tmnam20 commited on
Commit
0e9b9a4
1 Parent(s): 033a144

Training in progress, step 102000, checkpoint

Browse files
checkpoint-102000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-102000/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai-gpt",
3
+ "afn": "gelu",
4
+ "architectures": [
5
+ "OpenAIGPTForSequenceClassification"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "embd_pdrop": 0.1,
9
+ "id2label": {
10
+ "0": "entailment",
11
+ "1": "neutral",
12
+ "2": "contradiction"
13
+ },
14
+ "initializer_range": 0.02,
15
+ "label2id": {
16
+ "contradiction": 2,
17
+ "entailment": 0,
18
+ "neutral": 1
19
+ },
20
+ "layer_norm_epsilon": 1e-05,
21
+ "model_type": "openai-gpt",
22
+ "n_ctx": 512,
23
+ "n_embd": 768,
24
+ "n_head": 12,
25
+ "n_layer": 2,
26
+ "n_positions": 512,
27
+ "n_special": 0,
28
+ "pad_token_id": 40480,
29
+ "predict_special_tokens": true,
30
+ "problem_type": "single_label_classification",
31
+ "resid_pdrop": 0.1,
32
+ "summary_activation": null,
33
+ "summary_first_dropout": 0.1,
34
+ "summary_proj_to_labels": true,
35
+ "summary_type": "last",
36
+ "summary_use_proj": true,
37
+ "task_specific_params": {
38
+ "text-generation": {
39
+ "do_sample": true,
40
+ "max_length": 50
41
+ }
42
+ },
43
+ "torch_dtype": "float32",
44
+ "transformers_version": "4.33.0.dev0",
45
+ "vocab_size": 40481
46
+ }
checkpoint-102000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-102000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:acc819a2cec4fb995c7191a1afd895990da1f48f9d7ebc8ea32ef577daaa1f8f
3
+ size 365301299
checkpoint-102000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dbfd23dec6d4de3f960c73aa4365fe2ea3a0810d5c65cc9bcaca8bbcdaf69fd3
3
+ size 182651202
checkpoint-102000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0d10d0afdf9ba81db3936536253f38e16519f520e0fcad00df5e800bc9dbbd5
3
+ size 14575
checkpoint-102000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55be665208fa93994a0719f6cd16d99d8b2648f55763f86c04258cc324fdc73b
3
+ size 627
checkpoint-102000/special_tokens_map.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<bos>",
3
+ "eos_token": "<eos>",
4
+ "pad_token": "<pad>",
5
+ "unk_token": "<unk>"
6
+ }
checkpoint-102000/tokenizer_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "model_max_length": 512,
4
+ "padding_side": "left",
5
+ "tokenizer_class": "OpenAIGPTTokenizer",
6
+ "unk_token": "<unk>"
7
+ }
checkpoint-102000/trainer_state.json ADDED
@@ -0,0 +1,1555 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9834555387496948,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-100000",
4
+ "epoch": 4.15580182529335,
5
+ "eval_steps": 1000,
6
+ "global_step": 102000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "learning_rate": 2.7777777777777777e-11,
14
+ "loss": 1.1078,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.128,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.330938123752495,
26
+ "eval_loss": 1.1083118915557861,
27
+ "eval_runtime": 12.3979,
28
+ "eval_samples_per_second": 404.1,
29
+ "eval_steps_per_second": 50.573,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1164,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.3323353293413174,
41
+ "eval_loss": 1.102672815322876,
42
+ "eval_runtime": 12.2861,
43
+ "eval_samples_per_second": 407.778,
44
+ "eval_steps_per_second": 51.033,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1158,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.34211576846307384,
56
+ "eval_loss": 1.1024774312973022,
57
+ "eval_runtime": 12.3746,
58
+ "eval_samples_per_second": 404.863,
59
+ "eval_steps_per_second": 50.668,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1115,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.34890219560878244,
71
+ "eval_loss": 1.1012382507324219,
72
+ "eval_runtime": 12.5405,
73
+ "eval_samples_per_second": 399.505,
74
+ "eval_steps_per_second": 49.998,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3880555555555558e-07,
80
+ "loss": 1.1141,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.3534930139720559,
86
+ "eval_loss": 1.1002192497253418,
87
+ "eval_runtime": 12.3001,
88
+ "eval_samples_per_second": 407.314,
89
+ "eval_steps_per_second": 50.975,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.112,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3524950099800399,
101
+ "eval_loss": 1.100056767463684,
102
+ "eval_runtime": 12.4062,
103
+ "eval_samples_per_second": 403.831,
104
+ "eval_steps_per_second": 50.539,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9433333333333334e-07,
110
+ "loss": 1.1115,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.34331337325349304,
116
+ "eval_loss": 1.098986029624939,
117
+ "eval_runtime": 12.4456,
118
+ "eval_samples_per_second": 402.551,
119
+ "eval_steps_per_second": 50.379,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2211111111111114e-07,
125
+ "loss": 1.109,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.3347305389221557,
131
+ "eval_loss": 1.098678469657898,
132
+ "eval_runtime": 12.5978,
133
+ "eval_samples_per_second": 397.688,
134
+ "eval_steps_per_second": 49.771,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4986111111111113e-07,
140
+ "loss": 1.1117,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34311377245508984,
146
+ "eval_loss": 1.0984419584274292,
147
+ "eval_runtime": 12.313,
148
+ "eval_samples_per_second": 406.886,
149
+ "eval_steps_per_second": 50.922,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776388888888889e-07,
155
+ "loss": 1.1081,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.35429141716566864,
161
+ "eval_loss": 1.0968743562698364,
162
+ "eval_runtime": 12.4088,
163
+ "eval_samples_per_second": 403.744,
164
+ "eval_steps_per_second": 50.528,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.053888888888889e-07,
170
+ "loss": 1.107,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.3722554890219561,
176
+ "eval_loss": 1.0950665473937988,
177
+ "eval_runtime": 12.5775,
178
+ "eval_samples_per_second": 398.33,
179
+ "eval_steps_per_second": 49.851,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.331666666666667e-07,
185
+ "loss": 1.1057,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.37544910179640717,
191
+ "eval_loss": 1.0940738916397095,
192
+ "eval_runtime": 12.5186,
193
+ "eval_samples_per_second": 400.203,
194
+ "eval_steps_per_second": 50.085,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.609166666666667e-07,
200
+ "loss": 1.1069,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3401197604790419,
206
+ "eval_loss": 1.0937162637710571,
207
+ "eval_runtime": 12.2594,
208
+ "eval_samples_per_second": 408.666,
209
+ "eval_steps_per_second": 51.144,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.886944444444445e-07,
215
+ "loss": 1.1046,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.37604790419161677,
221
+ "eval_loss": 1.0926916599273682,
222
+ "eval_runtime": 12.256,
223
+ "eval_samples_per_second": 408.78,
224
+ "eval_steps_per_second": 51.159,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.164444444444445e-07,
230
+ "loss": 1.1046,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.39201596806387223,
236
+ "eval_loss": 1.0912976264953613,
237
+ "eval_runtime": 12.5019,
238
+ "eval_samples_per_second": 400.74,
239
+ "eval_steps_per_second": 50.152,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.442222222222223e-07,
245
+ "loss": 1.1052,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.393812375249501,
251
+ "eval_loss": 1.0901482105255127,
252
+ "eval_runtime": 12.3407,
253
+ "eval_samples_per_second": 405.973,
254
+ "eval_steps_per_second": 50.807,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7197222222222224e-07,
260
+ "loss": 1.1011,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.3932135728542914,
266
+ "eval_loss": 1.0901678800582886,
267
+ "eval_runtime": 12.3644,
268
+ "eval_samples_per_second": 405.197,
269
+ "eval_steps_per_second": 50.71,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.997500000000001e-07,
275
+ "loss": 1.1,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.38143712574850297,
281
+ "eval_loss": 1.0897998809814453,
282
+ "eval_runtime": 12.3516,
283
+ "eval_samples_per_second": 405.615,
284
+ "eval_steps_per_second": 50.763,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275277777777778e-07,
290
+ "loss": 1.1007,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.40678642714570856,
296
+ "eval_loss": 1.0876121520996094,
297
+ "eval_runtime": 12.6608,
298
+ "eval_samples_per_second": 395.708,
299
+ "eval_steps_per_second": 49.523,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.552777777777778e-07,
305
+ "loss": 1.1013,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4053892215568862,
311
+ "eval_loss": 1.0876350402832031,
312
+ "eval_runtime": 12.3243,
313
+ "eval_samples_per_second": 406.513,
314
+ "eval_steps_per_second": 50.875,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830555555555556e-07,
320
+ "loss": 1.0987,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.40858283433133735,
326
+ "eval_loss": 1.085940957069397,
327
+ "eval_runtime": 12.3217,
328
+ "eval_samples_per_second": 406.598,
329
+ "eval_steps_per_second": 50.886,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108055555555556e-07,
335
+ "loss": 1.0986,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.3972055888223553,
341
+ "eval_loss": 1.085978388786316,
342
+ "eval_runtime": 12.4285,
343
+ "eval_samples_per_second": 403.107,
344
+ "eval_steps_per_second": 50.449,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.385833333333334e-07,
350
+ "loss": 1.0944,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.41097804391217563,
356
+ "eval_loss": 1.0837496519088745,
357
+ "eval_runtime": 12.7175,
358
+ "eval_samples_per_second": 393.946,
359
+ "eval_steps_per_second": 49.302,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663611111111112e-07,
365
+ "loss": 1.095,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.4121756487025948,
371
+ "eval_loss": 1.0820704698562622,
372
+ "eval_runtime": 12.3323,
373
+ "eval_samples_per_second": 406.251,
374
+ "eval_steps_per_second": 50.842,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941111111111112e-07,
380
+ "loss": 1.0915,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4177644710578842,
386
+ "eval_loss": 1.0762931108474731,
387
+ "eval_runtime": 12.3887,
388
+ "eval_samples_per_second": 404.399,
389
+ "eval_steps_per_second": 50.61,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.218888888888889e-07,
395
+ "loss": 1.0889,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.43253493013972055,
401
+ "eval_loss": 1.0695326328277588,
402
+ "eval_runtime": 12.5611,
403
+ "eval_samples_per_second": 398.85,
404
+ "eval_steps_per_second": 49.916,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.49638888888889e-07,
410
+ "loss": 1.0827,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4343313373253493,
416
+ "eval_loss": 1.0662548542022705,
417
+ "eval_runtime": 12.641,
418
+ "eval_samples_per_second": 396.328,
419
+ "eval_steps_per_second": 49.6,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774166666666668e-07,
425
+ "loss": 1.0765,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.4315369261477046,
431
+ "eval_loss": 1.0609794855117798,
432
+ "eval_runtime": 12.4637,
433
+ "eval_samples_per_second": 401.967,
434
+ "eval_steps_per_second": 50.306,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.051666666666667e-07,
440
+ "loss": 1.0736,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.4377245508982036,
446
+ "eval_loss": 1.0587859153747559,
447
+ "eval_runtime": 12.3801,
448
+ "eval_samples_per_second": 404.683,
449
+ "eval_steps_per_second": 50.646,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.329444444444445e-07,
455
+ "loss": 1.0726,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.4249500998003992,
461
+ "eval_loss": 1.0559338331222534,
462
+ "eval_runtime": 12.5945,
463
+ "eval_samples_per_second": 397.792,
464
+ "eval_steps_per_second": 49.784,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.606944444444445e-07,
470
+ "loss": 1.0751,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.4427145708582834,
476
+ "eval_loss": 1.0576916933059692,
477
+ "eval_runtime": 12.6324,
478
+ "eval_samples_per_second": 396.599,
479
+ "eval_steps_per_second": 49.634,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.884722222222224e-07,
485
+ "loss": 1.0708,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.42634730538922155,
491
+ "eval_loss": 1.0571542978286743,
492
+ "eval_runtime": 12.4696,
493
+ "eval_samples_per_second": 401.776,
494
+ "eval_steps_per_second": 50.282,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.162222222222223e-07,
500
+ "loss": 1.0711,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.4419161676646707,
506
+ "eval_loss": 1.0544849634170532,
507
+ "eval_runtime": 12.415,
508
+ "eval_samples_per_second": 403.544,
509
+ "eval_steps_per_second": 50.503,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440000000000001e-07,
515
+ "loss": 1.0666,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.43333333333333335,
521
+ "eval_loss": 1.05018150806427,
522
+ "eval_runtime": 12.6226,
523
+ "eval_samples_per_second": 396.906,
524
+ "eval_steps_per_second": 49.673,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.7175e-07,
530
+ "loss": 1.0698,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.4377245508982036,
536
+ "eval_loss": 1.0488406419754028,
537
+ "eval_runtime": 12.6246,
538
+ "eval_samples_per_second": 396.844,
539
+ "eval_steps_per_second": 49.665,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995277777777778e-07,
545
+ "loss": 1.067,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.4407185628742515,
551
+ "eval_loss": 1.047359824180603,
552
+ "eval_runtime": 12.4039,
553
+ "eval_samples_per_second": 403.906,
554
+ "eval_steps_per_second": 50.549,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.027277777777778e-06,
560
+ "loss": 1.0636,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.44451097804391215,
566
+ "eval_loss": 1.0455204248428345,
567
+ "eval_runtime": 12.3836,
568
+ "eval_samples_per_second": 404.569,
569
+ "eval_steps_per_second": 50.632,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.0550555555555557e-06,
575
+ "loss": 1.0658,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.4407185628742515,
581
+ "eval_loss": 1.0461794137954712,
582
+ "eval_runtime": 12.5546,
583
+ "eval_samples_per_second": 399.057,
584
+ "eval_steps_per_second": 49.942,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0828333333333334e-06,
590
+ "loss": 1.0635,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.4367265469061876,
596
+ "eval_loss": 1.0461270809173584,
597
+ "eval_runtime": 12.5509,
598
+ "eval_samples_per_second": 399.173,
599
+ "eval_steps_per_second": 49.956,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1105833333333335e-06,
605
+ "loss": 1.0596,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.449500998003992,
611
+ "eval_loss": 1.0418727397918701,
612
+ "eval_runtime": 12.4044,
613
+ "eval_samples_per_second": 403.889,
614
+ "eval_steps_per_second": 50.547,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.1383611111111113e-06,
620
+ "loss": 1.0595,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.4409181636726547,
626
+ "eval_loss": 1.0417557954788208,
627
+ "eval_runtime": 12.3642,
628
+ "eval_samples_per_second": 405.202,
629
+ "eval_steps_per_second": 50.711,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.1661111111111111e-06,
635
+ "loss": 1.054,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.4357285429141717,
641
+ "eval_loss": 1.0383535623550415,
642
+ "eval_runtime": 12.3634,
643
+ "eval_samples_per_second": 405.23,
644
+ "eval_steps_per_second": 50.714,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.193888888888889e-06,
650
+ "loss": 1.0543,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.43812375249500995,
656
+ "eval_loss": 1.0463169813156128,
657
+ "eval_runtime": 12.6398,
658
+ "eval_samples_per_second": 396.366,
659
+ "eval_steps_per_second": 49.605,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.221638888888889e-06,
665
+ "loss": 1.0579,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4415169660678643,
671
+ "eval_loss": 1.0388957262039185,
672
+ "eval_runtime": 12.4542,
673
+ "eval_samples_per_second": 402.274,
674
+ "eval_steps_per_second": 50.344,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2494166666666668e-06,
680
+ "loss": 1.0561,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.46187624750499,
686
+ "eval_loss": 1.031552791595459,
687
+ "eval_runtime": 12.4082,
688
+ "eval_samples_per_second": 403.764,
689
+ "eval_steps_per_second": 50.531,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2771666666666668e-06,
695
+ "loss": 1.0516,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.4538922155688623,
701
+ "eval_loss": 1.0312999486923218,
702
+ "eval_runtime": 12.3695,
703
+ "eval_samples_per_second": 405.029,
704
+ "eval_steps_per_second": 50.689,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3049444444444446e-06,
710
+ "loss": 1.0471,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.46387225548902195,
716
+ "eval_loss": 1.027535319328308,
717
+ "eval_runtime": 12.5846,
718
+ "eval_samples_per_second": 398.107,
719
+ "eval_steps_per_second": 49.823,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.3326944444444447e-06,
725
+ "loss": 1.054,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.4568862275449102,
731
+ "eval_loss": 1.0308438539505005,
732
+ "eval_runtime": 12.5646,
733
+ "eval_samples_per_second": 398.74,
734
+ "eval_steps_per_second": 49.902,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3604722222222224e-06,
740
+ "loss": 1.0445,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.47045908183632734,
746
+ "eval_loss": 1.0230755805969238,
747
+ "eval_runtime": 12.4305,
748
+ "eval_samples_per_second": 403.041,
749
+ "eval_steps_per_second": 50.44,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.3882222222222223e-06,
755
+ "loss": 1.0479,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.47305389221556887,
761
+ "eval_loss": 1.0201445817947388,
762
+ "eval_runtime": 12.3672,
763
+ "eval_samples_per_second": 405.103,
764
+ "eval_steps_per_second": 50.698,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.416e-06,
770
+ "loss": 1.0482,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.4678642714570858,
776
+ "eval_loss": 1.0167714357376099,
777
+ "eval_runtime": 12.5982,
778
+ "eval_samples_per_second": 397.677,
779
+ "eval_steps_per_second": 49.769,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4437500000000002e-06,
785
+ "loss": 1.0406,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.4722554890219561,
791
+ "eval_loss": 1.0175116062164307,
792
+ "eval_runtime": 12.6041,
793
+ "eval_samples_per_second": 397.49,
794
+ "eval_steps_per_second": 49.746,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.471527777777778e-06,
800
+ "loss": 1.0457,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.4748502994011976,
806
+ "eval_loss": 1.010857343673706,
807
+ "eval_runtime": 12.3728,
808
+ "eval_samples_per_second": 404.921,
809
+ "eval_steps_per_second": 50.676,
810
+ "step": 53000
811
+ },
812
+ {
813
+ "epoch": 2.2,
814
+ "learning_rate": 1.4993055555555557e-06,
815
+ "loss": 1.0399,
816
+ "step": 54000
817
+ },
818
+ {
819
+ "epoch": 2.2,
820
+ "eval_accuracy": 0.48522954091816367,
821
+ "eval_loss": 1.0106219053268433,
822
+ "eval_runtime": 12.3952,
823
+ "eval_samples_per_second": 404.188,
824
+ "eval_steps_per_second": 50.584,
825
+ "step": 54000
826
+ },
827
+ {
828
+ "epoch": 2.24,
829
+ "learning_rate": 1.5270555555555558e-06,
830
+ "loss": 1.0412,
831
+ "step": 55000
832
+ },
833
+ {
834
+ "epoch": 2.24,
835
+ "eval_accuracy": 0.4750499001996008,
836
+ "eval_loss": 1.0111658573150635,
837
+ "eval_runtime": 12.4858,
838
+ "eval_samples_per_second": 401.257,
839
+ "eval_steps_per_second": 50.217,
840
+ "step": 55000
841
+ },
842
+ {
843
+ "epoch": 2.28,
844
+ "learning_rate": 1.5548333333333335e-06,
845
+ "loss": 1.0347,
846
+ "step": 56000
847
+ },
848
+ {
849
+ "epoch": 2.28,
850
+ "eval_accuracy": 0.47924151696606787,
851
+ "eval_loss": 1.0121456384658813,
852
+ "eval_runtime": 12.6206,
853
+ "eval_samples_per_second": 396.97,
854
+ "eval_steps_per_second": 49.681,
855
+ "step": 56000
856
+ },
857
+ {
858
+ "epoch": 2.32,
859
+ "learning_rate": 1.5825833333333334e-06,
860
+ "loss": 1.0324,
861
+ "step": 57000
862
+ },
863
+ {
864
+ "epoch": 2.32,
865
+ "eval_accuracy": 0.4738522954091816,
866
+ "eval_loss": 1.0100611448287964,
867
+ "eval_runtime": 12.3748,
868
+ "eval_samples_per_second": 404.857,
869
+ "eval_steps_per_second": 50.668,
870
+ "step": 57000
871
+ },
872
+ {
873
+ "epoch": 2.36,
874
+ "learning_rate": 1.6103611111111112e-06,
875
+ "loss": 1.0346,
876
+ "step": 58000
877
+ },
878
+ {
879
+ "epoch": 2.36,
880
+ "eval_accuracy": 0.4694610778443114,
881
+ "eval_loss": 1.0124626159667969,
882
+ "eval_runtime": 12.4716,
883
+ "eval_samples_per_second": 401.711,
884
+ "eval_steps_per_second": 50.274,
885
+ "step": 58000
886
+ },
887
+ {
888
+ "epoch": 2.4,
889
+ "learning_rate": 1.638111111111111e-06,
890
+ "loss": 1.0309,
891
+ "step": 59000
892
+ },
893
+ {
894
+ "epoch": 2.4,
895
+ "eval_accuracy": 0.48982035928143713,
896
+ "eval_loss": 1.0022122859954834,
897
+ "eval_runtime": 12.4734,
898
+ "eval_samples_per_second": 401.655,
899
+ "eval_steps_per_second": 50.267,
900
+ "step": 59000
901
+ },
902
+ {
903
+ "epoch": 2.44,
904
+ "learning_rate": 1.665888888888889e-06,
905
+ "loss": 1.0351,
906
+ "step": 60000
907
+ },
908
+ {
909
+ "epoch": 2.44,
910
+ "eval_accuracy": 0.4802395209580838,
911
+ "eval_loss": 1.004167914390564,
912
+ "eval_runtime": 12.6287,
913
+ "eval_samples_per_second": 396.714,
914
+ "eval_steps_per_second": 49.649,
915
+ "step": 60000
916
+ },
917
+ {
918
+ "epoch": 2.49,
919
+ "learning_rate": 1.693638888888889e-06,
920
+ "loss": 1.0312,
921
+ "step": 61000
922
+ },
923
+ {
924
+ "epoch": 2.49,
925
+ "eval_accuracy": 0.48323353293413174,
926
+ "eval_loss": 1.0065865516662598,
927
+ "eval_runtime": 12.4143,
928
+ "eval_samples_per_second": 403.567,
929
+ "eval_steps_per_second": 50.506,
930
+ "step": 61000
931
+ },
932
+ {
933
+ "epoch": 2.53,
934
+ "learning_rate": 1.7214166666666666e-06,
935
+ "loss": 1.0329,
936
+ "step": 62000
937
+ },
938
+ {
939
+ "epoch": 2.53,
940
+ "eval_accuracy": 0.47924151696606787,
941
+ "eval_loss": 1.0041630268096924,
942
+ "eval_runtime": 12.4326,
943
+ "eval_samples_per_second": 402.974,
944
+ "eval_steps_per_second": 50.432,
945
+ "step": 62000
946
+ },
947
+ {
948
+ "epoch": 2.57,
949
+ "learning_rate": 1.7491666666666667e-06,
950
+ "loss": 1.0355,
951
+ "step": 63000
952
+ },
953
+ {
954
+ "epoch": 2.57,
955
+ "eval_accuracy": 0.49261477045908186,
956
+ "eval_loss": 1.0007619857788086,
957
+ "eval_runtime": 12.4051,
958
+ "eval_samples_per_second": 403.866,
959
+ "eval_steps_per_second": 50.544,
960
+ "step": 63000
961
+ },
962
+ {
963
+ "epoch": 2.61,
964
+ "learning_rate": 1.7769444444444447e-06,
965
+ "loss": 1.0297,
966
+ "step": 64000
967
+ },
968
+ {
969
+ "epoch": 2.61,
970
+ "eval_accuracy": 0.4942115768463074,
971
+ "eval_loss": 0.9986936450004578,
972
+ "eval_runtime": 12.5727,
973
+ "eval_samples_per_second": 398.482,
974
+ "eval_steps_per_second": 49.87,
975
+ "step": 64000
976
+ },
977
+ {
978
+ "epoch": 2.65,
979
+ "learning_rate": 1.8046944444444446e-06,
980
+ "loss": 1.0277,
981
+ "step": 65000
982
+ },
983
+ {
984
+ "epoch": 2.65,
985
+ "eval_accuracy": 0.49121756487025947,
986
+ "eval_loss": 1.0008972883224487,
987
+ "eval_runtime": 12.4544,
988
+ "eval_samples_per_second": 402.267,
989
+ "eval_steps_per_second": 50.344,
990
+ "step": 65000
991
+ },
992
+ {
993
+ "epoch": 2.69,
994
+ "learning_rate": 1.8324722222222223e-06,
995
+ "loss": 1.0341,
996
+ "step": 66000
997
+ },
998
+ {
999
+ "epoch": 2.69,
1000
+ "eval_accuracy": 0.499001996007984,
1001
+ "eval_loss": 0.9953575134277344,
1002
+ "eval_runtime": 12.4973,
1003
+ "eval_samples_per_second": 400.886,
1004
+ "eval_steps_per_second": 50.171,
1005
+ "step": 66000
1006
+ },
1007
+ {
1008
+ "epoch": 2.73,
1009
+ "learning_rate": 1.8602222222222222e-06,
1010
+ "loss": 1.0279,
1011
+ "step": 67000
1012
+ },
1013
+ {
1014
+ "epoch": 2.73,
1015
+ "eval_accuracy": 0.493812375249501,
1016
+ "eval_loss": 0.998134195804596,
1017
+ "eval_runtime": 12.6589,
1018
+ "eval_samples_per_second": 395.769,
1019
+ "eval_steps_per_second": 49.53,
1020
+ "step": 67000
1021
+ },
1022
+ {
1023
+ "epoch": 2.77,
1024
+ "learning_rate": 1.8880000000000002e-06,
1025
+ "loss": 1.03,
1026
+ "step": 68000
1027
+ },
1028
+ {
1029
+ "epoch": 2.77,
1030
+ "eval_accuracy": 0.5051896207584831,
1031
+ "eval_loss": 0.9947481751441956,
1032
+ "eval_runtime": 12.7939,
1033
+ "eval_samples_per_second": 391.594,
1034
+ "eval_steps_per_second": 49.008,
1035
+ "step": 68000
1036
+ },
1037
+ {
1038
+ "epoch": 2.81,
1039
+ "learning_rate": 1.9157500000000003e-06,
1040
+ "loss": 1.0283,
1041
+ "step": 69000
1042
+ },
1043
+ {
1044
+ "epoch": 2.81,
1045
+ "eval_accuracy": 0.5001996007984032,
1046
+ "eval_loss": 0.997833251953125,
1047
+ "eval_runtime": 12.8129,
1048
+ "eval_samples_per_second": 391.012,
1049
+ "eval_steps_per_second": 48.935,
1050
+ "step": 69000
1051
+ },
1052
+ {
1053
+ "epoch": 2.85,
1054
+ "learning_rate": 1.943527777777778e-06,
1055
+ "loss": 1.0288,
1056
+ "step": 70000
1057
+ },
1058
+ {
1059
+ "epoch": 2.85,
1060
+ "eval_accuracy": 0.49181636726546907,
1061
+ "eval_loss": 1.00169038772583,
1062
+ "eval_runtime": 12.5936,
1063
+ "eval_samples_per_second": 397.822,
1064
+ "eval_steps_per_second": 49.787,
1065
+ "step": 70000
1066
+ },
1067
+ {
1068
+ "epoch": 2.89,
1069
+ "learning_rate": 1.9712777777777777e-06,
1070
+ "loss": 1.0238,
1071
+ "step": 71000
1072
+ },
1073
+ {
1074
+ "epoch": 2.89,
1075
+ "eval_accuracy": 0.4944111776447106,
1076
+ "eval_loss": 1.001316785812378,
1077
+ "eval_runtime": 12.485,
1078
+ "eval_samples_per_second": 401.282,
1079
+ "eval_steps_per_second": 50.22,
1080
+ "step": 71000
1081
+ },
1082
+ {
1083
+ "epoch": 2.93,
1084
+ "learning_rate": 1.9990555555555557e-06,
1085
+ "loss": 1.0268,
1086
+ "step": 72000
1087
+ },
1088
+ {
1089
+ "epoch": 2.93,
1090
+ "eval_accuracy": 0.49241516966067866,
1091
+ "eval_loss": 1.0005468130111694,
1092
+ "eval_runtime": 12.501,
1093
+ "eval_samples_per_second": 400.768,
1094
+ "eval_steps_per_second": 50.156,
1095
+ "step": 72000
1096
+ },
1097
+ {
1098
+ "epoch": 2.97,
1099
+ "learning_rate": 2.026805555555556e-06,
1100
+ "loss": 1.0266,
1101
+ "step": 73000
1102
+ },
1103
+ {
1104
+ "epoch": 2.97,
1105
+ "eval_accuracy": 0.4940119760479042,
1106
+ "eval_loss": 0.9998964071273804,
1107
+ "eval_runtime": 12.6227,
1108
+ "eval_samples_per_second": 396.904,
1109
+ "eval_steps_per_second": 49.672,
1110
+ "step": 73000
1111
+ },
1112
+ {
1113
+ "epoch": 3.01,
1114
+ "learning_rate": 2.0545833333333335e-06,
1115
+ "loss": 1.0222,
1116
+ "step": 74000
1117
+ },
1118
+ {
1119
+ "epoch": 3.01,
1120
+ "eval_accuracy": 0.501996007984032,
1121
+ "eval_loss": 0.9927905201911926,
1122
+ "eval_runtime": 12.7924,
1123
+ "eval_samples_per_second": 391.64,
1124
+ "eval_steps_per_second": 49.014,
1125
+ "step": 74000
1126
+ },
1127
+ {
1128
+ "epoch": 3.06,
1129
+ "learning_rate": 2.0823333333333334e-06,
1130
+ "loss": 1.0198,
1131
+ "step": 75000
1132
+ },
1133
+ {
1134
+ "epoch": 3.06,
1135
+ "eval_accuracy": 0.5025948103792415,
1136
+ "eval_loss": 0.9925487041473389,
1137
+ "eval_runtime": 12.6297,
1138
+ "eval_samples_per_second": 396.684,
1139
+ "eval_steps_per_second": 49.645,
1140
+ "step": 75000
1141
+ },
1142
+ {
1143
+ "epoch": 3.1,
1144
+ "learning_rate": 2.1101111111111113e-06,
1145
+ "loss": 1.0203,
1146
+ "step": 76000
1147
+ },
1148
+ {
1149
+ "epoch": 3.1,
1150
+ "eval_accuracy": 0.5027944111776447,
1151
+ "eval_loss": 0.9909006953239441,
1152
+ "eval_runtime": 12.4265,
1153
+ "eval_samples_per_second": 403.17,
1154
+ "eval_steps_per_second": 50.457,
1155
+ "step": 76000
1156
+ },
1157
+ {
1158
+ "epoch": 3.14,
1159
+ "learning_rate": 2.137861111111111e-06,
1160
+ "loss": 1.022,
1161
+ "step": 77000
1162
+ },
1163
+ {
1164
+ "epoch": 3.14,
1165
+ "eval_accuracy": 0.48303393213572854,
1166
+ "eval_loss": 1.010209083557129,
1167
+ "eval_runtime": 12.3999,
1168
+ "eval_samples_per_second": 404.037,
1169
+ "eval_steps_per_second": 50.565,
1170
+ "step": 77000
1171
+ },
1172
+ {
1173
+ "epoch": 3.18,
1174
+ "learning_rate": 2.1656388888888888e-06,
1175
+ "loss": 1.0133,
1176
+ "step": 78000
1177
+ },
1178
+ {
1179
+ "epoch": 3.18,
1180
+ "eval_accuracy": 0.5075848303393213,
1181
+ "eval_loss": 0.9907119274139404,
1182
+ "eval_runtime": 12.4429,
1183
+ "eval_samples_per_second": 402.64,
1184
+ "eval_steps_per_second": 50.39,
1185
+ "step": 78000
1186
+ },
1187
+ {
1188
+ "epoch": 3.22,
1189
+ "learning_rate": 2.193388888888889e-06,
1190
+ "loss": 1.0207,
1191
+ "step": 79000
1192
+ },
1193
+ {
1194
+ "epoch": 3.22,
1195
+ "eval_accuracy": 0.4944111776447106,
1196
+ "eval_loss": 1.0015305280685425,
1197
+ "eval_runtime": 12.7143,
1198
+ "eval_samples_per_second": 394.046,
1199
+ "eval_steps_per_second": 49.315,
1200
+ "step": 79000
1201
+ },
1202
+ {
1203
+ "epoch": 3.26,
1204
+ "learning_rate": 2.221166666666667e-06,
1205
+ "loss": 1.0172,
1206
+ "step": 80000
1207
+ },
1208
+ {
1209
+ "epoch": 3.26,
1210
+ "eval_accuracy": 0.49540918163672654,
1211
+ "eval_loss": 0.9991635680198669,
1212
+ "eval_runtime": 12.7122,
1213
+ "eval_samples_per_second": 394.109,
1214
+ "eval_steps_per_second": 49.323,
1215
+ "step": 80000
1216
+ },
1217
+ {
1218
+ "epoch": 3.3,
1219
+ "learning_rate": 2.248916666666667e-06,
1220
+ "loss": 1.0185,
1221
+ "step": 81000
1222
+ },
1223
+ {
1224
+ "epoch": 3.3,
1225
+ "eval_accuracy": 0.49001996007984033,
1226
+ "eval_loss": 1.0026898384094238,
1227
+ "eval_runtime": 12.6101,
1228
+ "eval_samples_per_second": 397.299,
1229
+ "eval_steps_per_second": 49.722,
1230
+ "step": 81000
1231
+ },
1232
+ {
1233
+ "epoch": 3.34,
1234
+ "learning_rate": 2.2766944444444444e-06,
1235
+ "loss": 1.0287,
1236
+ "step": 82000
1237
+ },
1238
+ {
1239
+ "epoch": 3.34,
1240
+ "eval_accuracy": 0.4998003992015968,
1241
+ "eval_loss": 0.9921432733535767,
1242
+ "eval_runtime": 12.4639,
1243
+ "eval_samples_per_second": 401.961,
1244
+ "eval_steps_per_second": 50.305,
1245
+ "step": 82000
1246
+ },
1247
+ {
1248
+ "epoch": 3.38,
1249
+ "learning_rate": 2.3044444444444443e-06,
1250
+ "loss": 1.0142,
1251
+ "step": 83000
1252
+ },
1253
+ {
1254
+ "epoch": 3.38,
1255
+ "eval_accuracy": 0.5017964071856288,
1256
+ "eval_loss": 0.9938461184501648,
1257
+ "eval_runtime": 12.4125,
1258
+ "eval_samples_per_second": 403.625,
1259
+ "eval_steps_per_second": 50.514,
1260
+ "step": 83000
1261
+ },
1262
+ {
1263
+ "epoch": 3.42,
1264
+ "learning_rate": 2.3322222222222223e-06,
1265
+ "loss": 1.0186,
1266
+ "step": 84000
1267
+ },
1268
+ {
1269
+ "epoch": 3.42,
1270
+ "eval_accuracy": 0.5089820359281437,
1271
+ "eval_loss": 0.9863815307617188,
1272
+ "eval_runtime": 12.6556,
1273
+ "eval_samples_per_second": 395.872,
1274
+ "eval_steps_per_second": 49.543,
1275
+ "step": 84000
1276
+ },
1277
+ {
1278
+ "epoch": 3.46,
1279
+ "learning_rate": 2.3599722222222226e-06,
1280
+ "loss": 1.0151,
1281
+ "step": 85000
1282
+ },
1283
+ {
1284
+ "epoch": 3.46,
1285
+ "eval_accuracy": 0.5041916167664671,
1286
+ "eval_loss": 0.9891662001609802,
1287
+ "eval_runtime": 12.6529,
1288
+ "eval_samples_per_second": 395.958,
1289
+ "eval_steps_per_second": 49.554,
1290
+ "step": 85000
1291
+ },
1292
+ {
1293
+ "epoch": 3.5,
1294
+ "learning_rate": 2.38775e-06,
1295
+ "loss": 1.0268,
1296
+ "step": 86000
1297
+ },
1298
+ {
1299
+ "epoch": 3.5,
1300
+ "eval_accuracy": 0.5043912175648703,
1301
+ "eval_loss": 0.9857804179191589,
1302
+ "eval_runtime": 12.4197,
1303
+ "eval_samples_per_second": 403.391,
1304
+ "eval_steps_per_second": 50.484,
1305
+ "step": 86000
1306
+ },
1307
+ {
1308
+ "epoch": 3.54,
1309
+ "learning_rate": 2.4155e-06,
1310
+ "loss": 1.0152,
1311
+ "step": 87000
1312
+ },
1313
+ {
1314
+ "epoch": 3.54,
1315
+ "eval_accuracy": 0.5075848303393213,
1316
+ "eval_loss": 0.9841725826263428,
1317
+ "eval_runtime": 12.4152,
1318
+ "eval_samples_per_second": 403.538,
1319
+ "eval_steps_per_second": 50.503,
1320
+ "step": 87000
1321
+ },
1322
+ {
1323
+ "epoch": 3.59,
1324
+ "learning_rate": 2.443277777777778e-06,
1325
+ "loss": 1.0201,
1326
+ "step": 88000
1327
+ },
1328
+ {
1329
+ "epoch": 3.59,
1330
+ "eval_accuracy": 0.5065868263473053,
1331
+ "eval_loss": 0.9841102957725525,
1332
+ "eval_runtime": 12.4716,
1333
+ "eval_samples_per_second": 401.711,
1334
+ "eval_steps_per_second": 50.274,
1335
+ "step": 88000
1336
+ },
1337
+ {
1338
+ "epoch": 3.63,
1339
+ "learning_rate": 2.471027777777778e-06,
1340
+ "loss": 1.0113,
1341
+ "step": 89000
1342
+ },
1343
+ {
1344
+ "epoch": 3.63,
1345
+ "eval_accuracy": 0.5079840319361277,
1346
+ "eval_loss": 0.9851917028427124,
1347
+ "eval_runtime": 12.6634,
1348
+ "eval_samples_per_second": 395.627,
1349
+ "eval_steps_per_second": 49.513,
1350
+ "step": 89000
1351
+ },
1352
+ {
1353
+ "epoch": 3.67,
1354
+ "learning_rate": 2.498805555555556e-06,
1355
+ "loss": 1.0195,
1356
+ "step": 90000
1357
+ },
1358
+ {
1359
+ "epoch": 3.67,
1360
+ "eval_accuracy": 0.5039920159680639,
1361
+ "eval_loss": 0.9855921864509583,
1362
+ "eval_runtime": 12.6226,
1363
+ "eval_samples_per_second": 396.908,
1364
+ "eval_steps_per_second": 49.673,
1365
+ "step": 90000
1366
+ },
1367
+ {
1368
+ "epoch": 3.71,
1369
+ "learning_rate": 2.5265555555555557e-06,
1370
+ "loss": 1.0202,
1371
+ "step": 91000
1372
+ },
1373
+ {
1374
+ "epoch": 3.71,
1375
+ "eval_accuracy": 0.5001996007984032,
1376
+ "eval_loss": 0.9933446645736694,
1377
+ "eval_runtime": 12.4993,
1378
+ "eval_samples_per_second": 400.821,
1379
+ "eval_steps_per_second": 50.163,
1380
+ "step": 91000
1381
+ },
1382
+ {
1383
+ "epoch": 3.75,
1384
+ "learning_rate": 2.5543333333333337e-06,
1385
+ "loss": 1.0183,
1386
+ "step": 92000
1387
+ },
1388
+ {
1389
+ "epoch": 3.75,
1390
+ "eval_accuracy": 0.5,
1391
+ "eval_loss": 0.9928790926933289,
1392
+ "eval_runtime": 12.4539,
1393
+ "eval_samples_per_second": 402.285,
1394
+ "eval_steps_per_second": 50.346,
1395
+ "step": 92000
1396
+ },
1397
+ {
1398
+ "epoch": 3.79,
1399
+ "learning_rate": 2.5820833333333335e-06,
1400
+ "loss": 1.02,
1401
+ "step": 93000
1402
+ },
1403
+ {
1404
+ "epoch": 3.79,
1405
+ "eval_accuracy": 0.5011976047904192,
1406
+ "eval_loss": 0.9867545962333679,
1407
+ "eval_runtime": 12.4054,
1408
+ "eval_samples_per_second": 403.855,
1409
+ "eval_steps_per_second": 50.542,
1410
+ "step": 93000
1411
+ },
1412
+ {
1413
+ "epoch": 3.83,
1414
+ "learning_rate": 2.609861111111111e-06,
1415
+ "loss": 1.0144,
1416
+ "step": 94000
1417
+ },
1418
+ {
1419
+ "epoch": 3.83,
1420
+ "eval_accuracy": 0.5027944111776447,
1421
+ "eval_loss": 0.9887515902519226,
1422
+ "eval_runtime": 12.8101,
1423
+ "eval_samples_per_second": 391.097,
1424
+ "eval_steps_per_second": 48.946,
1425
+ "step": 94000
1426
+ },
1427
+ {
1428
+ "epoch": 3.87,
1429
+ "learning_rate": 2.6376111111111114e-06,
1430
+ "loss": 1.0166,
1431
+ "step": 95000
1432
+ },
1433
+ {
1434
+ "epoch": 3.87,
1435
+ "eval_accuracy": 0.5077844311377245,
1436
+ "eval_loss": 0.9852330684661865,
1437
+ "eval_runtime": 12.6394,
1438
+ "eval_samples_per_second": 396.379,
1439
+ "eval_steps_per_second": 49.607,
1440
+ "step": 95000
1441
+ },
1442
+ {
1443
+ "epoch": 3.91,
1444
+ "learning_rate": 2.6653888888888894e-06,
1445
+ "loss": 1.0173,
1446
+ "step": 96000
1447
+ },
1448
+ {
1449
+ "epoch": 3.91,
1450
+ "eval_accuracy": 0.5053892215568863,
1451
+ "eval_loss": 0.9848083257675171,
1452
+ "eval_runtime": 12.5421,
1453
+ "eval_samples_per_second": 399.456,
1454
+ "eval_steps_per_second": 49.992,
1455
+ "step": 96000
1456
+ },
1457
+ {
1458
+ "epoch": 3.95,
1459
+ "learning_rate": 2.6931388888888892e-06,
1460
+ "loss": 1.0159,
1461
+ "step": 97000
1462
+ },
1463
+ {
1464
+ "epoch": 3.95,
1465
+ "eval_accuracy": 0.5051896207584831,
1466
+ "eval_loss": 0.9930784702301025,
1467
+ "eval_runtime": 12.6498,
1468
+ "eval_samples_per_second": 396.054,
1469
+ "eval_steps_per_second": 49.566,
1470
+ "step": 97000
1471
+ },
1472
+ {
1473
+ "epoch": 3.99,
1474
+ "learning_rate": 2.7209166666666668e-06,
1475
+ "loss": 1.0126,
1476
+ "step": 98000
1477
+ },
1478
+ {
1479
+ "epoch": 3.99,
1480
+ "eval_accuracy": 0.5037924151696607,
1481
+ "eval_loss": 0.9865526556968689,
1482
+ "eval_runtime": 12.4969,
1483
+ "eval_samples_per_second": 400.899,
1484
+ "eval_steps_per_second": 50.172,
1485
+ "step": 98000
1486
+ },
1487
+ {
1488
+ "epoch": 4.03,
1489
+ "learning_rate": 2.7486666666666666e-06,
1490
+ "loss": 1.0129,
1491
+ "step": 99000
1492
+ },
1493
+ {
1494
+ "epoch": 4.03,
1495
+ "eval_accuracy": 0.5073852295409181,
1496
+ "eval_loss": 0.9890230298042297,
1497
+ "eval_runtime": 12.7759,
1498
+ "eval_samples_per_second": 392.143,
1499
+ "eval_steps_per_second": 49.077,
1500
+ "step": 99000
1501
+ },
1502
+ {
1503
+ "epoch": 4.07,
1504
+ "learning_rate": 2.7764444444444446e-06,
1505
+ "loss": 1.0145,
1506
+ "step": 100000
1507
+ },
1508
+ {
1509
+ "epoch": 4.07,
1510
+ "eval_accuracy": 0.506187624750499,
1511
+ "eval_loss": 0.9834555387496948,
1512
+ "eval_runtime": 12.6734,
1513
+ "eval_samples_per_second": 395.316,
1514
+ "eval_steps_per_second": 49.474,
1515
+ "step": 100000
1516
+ },
1517
+ {
1518
+ "epoch": 4.12,
1519
+ "learning_rate": 2.804194444444445e-06,
1520
+ "loss": 1.0164,
1521
+ "step": 101000
1522
+ },
1523
+ {
1524
+ "epoch": 4.12,
1525
+ "eval_accuracy": 0.5081836327345309,
1526
+ "eval_loss": 0.9881391525268555,
1527
+ "eval_runtime": 12.4161,
1528
+ "eval_samples_per_second": 403.509,
1529
+ "eval_steps_per_second": 50.499,
1530
+ "step": 101000
1531
+ },
1532
+ {
1533
+ "epoch": 4.16,
1534
+ "learning_rate": 2.8319722222222225e-06,
1535
+ "loss": 1.0131,
1536
+ "step": 102000
1537
+ },
1538
+ {
1539
+ "epoch": 4.16,
1540
+ "eval_accuracy": 0.5007984031936128,
1541
+ "eval_loss": 0.9866752624511719,
1542
+ "eval_runtime": 12.6211,
1543
+ "eval_samples_per_second": 396.953,
1544
+ "eval_steps_per_second": 49.679,
1545
+ "step": 102000
1546
+ }
1547
+ ],
1548
+ "logging_steps": 1000,
1549
+ "max_steps": 10000000,
1550
+ "num_train_epochs": 408,
1551
+ "save_steps": 1000,
1552
+ "total_flos": 3.5540675960242176e+16,
1553
+ "trial_name": null,
1554
+ "trial_params": null
1555
+ }
checkpoint-102000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f48fd322744a41332d86e79cb68054df0053a36e9279c46ff138f13ef1aab365
3
+ size 4027
checkpoint-102000/vocab.json ADDED
The diff for this file is too large to render. See raw diff