mikhail-panzo commited on
Commit
72ed4dd
1 Parent(s): a2645b9

Training in progress, step 5000, checkpoint

Browse files
checkpoint-5000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-5000/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-5000/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-5000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37baba3b9d83f5cbb76d13a17843324d1336e43defeea671d37cc3e2540391a7
3
+ size 577789320
checkpoint-5000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68c57a43a16a76c7e6992f336097aef93e129001bbe0b84d6825f55d04daa591
3
+ size 1155772233
checkpoint-5000/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-5000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:916cc928436669a5d6c1a602bedc6a0d8d15b05474edb7f78b1a84ccd3394580
3
+ size 14244
checkpoint-5000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6ab38839505be77e8af214c9f2561c64ea029c3726893e214d680ff97dcf232
3
+ size 1064
checkpoint-5000/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-5000/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-5000/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-5000/trainer_state.json ADDED
@@ -0,0 +1,813 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.384044349193573,
3
+ "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_s12000/checkpoint-5000",
4
+ "epoch": 4.187604690117253,
5
+ "eval_steps": 500,
6
+ "global_step": 5000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.04187604690117253,
13
+ "grad_norm": 4.788218975067139,
14
+ "learning_rate": 2.4500000000000004e-07,
15
+ "loss": 1.1819,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.08375209380234507,
20
+ "grad_norm": 15.618935585021973,
21
+ "learning_rate": 4.95e-07,
22
+ "loss": 1.0042,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.12562814070351758,
27
+ "grad_norm": 2.7821555137634277,
28
+ "learning_rate": 7.4e-07,
29
+ "loss": 0.9676,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.16750418760469013,
34
+ "grad_norm": 3.759263038635254,
35
+ "learning_rate": 9.85e-07,
36
+ "loss": 0.9362,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.20938023450586266,
41
+ "grad_norm": 3.906439781188965,
42
+ "learning_rate": 1.235e-06,
43
+ "loss": 0.8667,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.25125628140703515,
48
+ "grad_norm": 4.397643566131592,
49
+ "learning_rate": 1.485e-06,
50
+ "loss": 0.8455,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.2931323283082077,
55
+ "grad_norm": 2.94077467918396,
56
+ "learning_rate": 1.7350000000000001e-06,
57
+ "loss": 0.8376,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.33500837520938026,
62
+ "grad_norm": 4.161074161529541,
63
+ "learning_rate": 1.985e-06,
64
+ "loss": 0.7521,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.3768844221105528,
69
+ "grad_norm": 2.8088552951812744,
70
+ "learning_rate": 2.235e-06,
71
+ "loss": 0.7444,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.4187604690117253,
76
+ "grad_norm": 2.62283992767334,
77
+ "learning_rate": 2.4850000000000003e-06,
78
+ "loss": 0.7129,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.4187604690117253,
83
+ "eval_loss": 0.6066474914550781,
84
+ "eval_runtime": 213.4911,
85
+ "eval_samples_per_second": 39.763,
86
+ "eval_steps_per_second": 4.974,
87
+ "step": 500
88
+ },
89
+ {
90
+ "epoch": 0.46063651591289784,
91
+ "grad_norm": 3.654665946960449,
92
+ "learning_rate": 2.7350000000000005e-06,
93
+ "loss": 0.6772,
94
+ "step": 550
95
+ },
96
+ {
97
+ "epoch": 0.5025125628140703,
98
+ "grad_norm": 5.1851887702941895,
99
+ "learning_rate": 2.9850000000000002e-06,
100
+ "loss": 0.6552,
101
+ "step": 600
102
+ },
103
+ {
104
+ "epoch": 0.5443886097152428,
105
+ "grad_norm": 4.017187595367432,
106
+ "learning_rate": 3.2350000000000004e-06,
107
+ "loss": 0.6506,
108
+ "step": 650
109
+ },
110
+ {
111
+ "epoch": 0.5862646566164154,
112
+ "grad_norm": 2.1161563396453857,
113
+ "learning_rate": 3.485e-06,
114
+ "loss": 0.6301,
115
+ "step": 700
116
+ },
117
+ {
118
+ "epoch": 0.628140703517588,
119
+ "grad_norm": 1.857313871383667,
120
+ "learning_rate": 3.7350000000000002e-06,
121
+ "loss": 0.6373,
122
+ "step": 750
123
+ },
124
+ {
125
+ "epoch": 0.6700167504187605,
126
+ "grad_norm": 3.6841330528259277,
127
+ "learning_rate": 3.985000000000001e-06,
128
+ "loss": 0.6305,
129
+ "step": 800
130
+ },
131
+ {
132
+ "epoch": 0.711892797319933,
133
+ "grad_norm": 2.6450912952423096,
134
+ "learning_rate": 4.235e-06,
135
+ "loss": 0.6169,
136
+ "step": 850
137
+ },
138
+ {
139
+ "epoch": 0.7537688442211056,
140
+ "grad_norm": 1.7241184711456299,
141
+ "learning_rate": 4.485e-06,
142
+ "loss": 0.5933,
143
+ "step": 900
144
+ },
145
+ {
146
+ "epoch": 0.7956448911222781,
147
+ "grad_norm": 2.602735996246338,
148
+ "learning_rate": 4.735e-06,
149
+ "loss": 0.583,
150
+ "step": 950
151
+ },
152
+ {
153
+ "epoch": 0.8375209380234506,
154
+ "grad_norm": 12.419028282165527,
155
+ "learning_rate": 4.9850000000000006e-06,
156
+ "loss": 0.5922,
157
+ "step": 1000
158
+ },
159
+ {
160
+ "epoch": 0.8375209380234506,
161
+ "eval_loss": 0.4979843497276306,
162
+ "eval_runtime": 214.0502,
163
+ "eval_samples_per_second": 39.659,
164
+ "eval_steps_per_second": 4.961,
165
+ "step": 1000
166
+ },
167
+ {
168
+ "epoch": 0.8793969849246231,
169
+ "grad_norm": 3.200178861618042,
170
+ "learning_rate": 5.235e-06,
171
+ "loss": 0.5678,
172
+ "step": 1050
173
+ },
174
+ {
175
+ "epoch": 0.9212730318257957,
176
+ "grad_norm": 2.458869218826294,
177
+ "learning_rate": 5.485e-06,
178
+ "loss": 0.5558,
179
+ "step": 1100
180
+ },
181
+ {
182
+ "epoch": 0.9631490787269682,
183
+ "grad_norm": 2.341752290725708,
184
+ "learning_rate": 5.735e-06,
185
+ "loss": 0.5666,
186
+ "step": 1150
187
+ },
188
+ {
189
+ "epoch": 1.0050251256281406,
190
+ "grad_norm": 2.9191603660583496,
191
+ "learning_rate": 5.985000000000001e-06,
192
+ "loss": 0.5345,
193
+ "step": 1200
194
+ },
195
+ {
196
+ "epoch": 1.0469011725293131,
197
+ "grad_norm": 2.68806529045105,
198
+ "learning_rate": 6.235000000000001e-06,
199
+ "loss": 0.5653,
200
+ "step": 1250
201
+ },
202
+ {
203
+ "epoch": 1.0887772194304857,
204
+ "grad_norm": 3.278820276260376,
205
+ "learning_rate": 6.485000000000001e-06,
206
+ "loss": 0.5565,
207
+ "step": 1300
208
+ },
209
+ {
210
+ "epoch": 1.1306532663316582,
211
+ "grad_norm": 3.066511869430542,
212
+ "learning_rate": 6.735000000000001e-06,
213
+ "loss": 0.5644,
214
+ "step": 1350
215
+ },
216
+ {
217
+ "epoch": 1.1725293132328307,
218
+ "grad_norm": 4.269862651824951,
219
+ "learning_rate": 6.985000000000001e-06,
220
+ "loss": 0.5356,
221
+ "step": 1400
222
+ },
223
+ {
224
+ "epoch": 1.2144053601340032,
225
+ "grad_norm": 3.6681156158447266,
226
+ "learning_rate": 7.235000000000001e-06,
227
+ "loss": 0.5246,
228
+ "step": 1450
229
+ },
230
+ {
231
+ "epoch": 1.2562814070351758,
232
+ "grad_norm": 3.1227505207061768,
233
+ "learning_rate": 7.485000000000001e-06,
234
+ "loss": 0.5168,
235
+ "step": 1500
236
+ },
237
+ {
238
+ "epoch": 1.2562814070351758,
239
+ "eval_loss": 0.45785218477249146,
240
+ "eval_runtime": 211.6749,
241
+ "eval_samples_per_second": 40.104,
242
+ "eval_steps_per_second": 5.017,
243
+ "step": 1500
244
+ },
245
+ {
246
+ "epoch": 1.2981574539363483,
247
+ "grad_norm": 2.6239662170410156,
248
+ "learning_rate": 7.735e-06,
249
+ "loss": 0.5153,
250
+ "step": 1550
251
+ },
252
+ {
253
+ "epoch": 1.3400335008375208,
254
+ "grad_norm": 2.881901979446411,
255
+ "learning_rate": 7.985e-06,
256
+ "loss": 0.518,
257
+ "step": 1600
258
+ },
259
+ {
260
+ "epoch": 1.3819095477386933,
261
+ "grad_norm": 2.8052310943603516,
262
+ "learning_rate": 8.235e-06,
263
+ "loss": 0.5272,
264
+ "step": 1650
265
+ },
266
+ {
267
+ "epoch": 1.4237855946398659,
268
+ "grad_norm": 2.6334052085876465,
269
+ "learning_rate": 8.485000000000001e-06,
270
+ "loss": 0.5188,
271
+ "step": 1700
272
+ },
273
+ {
274
+ "epoch": 1.4656616415410384,
275
+ "grad_norm": 3.038274049758911,
276
+ "learning_rate": 8.735000000000002e-06,
277
+ "loss": 0.5139,
278
+ "step": 1750
279
+ },
280
+ {
281
+ "epoch": 1.507537688442211,
282
+ "grad_norm": 3.2438852787017822,
283
+ "learning_rate": 8.985000000000001e-06,
284
+ "loss": 0.5092,
285
+ "step": 1800
286
+ },
287
+ {
288
+ "epoch": 1.5494137353433834,
289
+ "grad_norm": 4.168883800506592,
290
+ "learning_rate": 9.235e-06,
291
+ "loss": 0.5103,
292
+ "step": 1850
293
+ },
294
+ {
295
+ "epoch": 1.591289782244556,
296
+ "grad_norm": 2.6399967670440674,
297
+ "learning_rate": 9.485000000000002e-06,
298
+ "loss": 0.5025,
299
+ "step": 1900
300
+ },
301
+ {
302
+ "epoch": 1.6331658291457285,
303
+ "grad_norm": 1.6828950643539429,
304
+ "learning_rate": 9.735e-06,
305
+ "loss": 0.4963,
306
+ "step": 1950
307
+ },
308
+ {
309
+ "epoch": 1.675041876046901,
310
+ "grad_norm": 1.8955535888671875,
311
+ "learning_rate": 9.985000000000002e-06,
312
+ "loss": 0.4954,
313
+ "step": 2000
314
+ },
315
+ {
316
+ "epoch": 1.675041876046901,
317
+ "eval_loss": 0.43827974796295166,
318
+ "eval_runtime": 212.1455,
319
+ "eval_samples_per_second": 40.015,
320
+ "eval_steps_per_second": 5.006,
321
+ "step": 2000
322
+ },
323
+ {
324
+ "epoch": 1.7169179229480735,
325
+ "grad_norm": 3.9634511470794678,
326
+ "learning_rate": 9.953000000000001e-06,
327
+ "loss": 0.4891,
328
+ "step": 2050
329
+ },
330
+ {
331
+ "epoch": 1.758793969849246,
332
+ "grad_norm": 3.5446126461029053,
333
+ "learning_rate": 9.903e-06,
334
+ "loss": 0.5141,
335
+ "step": 2100
336
+ },
337
+ {
338
+ "epoch": 1.8006700167504186,
339
+ "grad_norm": 3.1621015071868896,
340
+ "learning_rate": 9.853e-06,
341
+ "loss": 0.4941,
342
+ "step": 2150
343
+ },
344
+ {
345
+ "epoch": 1.8425460636515911,
346
+ "grad_norm": 2.6622440814971924,
347
+ "learning_rate": 9.803e-06,
348
+ "loss": 0.4979,
349
+ "step": 2200
350
+ },
351
+ {
352
+ "epoch": 1.8844221105527639,
353
+ "grad_norm": 1.7774969339370728,
354
+ "learning_rate": 9.753e-06,
355
+ "loss": 0.4768,
356
+ "step": 2250
357
+ },
358
+ {
359
+ "epoch": 1.9262981574539364,
360
+ "grad_norm": 2.1141610145568848,
361
+ "learning_rate": 9.703000000000002e-06,
362
+ "loss": 0.4834,
363
+ "step": 2300
364
+ },
365
+ {
366
+ "epoch": 1.968174204355109,
367
+ "grad_norm": 3.4576170444488525,
368
+ "learning_rate": 9.653e-06,
369
+ "loss": 0.493,
370
+ "step": 2350
371
+ },
372
+ {
373
+ "epoch": 2.0100502512562812,
374
+ "grad_norm": 2.4238383769989014,
375
+ "learning_rate": 9.603000000000001e-06,
376
+ "loss": 0.4836,
377
+ "step": 2400
378
+ },
379
+ {
380
+ "epoch": 2.0519262981574538,
381
+ "grad_norm": 2.2526774406433105,
382
+ "learning_rate": 9.553000000000002e-06,
383
+ "loss": 0.4832,
384
+ "step": 2450
385
+ },
386
+ {
387
+ "epoch": 2.0938023450586263,
388
+ "grad_norm": 2.8719518184661865,
389
+ "learning_rate": 9.503e-06,
390
+ "loss": 0.4836,
391
+ "step": 2500
392
+ },
393
+ {
394
+ "epoch": 2.0938023450586263,
395
+ "eval_loss": 0.4208170473575592,
396
+ "eval_runtime": 214.437,
397
+ "eval_samples_per_second": 39.587,
398
+ "eval_steps_per_second": 4.953,
399
+ "step": 2500
400
+ },
401
+ {
402
+ "epoch": 2.135678391959799,
403
+ "grad_norm": 3.2671754360198975,
404
+ "learning_rate": 9.453e-06,
405
+ "loss": 0.4775,
406
+ "step": 2550
407
+ },
408
+ {
409
+ "epoch": 2.1775544388609713,
410
+ "grad_norm": 2.454883337020874,
411
+ "learning_rate": 9.403000000000001e-06,
412
+ "loss": 0.4777,
413
+ "step": 2600
414
+ },
415
+ {
416
+ "epoch": 2.219430485762144,
417
+ "grad_norm": 2.901097297668457,
418
+ "learning_rate": 9.353000000000002e-06,
419
+ "loss": 0.4757,
420
+ "step": 2650
421
+ },
422
+ {
423
+ "epoch": 2.2613065326633164,
424
+ "grad_norm": 2.980057954788208,
425
+ "learning_rate": 9.303e-06,
426
+ "loss": 0.4644,
427
+ "step": 2700
428
+ },
429
+ {
430
+ "epoch": 2.3031825795644894,
431
+ "grad_norm": 1.7162193059921265,
432
+ "learning_rate": 9.253000000000001e-06,
433
+ "loss": 0.4714,
434
+ "step": 2750
435
+ },
436
+ {
437
+ "epoch": 2.3450586264656614,
438
+ "grad_norm": 2.912874937057495,
439
+ "learning_rate": 9.203000000000002e-06,
440
+ "loss": 0.4596,
441
+ "step": 2800
442
+ },
443
+ {
444
+ "epoch": 2.3869346733668344,
445
+ "grad_norm": 1.7781060934066772,
446
+ "learning_rate": 9.153e-06,
447
+ "loss": 0.4591,
448
+ "step": 2850
449
+ },
450
+ {
451
+ "epoch": 2.4288107202680065,
452
+ "grad_norm": 5.2557902336120605,
453
+ "learning_rate": 9.103e-06,
454
+ "loss": 0.4652,
455
+ "step": 2900
456
+ },
457
+ {
458
+ "epoch": 2.4706867671691795,
459
+ "grad_norm": 1.7230316400527954,
460
+ "learning_rate": 9.053000000000001e-06,
461
+ "loss": 0.46,
462
+ "step": 2950
463
+ },
464
+ {
465
+ "epoch": 2.5125628140703515,
466
+ "grad_norm": 4.071449279785156,
467
+ "learning_rate": 9.003e-06,
468
+ "loss": 0.4623,
469
+ "step": 3000
470
+ },
471
+ {
472
+ "epoch": 2.5125628140703515,
473
+ "eval_loss": 0.4098353981971741,
474
+ "eval_runtime": 214.5603,
475
+ "eval_samples_per_second": 39.565,
476
+ "eval_steps_per_second": 4.95,
477
+ "step": 3000
478
+ },
479
+ {
480
+ "epoch": 2.5544388609715245,
481
+ "grad_norm": 3.5458014011383057,
482
+ "learning_rate": 8.953e-06,
483
+ "loss": 0.4488,
484
+ "step": 3050
485
+ },
486
+ {
487
+ "epoch": 2.5963149078726966,
488
+ "grad_norm": 1.726624846458435,
489
+ "learning_rate": 8.903000000000001e-06,
490
+ "loss": 0.4536,
491
+ "step": 3100
492
+ },
493
+ {
494
+ "epoch": 2.6381909547738696,
495
+ "grad_norm": 3.204106092453003,
496
+ "learning_rate": 8.853e-06,
497
+ "loss": 0.4654,
498
+ "step": 3150
499
+ },
500
+ {
501
+ "epoch": 2.6800670016750416,
502
+ "grad_norm": 2.245302677154541,
503
+ "learning_rate": 8.803e-06,
504
+ "loss": 0.462,
505
+ "step": 3200
506
+ },
507
+ {
508
+ "epoch": 2.7219430485762146,
509
+ "grad_norm": 2.0797626972198486,
510
+ "learning_rate": 8.753e-06,
511
+ "loss": 0.4599,
512
+ "step": 3250
513
+ },
514
+ {
515
+ "epoch": 2.7638190954773867,
516
+ "grad_norm": 1.9437323808670044,
517
+ "learning_rate": 8.703e-06,
518
+ "loss": 0.4473,
519
+ "step": 3300
520
+ },
521
+ {
522
+ "epoch": 2.8056951423785597,
523
+ "grad_norm": 2.5299806594848633,
524
+ "learning_rate": 8.653e-06,
525
+ "loss": 0.4442,
526
+ "step": 3350
527
+ },
528
+ {
529
+ "epoch": 2.8475711892797317,
530
+ "grad_norm": 3.208460807800293,
531
+ "learning_rate": 8.603e-06,
532
+ "loss": 0.4553,
533
+ "step": 3400
534
+ },
535
+ {
536
+ "epoch": 2.8894472361809047,
537
+ "grad_norm": 2.703047037124634,
538
+ "learning_rate": 8.553000000000001e-06,
539
+ "loss": 0.4532,
540
+ "step": 3450
541
+ },
542
+ {
543
+ "epoch": 2.931323283082077,
544
+ "grad_norm": 1.9532470703125,
545
+ "learning_rate": 8.503e-06,
546
+ "loss": 0.4499,
547
+ "step": 3500
548
+ },
549
+ {
550
+ "epoch": 2.931323283082077,
551
+ "eval_loss": 0.4003259837627411,
552
+ "eval_runtime": 213.8518,
553
+ "eval_samples_per_second": 39.696,
554
+ "eval_steps_per_second": 4.966,
555
+ "step": 3500
556
+ },
557
+ {
558
+ "epoch": 2.9731993299832498,
559
+ "grad_norm": 1.9028701782226562,
560
+ "learning_rate": 8.453000000000002e-06,
561
+ "loss": 0.4391,
562
+ "step": 3550
563
+ },
564
+ {
565
+ "epoch": 3.0150753768844223,
566
+ "grad_norm": 10.080265045166016,
567
+ "learning_rate": 8.403e-06,
568
+ "loss": 0.4462,
569
+ "step": 3600
570
+ },
571
+ {
572
+ "epoch": 3.056951423785595,
573
+ "grad_norm": 2.996328115463257,
574
+ "learning_rate": 8.353000000000001e-06,
575
+ "loss": 0.45,
576
+ "step": 3650
577
+ },
578
+ {
579
+ "epoch": 3.0988274706867673,
580
+ "grad_norm": 2.221226930618286,
581
+ "learning_rate": 8.303000000000002e-06,
582
+ "loss": 0.4469,
583
+ "step": 3700
584
+ },
585
+ {
586
+ "epoch": 3.14070351758794,
587
+ "grad_norm": 2.961524248123169,
588
+ "learning_rate": 8.253e-06,
589
+ "loss": 0.4535,
590
+ "step": 3750
591
+ },
592
+ {
593
+ "epoch": 3.1825795644891124,
594
+ "grad_norm": 2.487060785293579,
595
+ "learning_rate": 8.203000000000001e-06,
596
+ "loss": 0.4444,
597
+ "step": 3800
598
+ },
599
+ {
600
+ "epoch": 3.224455611390285,
601
+ "grad_norm": 1.5449366569519043,
602
+ "learning_rate": 8.153000000000001e-06,
603
+ "loss": 0.449,
604
+ "step": 3850
605
+ },
606
+ {
607
+ "epoch": 3.2663316582914574,
608
+ "grad_norm": 1.6193325519561768,
609
+ "learning_rate": 8.103e-06,
610
+ "loss": 0.4423,
611
+ "step": 3900
612
+ },
613
+ {
614
+ "epoch": 3.30820770519263,
615
+ "grad_norm": 2.442594528198242,
616
+ "learning_rate": 8.053e-06,
617
+ "loss": 0.4456,
618
+ "step": 3950
619
+ },
620
+ {
621
+ "epoch": 3.3500837520938025,
622
+ "grad_norm": 1.759829044342041,
623
+ "learning_rate": 8.003000000000001e-06,
624
+ "loss": 0.4421,
625
+ "step": 4000
626
+ },
627
+ {
628
+ "epoch": 3.3500837520938025,
629
+ "eval_loss": 0.39415255188941956,
630
+ "eval_runtime": 212.9896,
631
+ "eval_samples_per_second": 39.856,
632
+ "eval_steps_per_second": 4.986,
633
+ "step": 4000
634
+ },
635
+ {
636
+ "epoch": 3.391959798994975,
637
+ "grad_norm": 1.6665236949920654,
638
+ "learning_rate": 7.953e-06,
639
+ "loss": 0.4626,
640
+ "step": 4050
641
+ },
642
+ {
643
+ "epoch": 3.4338358458961475,
644
+ "grad_norm": 2.2765064239501953,
645
+ "learning_rate": 7.903e-06,
646
+ "loss": 0.4467,
647
+ "step": 4100
648
+ },
649
+ {
650
+ "epoch": 3.47571189279732,
651
+ "grad_norm": 2.401715040206909,
652
+ "learning_rate": 7.853000000000001e-06,
653
+ "loss": 0.4523,
654
+ "step": 4150
655
+ },
656
+ {
657
+ "epoch": 3.5175879396984926,
658
+ "grad_norm": 4.074472427368164,
659
+ "learning_rate": 7.804e-06,
660
+ "loss": 0.4422,
661
+ "step": 4200
662
+ },
663
+ {
664
+ "epoch": 3.559463986599665,
665
+ "grad_norm": 2.0492303371429443,
666
+ "learning_rate": 7.754e-06,
667
+ "loss": 0.4417,
668
+ "step": 4250
669
+ },
670
+ {
671
+ "epoch": 3.6013400335008376,
672
+ "grad_norm": 2.8641483783721924,
673
+ "learning_rate": 7.704000000000001e-06,
674
+ "loss": 0.4364,
675
+ "step": 4300
676
+ },
677
+ {
678
+ "epoch": 3.64321608040201,
679
+ "grad_norm": 2.1105141639709473,
680
+ "learning_rate": 7.654e-06,
681
+ "loss": 0.4456,
682
+ "step": 4350
683
+ },
684
+ {
685
+ "epoch": 3.6850921273031827,
686
+ "grad_norm": 3.2357890605926514,
687
+ "learning_rate": 7.604e-06,
688
+ "loss": 0.4314,
689
+ "step": 4400
690
+ },
691
+ {
692
+ "epoch": 3.726968174204355,
693
+ "grad_norm": 1.3265612125396729,
694
+ "learning_rate": 7.554e-06,
695
+ "loss": 0.4309,
696
+ "step": 4450
697
+ },
698
+ {
699
+ "epoch": 3.7688442211055277,
700
+ "grad_norm": 1.558438777923584,
701
+ "learning_rate": 7.5040000000000005e-06,
702
+ "loss": 0.4345,
703
+ "step": 4500
704
+ },
705
+ {
706
+ "epoch": 3.7688442211055277,
707
+ "eval_loss": 0.3893836736679077,
708
+ "eval_runtime": 214.7943,
709
+ "eval_samples_per_second": 39.522,
710
+ "eval_steps_per_second": 4.944,
711
+ "step": 4500
712
+ },
713
+ {
714
+ "epoch": 3.8107202680067003,
715
+ "grad_norm": 1.6830787658691406,
716
+ "learning_rate": 7.454e-06,
717
+ "loss": 0.4279,
718
+ "step": 4550
719
+ },
720
+ {
721
+ "epoch": 3.852596314907873,
722
+ "grad_norm": 2.7424182891845703,
723
+ "learning_rate": 7.404e-06,
724
+ "loss": 0.4312,
725
+ "step": 4600
726
+ },
727
+ {
728
+ "epoch": 3.8944723618090453,
729
+ "grad_norm": 1.8119078874588013,
730
+ "learning_rate": 7.354000000000001e-06,
731
+ "loss": 0.4262,
732
+ "step": 4650
733
+ },
734
+ {
735
+ "epoch": 3.936348408710218,
736
+ "grad_norm": 1.8333683013916016,
737
+ "learning_rate": 7.304000000000001e-06,
738
+ "loss": 0.4225,
739
+ "step": 4700
740
+ },
741
+ {
742
+ "epoch": 3.9782244556113904,
743
+ "grad_norm": 1.8040356636047363,
744
+ "learning_rate": 7.254000000000001e-06,
745
+ "loss": 0.4239,
746
+ "step": 4750
747
+ },
748
+ {
749
+ "epoch": 4.0201005025125625,
750
+ "grad_norm": 1.7291460037231445,
751
+ "learning_rate": 7.204000000000001e-06,
752
+ "loss": 0.4327,
753
+ "step": 4800
754
+ },
755
+ {
756
+ "epoch": 4.061976549413735,
757
+ "grad_norm": 1.8311458826065063,
758
+ "learning_rate": 7.154000000000001e-06,
759
+ "loss": 0.4297,
760
+ "step": 4850
761
+ },
762
+ {
763
+ "epoch": 4.1038525963149075,
764
+ "grad_norm": 2.10210919380188,
765
+ "learning_rate": 7.104000000000001e-06,
766
+ "loss": 0.4207,
767
+ "step": 4900
768
+ },
769
+ {
770
+ "epoch": 4.1457286432160805,
771
+ "grad_norm": 2.5385794639587402,
772
+ "learning_rate": 7.0540000000000006e-06,
773
+ "loss": 0.4342,
774
+ "step": 4950
775
+ },
776
+ {
777
+ "epoch": 4.187604690117253,
778
+ "grad_norm": 1.8550605773925781,
779
+ "learning_rate": 7.004000000000001e-06,
780
+ "loss": 0.4233,
781
+ "step": 5000
782
+ },
783
+ {
784
+ "epoch": 4.187604690117253,
785
+ "eval_loss": 0.384044349193573,
786
+ "eval_runtime": 215.1398,
787
+ "eval_samples_per_second": 39.458,
788
+ "eval_steps_per_second": 4.936,
789
+ "step": 5000
790
+ }
791
+ ],
792
+ "logging_steps": 50,
793
+ "max_steps": 12000,
794
+ "num_input_tokens_seen": 0,
795
+ "num_train_epochs": 11,
796
+ "save_steps": 500,
797
+ "stateful_callbacks": {
798
+ "TrainerControl": {
799
+ "args": {
800
+ "should_epoch_stop": false,
801
+ "should_evaluate": false,
802
+ "should_log": false,
803
+ "should_save": true,
804
+ "should_training_stop": false
805
+ },
806
+ "attributes": {}
807
+ }
808
+ },
809
+ "total_flos": 4.86925862796743e+16,
810
+ "train_batch_size": 32,
811
+ "trial_name": null,
812
+ "trial_params": null
813
+ }
checkpoint-5000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4687cdb2b9124db8f1d7d7525bb9403c3df24fb82ca7d74640c10781e1e94be
3
+ size 5304