
+ "step": 2900
456
+ },
457
+ {
458
+ "epoch": 260.77348066298345,
459
+ "grad_norm": 1.1075533628463745,
460
+ "learning_rate": 8.418333333333334e-05,
461
+ "loss": 0.3691,
462
+ "step": 2950
463
+ },
464
+ {
465
+ "epoch": 265.1933701657459,
466
+ "grad_norm": 1.145349383354187,
467
+ "learning_rate": 8.335e-05,
468
+ "loss": 0.3712,
469
+ "step": 3000
470
+ },
471
+ {
472
+ "epoch": 265.1933701657459,
473
+ "eval_loss": 0.40815719962120056,
474
+ "eval_runtime": 15.0404,
475
+ "eval_samples_per_second": 10.705,
476
+ "eval_steps_per_second": 1.396,
477
+ "step": 3000
478
+ },
479
+ {
480
+ "epoch": 269.6132596685083,
481
+ "grad_norm": 0.788912832736969,
482
+ "learning_rate": 8.251666666666668e-05,
483
+ "loss": 0.3691,
484
+ "step": 3050
485
+ },
486
+ {
487
+ "epoch": 274.03314917127074,
488
+ "grad_norm": 1.5521596670150757,
489
+ "learning_rate": 8.168333333333333e-05,
490
+ "loss": 0.3687,
491
+ "step": 3100
492
+ },
493
+ {
494
+ "epoch": 278.45303867403317,
495
+ "grad_norm": 1.2638827562332153,
496
+ "learning_rate": 8.085e-05,
497
+ "loss": 0.3681,
498
+ "step": 3150
499
+ },
500
+ {
501
+ "epoch": 282.8729281767956,
502
+ "grad_norm": 1.6849925518035889,
503
+ "learning_rate": 8.001666666666667e-05,
504
+ "loss": 0.3664,
505
+ "step": 3200
506
+ },
507
+ {
508
+ "epoch": 287.292817679558,
509
+ "grad_norm": 1.9946844577789307,
510
+ "learning_rate": 7.918333333333334e-05,
511
+ "loss": 0.3671,
512
+ "step": 3250
513
+ },
514
+ {
515
+ "epoch": 291.71270718232046,
516
+ "grad_norm": 1.3560407161712646,
517
+ "learning_rate": 7.835000000000001e-05,
518
+ "loss": 0.3631,
519
+ "step": 3300
520
+ },
521
+ {
522
+ "epoch": 296.1325966850829,
523
+ "grad_norm": 1.4300700426101685,
524
+ "learning_rate": 7.751666666666668e-05,
525
+ "loss": 0.3616,
526
+ "step": 3350
527
+ },
528
+ {
529
+ "epoch": 300.5524861878453,
530
+ "grad_norm": 1.0440477132797241,
531
+ "learning_rate": 7.668333333333335e-05,
532
+ "loss": 0.3627,
533
+ "step": 3400
534
+ },
535
+ {
536
+ "epoch": 304.97237569060775,
537
+ "grad_norm": 0.78243088722229,
538
+ "learning_rate": 7.585e-05,
539
+ "loss": 0.3624,
540
+ "step": 3450
541
+ },
542
+ {
543
+ "epoch": 309.3922651933702,
544
+ "grad_norm": 1.4992115497589111,
545
+ "learning_rate": 7.501666666666667e-05,
546
+ "loss": 0.3589,
547
+ "step": 3500
548
+ },
549
+ {
550
+ "epoch": 309.3922651933702,
551
+ "eval_loss": 0.4111907482147217,
552
+ "eval_runtime": 15.163,
553
+ "eval_samples_per_second": 10.618,
554
+ "eval_steps_per_second": 1.385,
555
+ "step": 3500
556
+ },
557
+ {
558
+ "epoch": 313.8121546961326,
559
+ "grad_norm": 1.0604934692382812,
560
+ "learning_rate": 7.418333333333334e-05,
561
+ "loss": 0.3599,
562
+ "step": 3550
563
+ },
564
+ {
565
+ "epoch": 318.23204419889504,
566
+ "grad_norm": 0.8896504640579224,
567
+ "learning_rate": 7.335000000000001e-05,
568
+ "loss": 0.361,
569
+ "step": 3600
570
+ },
571
+ {
572
+ "epoch": 322.65193370165747,
573
+ "grad_norm": 1.3247913122177124,
574
+ "learning_rate": 7.251666666666666e-05,
575
+ "loss": 0.3575,
576
+ "step": 3650
577
+ },
578
+ {
579
+ "epoch": 327.0718232044199,
580
+ "grad_norm": 0.6475388407707214,
581
+ "learning_rate": 7.168333333333333e-05,
582
+ "loss": 0.3575,
583
+ "step": 3700
584
+ },
585
+ {
586
+ "epoch": 331.49171270718233,
587
+ "grad_norm": 0.6724014282226562,
588
+ "learning_rate": 7.085e-05,
589
+ "loss": 0.3558,
590
+ "step": 3750
591
+ },
592
+ {
593
+ "epoch": 335.91160220994476,
594
+ "grad_norm": 1.0705429315567017,
595
+ "learning_rate": 7.001666666666667e-05,
596
+ "loss": 0.3564,
597
+ "step": 3800
598
+ },
599
+ {
600
+ "epoch": 340.3314917127072,
601
+ "grad_norm": 1.0692503452301025,
602
+ "learning_rate": 6.918333333333334e-05,
603
+ "loss": 0.3601,
604
+ "step": 3850
605
+ },
606
+ {
607
+ "epoch": 344.7513812154696,
608
+ "grad_norm": 0.8828334212303162,
609
+ "learning_rate": 6.835000000000001e-05,
610
+ "loss": 0.3548,
611
+ "step": 3900
612
+ },
613
+ {
614
+ "epoch": 349.17127071823205,
615
+ "grad_norm": 0.6390587091445923,
616
+ "learning_rate": 6.751666666666668e-05,
617
+ "loss": 0.3529,
618
+ "step": 3950
619
+ },
620
+ {
621
+ "epoch": 353.5911602209945,
622
+ "grad_norm": 0.7502044439315796,
623
+ "learning_rate": 6.668333333333333e-05,
624
+ "loss": 0.3544,
625
+ "step": 4000
626
+ },
627
+ {
628
+ "epoch": 353.5911602209945,
629
+ "eval_loss": 0.4128935635089874,
630
+ "eval_runtime": 16.0241,
631
+ "eval_samples_per_second": 10.047,
632
+ "eval_steps_per_second": 1.311,
633
+ "step": 4000
634
+ },
635
+ {
636
+ "epoch": 358.0110497237569,
637
+ "grad_norm": 0.8119373917579651,
638
+ "learning_rate": 6.585e-05,
639
+ "loss": 0.3527,
640
+ "step": 4050
641
+ },
642
+ {
643
+ "epoch": 362.43093922651934,
644
+ "grad_norm": 0.7546930313110352,
645
+ "learning_rate": 6.501666666666667e-05,
646
+ "loss": 0.352,
647
+ "step": 4100
648
+ },
649
+ {
650
+ "epoch": 366.8508287292818,
651
+ "grad_norm": 1.1060476303100586,
652
+ "learning_rate": 6.418333333333334e-05,
653
+ "loss": 0.3514,
654
+ "step": 4150
655
+ },
656
+ {
657
+ "epoch": 371.2707182320442,
658
+ "grad_norm": 1.3087377548217773,
659
+ "learning_rate": 6.335e-05,
660
+ "loss": 0.3508,
661
+ "step": 4200
662
+ },
663
+ {
664
+ "epoch": 375.69060773480663,
665
+ "grad_norm": 0.7953481078147888,
666
+ "learning_rate": 6.251666666666666e-05,
667
+ "loss": 0.3493,
668
+ "step": 4250
669
+ },
670
+ {
671
+ "epoch": 380.11049723756906,
672
+ "grad_norm": 1.1858787536621094,
673
+ "learning_rate": 6.168333333333333e-05,
674
+ "loss": 0.3493,
675
+ "step": 4300
676
+ },
677
+ {
678
+ "epoch": 384.5303867403315,
679
+ "grad_norm": 0.9929729700088501,
680
+ "learning_rate": 6.085000000000001e-05,
681
+ "loss": 0.3485,
682
+ "step": 4350
683
+ },
684
+ {
685
+ "epoch": 388.9502762430939,
686
+ "grad_norm": 0.8156938552856445,
687
+ "learning_rate": 6.0016666666666664e-05,
688
+ "loss": 0.347,
689
+ "step": 4400
690
+ },
691
+ {
692
+ "epoch": 393.37016574585635,
693
+ "grad_norm": 1.0693029165267944,
694
+ "learning_rate": 5.918333333333333e-05,
695
+ "loss": 0.3489,
696
+ "step": 4450
697
+ },
698
+ {
699
+ "epoch": 397.7900552486188,
700
+ "grad_norm": 1.1495543718338013,
701
+ "learning_rate": 5.835e-05,
702
+ "loss": 0.347,
703
+ "step": 4500
704
+ },
705
+ {
706
+ "epoch": 397.7900552486188,
707
+ "eval_loss": 0.41448456048965454,
708
+ "eval_runtime": 15.2027,
709
+ "eval_samples_per_second": 10.59,
710
+ "eval_steps_per_second": 1.381,
711
+ "step": 4500
712
+ },
713
+ {
714
+ "epoch": 402.2099447513812,
715
+ "grad_norm": 0.9324333071708679,
716
+ "learning_rate": 5.751666666666667e-05,
717
+ "loss": 0.3473,
718
+ "step": 4550
719
+ },
720
+ {
721
+ "epoch": 406.62983425414365,
722
+ "grad_norm": 0.6897173523902893,
723
+ "learning_rate": 5.668333333333333e-05,
724
+ "loss": 0.3441,
725
+ "step": 4600
726
+ },
727
+ {
728
+ "epoch": 411.0497237569061,
729
+ "grad_norm": 1.1474602222442627,
730
+ "learning_rate": 5.585e-05,
731
+ "loss": 0.346,
732
+ "step": 4650
733
+ },
734
+ {
735
+ "epoch": 415.4696132596685,
736
+ "grad_norm": 1.4295645952224731,
737
+ "learning_rate": 5.501666666666667e-05,
738
+ "loss": 0.3442,
739
+ "step": 4700
740
+ },
741
+ {
742
+ "epoch": 419.88950276243094,
743
+ "grad_norm": 0.8301011323928833,
744
+ "learning_rate": 5.4183333333333334e-05,
745
+ "loss": 0.344,
746
+ "step": 4750
747
+ },
748
+ {
749
+ "epoch": 424.30939226519337,
750
+ "grad_norm": 1.2894037961959839,
751
+ "learning_rate": 5.335e-05,
752
+ "loss": 0.3431,
753
+ "step": 4800
754
+ },
755
+ {
756
+ "epoch": 428.7292817679558,
757
+ "grad_norm": 0.6597224473953247,
758
+ "learning_rate": 5.251666666666667e-05,
759
+ "loss": 0.3443,
760
+ "step": 4850
761
+ },
762
+ {
763
+ "epoch": 433.1491712707182,
764
+ "grad_norm": 0.8772273063659668,
765
+ "learning_rate": 5.168333333333334e-05,
766
+ "loss": 0.3447,
767
+ "step": 4900
768
+ },
769
+ {
770
+ "epoch": 437.56906077348066,
771
+ "grad_norm": 0.6241065263748169,
772
+ "learning_rate": 5.0849999999999996e-05,
773
+ "loss": 0.3412,
774
+ "step": 4950
775
+ },
776
+ {
777
+ "epoch": 441.9889502762431,
778
+ "grad_norm": 0.8275649547576904,
779
+ "learning_rate": 5.0016666666666665e-05,
780
+ "loss": 0.3417,
781
+ "step": 5000
782
+ },
783
+ {
784
+ "epoch": 441.9889502762431,
785
+ "eval_loss": 0.42017292976379395,
786
+ "eval_runtime": 15.281,
787
+ "eval_samples_per_second": 10.536,
788
+ "eval_steps_per_second": 1.374,
789
+ "step": 5000
790
+ },
791
+ {
792
+ "epoch": 446.4088397790055,
793
+ "grad_norm": 0.9026231169700623,
794
+ "learning_rate": 4.9183333333333334e-05,
795
+ "loss": 0.3417,
796
+ "step": 5050
797
+ },
798
+ {
799
+ "epoch": 450.82872928176795,
800
+ "grad_norm": 1.248936414718628,
801
+ "learning_rate": 4.835e-05,
802
+ "loss": 0.3396,
803
+ "step": 5100
804
+ },
805
+ {
806
+ "epoch": 455.2486187845304,
807
+ "grad_norm": 1.257866621017456,
808
+ "learning_rate": 4.751666666666667e-05,
809
+ "loss": 0.3414,
810
+ "step": 5150
811
+ },
812
+ {
813
+ "epoch": 459.6685082872928,
814
+ "grad_norm": 1.029221534729004,
815
+ "learning_rate": 4.6683333333333334e-05,
816
+ "loss": 0.3396,
817
+ "step": 5200
818
+ },
819
+ {
820
+ "epoch": 464.08839779005524,
821
+ "grad_norm": 1.1000345945358276,
822
+ "learning_rate": 4.585e-05,
823
+ "loss": 0.3392,
824
+ "step": 5250
825
+ },
826
+ {
827
+ "epoch": 468.50828729281767,
828
+ "grad_norm": 0.7120194435119629,
829
+ "learning_rate": 4.5016666666666665e-05,
830
+ "loss": 0.3412,
831
+ "step": 5300
832
+ },
833
+ {
834
+ "epoch": 472.9281767955801,
835
+ "grad_norm": 0.5869295597076416,
836
+ "learning_rate": 4.4183333333333334e-05,
837
+ "loss": 0.3398,
838
+ "step": 5350
839
+ },
840
+ {
841
+ "epoch": 477.34806629834253,
842
+ "grad_norm": 0.8485609292984009,
843
+ "learning_rate": 4.335e-05,
844
+ "loss": 0.3385,
845
+ "step": 5400
846
+ },
847
+ {
848
+ "epoch": 481.76795580110496,
849
+ "grad_norm": 0.520220160484314,
850
+ "learning_rate": 4.251666666666667e-05,
851
+ "loss": 0.3366,
852
+ "step": 5450
853
+ },
854
+ {
855
+ "epoch": 486.1878453038674,
856
+ "grad_norm": 0.6742538809776306,
857
+ "learning_rate": 4.1683333333333335e-05,
858
+ "loss": 0.3372,
859
+ "step": 5500
860
+ },
861
+ {
862
+ "epoch": 486.1878453038674,
863
+ "eval_loss": 0.41964003443717957,
864
+ "eval_runtime": 16.3724,
865
+ "eval_samples_per_second": 9.834,
866
+ "eval_steps_per_second": 1.283,
867
+ "step": 5500
868
+ },
869
+ {
870
+ "epoch": 490.6077348066298,
871
+ "grad_norm": 0.8028594255447388,
872
+ "learning_rate": 4.085e-05,
873
+ "loss": 0.3378,
874
+ "step": 5550
875
+ },
876
+ {
877
+ "epoch": 495.02762430939225,
878
+ "grad_norm": 0.6300519704818726,
879
+ "learning_rate": 4.0016666666666666e-05,
880
+ "loss": 0.3373,
881
+ "step": 5600
882
+ },
883
+ {
884
+ "epoch": 499.4475138121547,
885
+ "grad_norm": 0.6699241399765015,
886
+ "learning_rate": 3.9183333333333335e-05,
887
+ "loss": 0.3365,
888
+ "step": 5650
889
+ },
890
+ {
891
+ "epoch": 503.8674033149171,
892
+ "grad_norm": 0.6375143527984619,
893
+ "learning_rate": 3.8350000000000004e-05,
894
+ "loss": 0.3358,
895
+ "step": 5700
896
+ },
897
+ {
898
+ "epoch": 508.28729281767954,
899
+ "grad_norm": 0.6259489059448242,
900
+ "learning_rate": 3.7516666666666666e-05,
901
+ "loss": 0.3363,
902
+ "step": 5750
903
+ },
904
+ {
905
+ "epoch": 512.707182320442,
906
+ "grad_norm": 0.5885080695152283,
907
+ "learning_rate": 3.6683333333333335e-05,
908
+ "loss": 0.3353,
909
+ "step": 5800
910
+ },
911
+ {
912
+ "epoch": 517.1270718232045,
913
+ "grad_norm": 0.6459413766860962,
914
+ "learning_rate": 3.585e-05,
915
+ "loss": 0.3344,
916
+ "step": 5850
917
+ },
918
+ {
919
+ "epoch": 521.5469613259669,
920
+ "grad_norm": 0.6893390417098999,
921
+ "learning_rate": 3.501666666666667e-05,
922
+ "loss": 0.3353,
923
+ "step": 5900
924
+ },
925
+ {
926
+ "epoch": 525.9668508287293,
927
+ "grad_norm": 0.644479513168335,
928
+ "learning_rate": 3.4183333333333335e-05,
929
+ "loss": 0.3354,
930
+ "step": 5950
931
+ },
932
+ {
933
+ "epoch": 530.3867403314918,
934
+ "grad_norm": 0.5544880628585815,
935
+ "learning_rate": 3.3350000000000004e-05,
936
+ "loss": 0.3335,
937
+ "step": 6000
938
+ },
939
+ {
940
+ "epoch": 530.3867403314918,
941
+ "eval_loss": 0.4183831810951233,
942
+ "eval_runtime": 15.2687,
943
+ "eval_samples_per_second": 10.544,
944
+ "eval_steps_per_second": 1.375,
945
+ "step": 6000
946
+ }
947
+ ],
948
+ "logging_steps": 50,
949
+ "max_steps": 8000,
950
+ "num_input_tokens_seen": 0,
951
+ "num_train_epochs": 728,
952
+ "save_steps": 500,
953
+ "stateful_callbacks": {
954
+ "TrainerControl": {
955
+ "args": {
956
+ "should_epoch_stop": false,
957
+ "should_evaluate": false,
958
+ "should_log": false,
959
+ "should_save": true,
960
+ "should_training_stop": false
961
+ },
962
+ "attributes": {}
963
+ }
964
+ },
965
+ "total_flos": 1.358564139384107e+17,
966
+ "train_batch_size": 8,
967
+ "trial_name": null,
968
+ "trial_params": null
969
+ }
checkpoint-6000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:067941ce54247f638979fecfb22f3bf44cd9bc2df99dac3a436b8b51c9f0dbc4
3
+ size 5304