mikhail-panzo commited on
Commit
5795bcb
1 Parent(s): ed04c91

Training in progress, step 4500, checkpoint

Browse files
checkpoint-4500/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<ctc_blank>": 80,
3
+ "<mask>": 79
4
+ }
checkpoint-4500/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-4500/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-4500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28bd8c4540407750715005702e248f5a1bbcb47e12075aac89498f4e898f7897
3
+ size 577789320
checkpoint-4500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1133bc2dd826fdbc1096ba73f7076059a8bab3ae057d17dc06a42e69b7ee279
3
+ size 1155772233
checkpoint-4500/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-4500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55b41ee41df1b95daa7c247b046435c9be54cceaa08049c6c5013204de4ef1d1
3
+ size 14244
checkpoint-4500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4066aea012cd08c3c9f61bfa0e75b9c47febdda1cf36d357485fed6771cb1982
3
+ size 1064
checkpoint-4500/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-4500/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-4500/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-4500/trainer_state.json ADDED
@@ -0,0 +1,735 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.3396616578102112,
3
+ "best_model_checkpoint": "mikhail_panzo/zlm_b64_le4_s12000/checkpoint-4500",
4
+ "epoch": 3.770261780104712,
5
+ "eval_steps": 500,
6
+ "global_step": 4500,
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": 2.697540283203125,
14
+ "learning_rate": 2.4000000000000003e-06,
15
+ "loss": 1.0526,
16
+ "step": 50
17
+ },
18
+ {
19
+ "epoch": 0.08376963350785341,
20
+ "grad_norm": 4.0351996421813965,
21
+ "learning_rate": 4.9000000000000005e-06,
22
+ "loss": 0.8484,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 0.1256544502617801,
27
+ "grad_norm": 4.858840465545654,
28
+ "learning_rate": 7.4e-06,
29
+ "loss": 0.7505,
30
+ "step": 150
31
+ },
32
+ {
33
+ "epoch": 0.16753926701570682,
34
+ "grad_norm": 2.396157741546631,
35
+ "learning_rate": 9.900000000000002e-06,
36
+ "loss": 0.6889,
37
+ "step": 200
38
+ },
39
+ {
40
+ "epoch": 0.2094240837696335,
41
+ "grad_norm": 2.660891056060791,
42
+ "learning_rate": 1.24e-05,
43
+ "loss": 0.6363,
44
+ "step": 250
45
+ },
46
+ {
47
+ "epoch": 0.2513089005235602,
48
+ "grad_norm": 2.848365068435669,
49
+ "learning_rate": 1.49e-05,
50
+ "loss": 0.6107,
51
+ "step": 300
52
+ },
53
+ {
54
+ "epoch": 0.2931937172774869,
55
+ "grad_norm": 2.0031681060791016,
56
+ "learning_rate": 1.74e-05,
57
+ "loss": 0.5673,
58
+ "step": 350
59
+ },
60
+ {
61
+ "epoch": 0.33507853403141363,
62
+ "grad_norm": 4.771584510803223,
63
+ "learning_rate": 1.9900000000000003e-05,
64
+ "loss": 0.5566,
65
+ "step": 400
66
+ },
67
+ {
68
+ "epoch": 0.3769633507853403,
69
+ "grad_norm": 9.876974105834961,
70
+ "learning_rate": 2.2400000000000002e-05,
71
+ "loss": 0.5487,
72
+ "step": 450
73
+ },
74
+ {
75
+ "epoch": 0.418848167539267,
76
+ "grad_norm": 2.113546848297119,
77
+ "learning_rate": 2.4900000000000002e-05,
78
+ "loss": 0.5268,
79
+ "step": 500
80
+ },
81
+ {
82
+ "epoch": 0.418848167539267,
83
+ "eval_loss": 0.46557939052581787,
84
+ "eval_runtime": 297.9193,
85
+ "eval_samples_per_second": 28.494,
86
+ "eval_steps_per_second": 3.565,
87
+ "step": 500
88
+ },
89
+ {
90
+ "epoch": 0.4607329842931937,
91
+ "grad_norm": 2.7003655433654785,
92
+ "learning_rate": 2.7400000000000002e-05,
93
+ "loss": 0.5212,
94
+ "step": 550
95
+ },
96
+ {
97
+ "epoch": 0.5026178010471204,
98
+ "grad_norm": 5.649191379547119,
99
+ "learning_rate": 2.9900000000000002e-05,
100
+ "loss": 0.5138,
101
+ "step": 600
102
+ },
103
+ {
104
+ "epoch": 0.5445026178010471,
105
+ "grad_norm": 2.272181749343872,
106
+ "learning_rate": 3.24e-05,
107
+ "loss": 0.5121,
108
+ "step": 650
109
+ },
110
+ {
111
+ "epoch": 0.5863874345549738,
112
+ "grad_norm": 2.957671642303467,
113
+ "learning_rate": 3.49e-05,
114
+ "loss": 0.496,
115
+ "step": 700
116
+ },
117
+ {
118
+ "epoch": 0.6282722513089005,
119
+ "grad_norm": 3.577693462371826,
120
+ "learning_rate": 3.74e-05,
121
+ "loss": 0.4886,
122
+ "step": 750
123
+ },
124
+ {
125
+ "epoch": 0.6701570680628273,
126
+ "grad_norm": 2.466113328933716,
127
+ "learning_rate": 3.99e-05,
128
+ "loss": 0.4879,
129
+ "step": 800
130
+ },
131
+ {
132
+ "epoch": 0.7120418848167539,
133
+ "grad_norm": 2.941208600997925,
134
+ "learning_rate": 4.24e-05,
135
+ "loss": 0.4865,
136
+ "step": 850
137
+ },
138
+ {
139
+ "epoch": 0.7539267015706806,
140
+ "grad_norm": 2.3843142986297607,
141
+ "learning_rate": 4.49e-05,
142
+ "loss": 0.4689,
143
+ "step": 900
144
+ },
145
+ {
146
+ "epoch": 0.7958115183246073,
147
+ "grad_norm": 2.232208728790283,
148
+ "learning_rate": 4.74e-05,
149
+ "loss": 0.4768,
150
+ "step": 950
151
+ },
152
+ {
153
+ "epoch": 0.837696335078534,
154
+ "grad_norm": 2.1702969074249268,
155
+ "learning_rate": 4.99e-05,
156
+ "loss": 0.4583,
157
+ "step": 1000
158
+ },
159
+ {
160
+ "epoch": 0.837696335078534,
161
+ "eval_loss": 0.4044812321662903,
162
+ "eval_runtime": 296.3081,
163
+ "eval_samples_per_second": 28.649,
164
+ "eval_steps_per_second": 3.584,
165
+ "step": 1000
166
+ },
167
+ {
168
+ "epoch": 0.8795811518324608,
169
+ "grad_norm": 2.7613942623138428,
170
+ "learning_rate": 5.2400000000000007e-05,
171
+ "loss": 0.4664,
172
+ "step": 1050
173
+ },
174
+ {
175
+ "epoch": 0.9214659685863874,
176
+ "grad_norm": 2.4752321243286133,
177
+ "learning_rate": 5.4900000000000006e-05,
178
+ "loss": 0.442,
179
+ "step": 1100
180
+ },
181
+ {
182
+ "epoch": 0.9633507853403142,
183
+ "grad_norm": 2.27407169342041,
184
+ "learning_rate": 5.74e-05,
185
+ "loss": 0.4498,
186
+ "step": 1150
187
+ },
188
+ {
189
+ "epoch": 1.0052356020942408,
190
+ "grad_norm": 1.8998862504959106,
191
+ "learning_rate": 5.99e-05,
192
+ "loss": 0.4433,
193
+ "step": 1200
194
+ },
195
+ {
196
+ "epoch": 1.0471204188481675,
197
+ "grad_norm": 2.8479738235473633,
198
+ "learning_rate": 6.24e-05,
199
+ "loss": 0.4417,
200
+ "step": 1250
201
+ },
202
+ {
203
+ "epoch": 1.0890052356020943,
204
+ "grad_norm": 1.9095953702926636,
205
+ "learning_rate": 6.49e-05,
206
+ "loss": 0.444,
207
+ "step": 1300
208
+ },
209
+ {
210
+ "epoch": 1.130890052356021,
211
+ "grad_norm": 1.368077278137207,
212
+ "learning_rate": 6.740000000000001e-05,
213
+ "loss": 0.4519,
214
+ "step": 1350
215
+ },
216
+ {
217
+ "epoch": 1.1727748691099475,
218
+ "grad_norm": 2.0456721782684326,
219
+ "learning_rate": 6.99e-05,
220
+ "loss": 0.4411,
221
+ "step": 1400
222
+ },
223
+ {
224
+ "epoch": 1.2146596858638743,
225
+ "grad_norm": 1.6217619180679321,
226
+ "learning_rate": 7.24e-05,
227
+ "loss": 0.4324,
228
+ "step": 1450
229
+ },
230
+ {
231
+ "epoch": 1.256544502617801,
232
+ "grad_norm": 2.961646795272827,
233
+ "learning_rate": 7.49e-05,
234
+ "loss": 0.4321,
235
+ "step": 1500
236
+ },
237
+ {
238
+ "epoch": 1.256544502617801,
239
+ "eval_loss": 0.3930450975894928,
240
+ "eval_runtime": 301.1428,
241
+ "eval_samples_per_second": 28.189,
242
+ "eval_steps_per_second": 3.527,
243
+ "step": 1500
244
+ },
245
+ {
246
+ "epoch": 1.2984293193717278,
247
+ "grad_norm": 2.543201446533203,
248
+ "learning_rate": 7.740000000000001e-05,
249
+ "loss": 0.4364,
250
+ "step": 1550
251
+ },
252
+ {
253
+ "epoch": 1.3403141361256545,
254
+ "grad_norm": 1.5807560682296753,
255
+ "learning_rate": 7.99e-05,
256
+ "loss": 0.433,
257
+ "step": 1600
258
+ },
259
+ {
260
+ "epoch": 1.3821989528795813,
261
+ "grad_norm": 1.587385892868042,
262
+ "learning_rate": 8.24e-05,
263
+ "loss": 0.4285,
264
+ "step": 1650
265
+ },
266
+ {
267
+ "epoch": 1.4240837696335078,
268
+ "grad_norm": 2.1177687644958496,
269
+ "learning_rate": 8.49e-05,
270
+ "loss": 0.4209,
271
+ "step": 1700
272
+ },
273
+ {
274
+ "epoch": 1.4659685863874345,
275
+ "grad_norm": 1.6078702211380005,
276
+ "learning_rate": 8.740000000000001e-05,
277
+ "loss": 0.426,
278
+ "step": 1750
279
+ },
280
+ {
281
+ "epoch": 1.5078534031413613,
282
+ "grad_norm": 2.10492205619812,
283
+ "learning_rate": 8.99e-05,
284
+ "loss": 0.4191,
285
+ "step": 1800
286
+ },
287
+ {
288
+ "epoch": 1.5497382198952878,
289
+ "grad_norm": 2.169309616088867,
290
+ "learning_rate": 9.240000000000001e-05,
291
+ "loss": 0.4178,
292
+ "step": 1850
293
+ },
294
+ {
295
+ "epoch": 1.5916230366492146,
296
+ "grad_norm": 1.559006929397583,
297
+ "learning_rate": 9.49e-05,
298
+ "loss": 0.4133,
299
+ "step": 1900
300
+ },
301
+ {
302
+ "epoch": 1.6335078534031413,
303
+ "grad_norm": 2.0436346530914307,
304
+ "learning_rate": 9.74e-05,
305
+ "loss": 0.4234,
306
+ "step": 1950
307
+ },
308
+ {
309
+ "epoch": 1.675392670157068,
310
+ "grad_norm": 2.089486837387085,
311
+ "learning_rate": 9.99e-05,
312
+ "loss": 0.4092,
313
+ "step": 2000
314
+ },
315
+ {
316
+ "epoch": 1.675392670157068,
317
+ "eval_loss": 0.37686464190483093,
318
+ "eval_runtime": 310.3113,
319
+ "eval_samples_per_second": 27.356,
320
+ "eval_steps_per_second": 3.422,
321
+ "step": 2000
322
+ },
323
+ {
324
+ "epoch": 1.7179057591623037,
325
+ "grad_norm": 2.684563398361206,
326
+ "learning_rate": 9.952e-05,
327
+ "loss": 0.4099,
328
+ "step": 2050
329
+ },
330
+ {
331
+ "epoch": 1.7597905759162304,
332
+ "grad_norm": 1.7815254926681519,
333
+ "learning_rate": 9.902e-05,
334
+ "loss": 0.4093,
335
+ "step": 2100
336
+ },
337
+ {
338
+ "epoch": 1.8016753926701572,
339
+ "grad_norm": 1.9900892972946167,
340
+ "learning_rate": 9.852e-05,
341
+ "loss": 0.4116,
342
+ "step": 2150
343
+ },
344
+ {
345
+ "epoch": 1.8435602094240837,
346
+ "grad_norm": 1.8637460470199585,
347
+ "learning_rate": 9.802e-05,
348
+ "loss": 0.4089,
349
+ "step": 2200
350
+ },
351
+ {
352
+ "epoch": 1.8854450261780105,
353
+ "grad_norm": 1.58171546459198,
354
+ "learning_rate": 9.752e-05,
355
+ "loss": 0.4053,
356
+ "step": 2250
357
+ },
358
+ {
359
+ "epoch": 1.9273298429319372,
360
+ "grad_norm": 1.4995630979537964,
361
+ "learning_rate": 9.702e-05,
362
+ "loss": 0.4049,
363
+ "step": 2300
364
+ },
365
+ {
366
+ "epoch": 1.9692146596858637,
367
+ "grad_norm": 2.447049379348755,
368
+ "learning_rate": 9.652e-05,
369
+ "loss": 0.4029,
370
+ "step": 2350
371
+ },
372
+ {
373
+ "epoch": 2.0110994764397905,
374
+ "grad_norm": 1.8776074647903442,
375
+ "learning_rate": 9.602e-05,
376
+ "loss": 0.3988,
377
+ "step": 2400
378
+ },
379
+ {
380
+ "epoch": 2.0529842931937172,
381
+ "grad_norm": 3.1785056591033936,
382
+ "learning_rate": 9.552000000000001e-05,
383
+ "loss": 0.4023,
384
+ "step": 2450
385
+ },
386
+ {
387
+ "epoch": 2.094869109947644,
388
+ "grad_norm": 1.4416707754135132,
389
+ "learning_rate": 9.502000000000001e-05,
390
+ "loss": 0.4008,
391
+ "step": 2500
392
+ },
393
+ {
394
+ "epoch": 2.094869109947644,
395
+ "eval_loss": 0.35975897312164307,
396
+ "eval_runtime": 275.2855,
397
+ "eval_samples_per_second": 30.837,
398
+ "eval_steps_per_second": 3.858,
399
+ "step": 2500
400
+ },
401
+ {
402
+ "epoch": 2.1367539267015707,
403
+ "grad_norm": 1.3332245349884033,
404
+ "learning_rate": 9.452000000000001e-05,
405
+ "loss": 0.3985,
406
+ "step": 2550
407
+ },
408
+ {
409
+ "epoch": 2.1786387434554975,
410
+ "grad_norm": 1.291338324546814,
411
+ "learning_rate": 9.402000000000001e-05,
412
+ "loss": 0.4021,
413
+ "step": 2600
414
+ },
415
+ {
416
+ "epoch": 2.220523560209424,
417
+ "grad_norm": 1.6172423362731934,
418
+ "learning_rate": 9.352000000000001e-05,
419
+ "loss": 0.4006,
420
+ "step": 2650
421
+ },
422
+ {
423
+ "epoch": 2.262408376963351,
424
+ "grad_norm": 1.8729867935180664,
425
+ "learning_rate": 9.302e-05,
426
+ "loss": 0.3966,
427
+ "step": 2700
428
+ },
429
+ {
430
+ "epoch": 2.3042931937172773,
431
+ "grad_norm": 1.623901605606079,
432
+ "learning_rate": 9.252e-05,
433
+ "loss": 0.3951,
434
+ "step": 2750
435
+ },
436
+ {
437
+ "epoch": 2.346178010471204,
438
+ "grad_norm": 1.032320261001587,
439
+ "learning_rate": 9.202e-05,
440
+ "loss": 0.3904,
441
+ "step": 2800
442
+ },
443
+ {
444
+ "epoch": 2.3880628272251307,
445
+ "grad_norm": 2.0027778148651123,
446
+ "learning_rate": 9.152e-05,
447
+ "loss": 0.3933,
448
+ "step": 2850
449
+ },
450
+ {
451
+ "epoch": 2.4299476439790575,
452
+ "grad_norm": 1.5681087970733643,
453
+ "learning_rate": 9.102e-05,
454
+ "loss": 0.4024,
455
+ "step": 2900
456
+ },
457
+ {
458
+ "epoch": 2.4718324607329842,
459
+ "grad_norm": 1.2643622159957886,
460
+ "learning_rate": 9.052000000000001e-05,
461
+ "loss": 0.3939,
462
+ "step": 2950
463
+ },
464
+ {
465
+ "epoch": 2.513717277486911,
466
+ "grad_norm": 1.4029465913772583,
467
+ "learning_rate": 9.002000000000001e-05,
468
+ "loss": 0.3941,
469
+ "step": 3000
470
+ },
471
+ {
472
+ "epoch": 2.513717277486911,
473
+ "eval_loss": 0.3533318042755127,
474
+ "eval_runtime": 271.9646,
475
+ "eval_samples_per_second": 31.214,
476
+ "eval_steps_per_second": 3.905,
477
+ "step": 3000
478
+ },
479
+ {
480
+ "epoch": 2.5556020942408377,
481
+ "grad_norm": 1.6669938564300537,
482
+ "learning_rate": 8.952000000000001e-05,
483
+ "loss": 0.3908,
484
+ "step": 3050
485
+ },
486
+ {
487
+ "epoch": 2.5974869109947645,
488
+ "grad_norm": 1.6020203828811646,
489
+ "learning_rate": 8.902e-05,
490
+ "loss": 0.3905,
491
+ "step": 3100
492
+ },
493
+ {
494
+ "epoch": 2.639371727748691,
495
+ "grad_norm": 2.2321176528930664,
496
+ "learning_rate": 8.852e-05,
497
+ "loss": 0.389,
498
+ "step": 3150
499
+ },
500
+ {
501
+ "epoch": 2.681256544502618,
502
+ "grad_norm": 1.0864609479904175,
503
+ "learning_rate": 8.802e-05,
504
+ "loss": 0.3906,
505
+ "step": 3200
506
+ },
507
+ {
508
+ "epoch": 2.7231413612565447,
509
+ "grad_norm": 2.0679004192352295,
510
+ "learning_rate": 8.752e-05,
511
+ "loss": 0.3869,
512
+ "step": 3250
513
+ },
514
+ {
515
+ "epoch": 2.765026178010471,
516
+ "grad_norm": 2.6636288166046143,
517
+ "learning_rate": 8.702e-05,
518
+ "loss": 0.3841,
519
+ "step": 3300
520
+ },
521
+ {
522
+ "epoch": 2.8069109947643978,
523
+ "grad_norm": 1.435549259185791,
524
+ "learning_rate": 8.652e-05,
525
+ "loss": 0.3841,
526
+ "step": 3350
527
+ },
528
+ {
529
+ "epoch": 2.8487958115183245,
530
+ "grad_norm": 2.208448648452759,
531
+ "learning_rate": 8.602e-05,
532
+ "loss": 0.3828,
533
+ "step": 3400
534
+ },
535
+ {
536
+ "epoch": 2.8906806282722513,
537
+ "grad_norm": 1.731292963027954,
538
+ "learning_rate": 8.552e-05,
539
+ "loss": 0.3842,
540
+ "step": 3450
541
+ },
542
+ {
543
+ "epoch": 2.932565445026178,
544
+ "grad_norm": 1.7065523862838745,
545
+ "learning_rate": 8.502e-05,
546
+ "loss": 0.3834,
547
+ "step": 3500
548
+ },
549
+ {
550
+ "epoch": 2.932565445026178,
551
+ "eval_loss": 0.3439213037490845,
552
+ "eval_runtime": 274.3567,
553
+ "eval_samples_per_second": 30.941,
554
+ "eval_steps_per_second": 3.871,
555
+ "step": 3500
556
+ },
557
+ {
558
+ "epoch": 2.9744502617801047,
559
+ "grad_norm": 1.5389653444290161,
560
+ "learning_rate": 8.452e-05,
561
+ "loss": 0.3915,
562
+ "step": 3550
563
+ },
564
+ {
565
+ "epoch": 3.0163350785340315,
566
+ "grad_norm": 1.3524985313415527,
567
+ "learning_rate": 8.402e-05,
568
+ "loss": 0.3843,
569
+ "step": 3600
570
+ },
571
+ {
572
+ "epoch": 3.0582198952879582,
573
+ "grad_norm": 1.5389949083328247,
574
+ "learning_rate": 8.352000000000001e-05,
575
+ "loss": 0.3822,
576
+ "step": 3650
577
+ },
578
+ {
579
+ "epoch": 3.100104712041885,
580
+ "grad_norm": 1.6620662212371826,
581
+ "learning_rate": 8.302000000000001e-05,
582
+ "loss": 0.38,
583
+ "step": 3700
584
+ },
585
+ {
586
+ "epoch": 3.1419895287958117,
587
+ "grad_norm": 1.187556505203247,
588
+ "learning_rate": 8.252e-05,
589
+ "loss": 0.3768,
590
+ "step": 3750
591
+ },
592
+ {
593
+ "epoch": 3.183874345549738,
594
+ "grad_norm": 2.6059765815734863,
595
+ "learning_rate": 8.202e-05,
596
+ "loss": 0.3799,
597
+ "step": 3800
598
+ },
599
+ {
600
+ "epoch": 3.2257591623036648,
601
+ "grad_norm": 1.3912620544433594,
602
+ "learning_rate": 8.152e-05,
603
+ "loss": 0.3776,
604
+ "step": 3850
605
+ },
606
+ {
607
+ "epoch": 3.2676439790575915,
608
+ "grad_norm": 1.4691671133041382,
609
+ "learning_rate": 8.102000000000001e-05,
610
+ "loss": 0.3755,
611
+ "step": 3900
612
+ },
613
+ {
614
+ "epoch": 3.3095287958115183,
615
+ "grad_norm": 1.452143669128418,
616
+ "learning_rate": 8.052000000000001e-05,
617
+ "loss": 0.3821,
618
+ "step": 3950
619
+ },
620
+ {
621
+ "epoch": 3.351413612565445,
622
+ "grad_norm": 0.9386736750602722,
623
+ "learning_rate": 8.002000000000001e-05,
624
+ "loss": 0.3699,
625
+ "step": 4000
626
+ },
627
+ {
628
+ "epoch": 3.351413612565445,
629
+ "eval_loss": 0.342171847820282,
630
+ "eval_runtime": 268.7667,
631
+ "eval_samples_per_second": 31.585,
632
+ "eval_steps_per_second": 3.951,
633
+ "step": 4000
634
+ },
635
+ {
636
+ "epoch": 3.3932984293193718,
637
+ "grad_norm": 1.6316275596618652,
638
+ "learning_rate": 7.952000000000001e-05,
639
+ "loss": 0.3741,
640
+ "step": 4050
641
+ },
642
+ {
643
+ "epoch": 3.4351832460732985,
644
+ "grad_norm": 1.2369744777679443,
645
+ "learning_rate": 7.902e-05,
646
+ "loss": 0.3728,
647
+ "step": 4100
648
+ },
649
+ {
650
+ "epoch": 3.4770680628272252,
651
+ "grad_norm": 1.48318350315094,
652
+ "learning_rate": 7.853000000000001e-05,
653
+ "loss": 0.3772,
654
+ "step": 4150
655
+ },
656
+ {
657
+ "epoch": 3.518952879581152,
658
+ "grad_norm": 1.4171572923660278,
659
+ "learning_rate": 7.803e-05,
660
+ "loss": 0.3717,
661
+ "step": 4200
662
+ },
663
+ {
664
+ "epoch": 3.5608376963350787,
665
+ "grad_norm": 1.5374139547348022,
666
+ "learning_rate": 7.753e-05,
667
+ "loss": 0.3774,
668
+ "step": 4250
669
+ },
670
+ {
671
+ "epoch": 3.602722513089005,
672
+ "grad_norm": 1.6316169500350952,
673
+ "learning_rate": 7.703e-05,
674
+ "loss": 0.3731,
675
+ "step": 4300
676
+ },
677
+ {
678
+ "epoch": 3.644607329842932,
679
+ "grad_norm": 1.749531865119934,
680
+ "learning_rate": 7.653e-05,
681
+ "loss": 0.3729,
682
+ "step": 4350
683
+ },
684
+ {
685
+ "epoch": 3.6864921465968585,
686
+ "grad_norm": 1.4812896251678467,
687
+ "learning_rate": 7.603e-05,
688
+ "loss": 0.375,
689
+ "step": 4400
690
+ },
691
+ {
692
+ "epoch": 3.7283769633507853,
693
+ "grad_norm": 1.9387775659561157,
694
+ "learning_rate": 7.553e-05,
695
+ "loss": 0.375,
696
+ "step": 4450
697
+ },
698
+ {
699
+ "epoch": 3.770261780104712,
700
+ "grad_norm": 1.4758845567703247,
701
+ "learning_rate": 7.503e-05,
702
+ "loss": 0.3774,
703
+ "step": 4500
704
+ },
705
+ {
706
+ "epoch": 3.770261780104712,
707
+ "eval_loss": 0.3396616578102112,
708
+ "eval_runtime": 265.2963,
709
+ "eval_samples_per_second": 31.998,
710
+ "eval_steps_per_second": 4.003,
711
+ "step": 4500
712
+ }
713
+ ],
714
+ "logging_steps": 50,
715
+ "max_steps": 12000,
716
+ "num_input_tokens_seen": 0,
717
+ "num_train_epochs": 11,
718
+ "save_steps": 500,
719
+ "stateful_callbacks": {
720
+ "TrainerControl": {
721
+ "args": {
722
+ "should_epoch_stop": false,
723
+ "should_evaluate": false,
724
+ "should_log": false,
725
+ "should_save": true,
726
+ "should_training_stop": false
727
+ },
728
+ "attributes": {}
729
+ }
730
+ },
731
+ "total_flos": 4.040100442105517e+16,
732
+ "train_batch_size": 16,
733
+ "trial_name": null,
734
+ "trial_params": null
735
+ }
checkpoint-4500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa8f41edfc5db87ff9f21c8561c4e1b9e9e17c2b5ee34213317807a5e0e41371
3
+ size 5304