tmnam20 commited on
Commit
c596db8
1 Parent(s): 84ed4cb

Training in progress, step 100000, checkpoint

Browse files
checkpoint-100000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-100000/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": 6,
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-100000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-100000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d4397b92407969802760412ba86ad8c6bb201a5f489443b539f00e28edb9b48
3
+ size 592141381
checkpoint-100000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b842dd0ef9f49aad98bb46dab758c20897f779f38226ee69dcb9fa594970e14f
3
+ size 296072722
checkpoint-100000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:178905242413bb8c8975bad1110aeb5c7269572f1ac3244c7db9d31512ef2056
3
+ size 14575
checkpoint-100000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac41a2d602b48101322cb13f159ff0689d7030cacb2db2df9532911a4002f4ce
3
+ size 627
checkpoint-100000/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-100000/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-100000/trainer_state.json ADDED
@@ -0,0 +1,1525 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9735556244850159,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-100000",
4
+ "epoch": 4.074315514993481,
5
+ "eval_steps": 1000,
6
+ "global_step": 100000,
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": 1.2358,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1519,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.3201596806387226,
26
+ "eval_loss": 1.1120171546936035,
27
+ "eval_runtime": 29.027,
28
+ "eval_samples_per_second": 172.598,
29
+ "eval_steps_per_second": 21.601,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.128,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.31796407185628744,
41
+ "eval_loss": 1.108055830001831,
42
+ "eval_runtime": 28.9944,
43
+ "eval_samples_per_second": 172.792,
44
+ "eval_steps_per_second": 21.625,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.122,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3345309381237525,
56
+ "eval_loss": 1.1074285507202148,
57
+ "eval_runtime": 28.9901,
58
+ "eval_samples_per_second": 172.818,
59
+ "eval_steps_per_second": 21.628,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1196,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.3337325349301397,
71
+ "eval_loss": 1.1045424938201904,
72
+ "eval_runtime": 28.9791,
73
+ "eval_samples_per_second": 172.883,
74
+ "eval_steps_per_second": 21.636,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3880555555555558e-07,
80
+ "loss": 1.1169,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.3397205588822355,
86
+ "eval_loss": 1.1033179759979248,
87
+ "eval_runtime": 29.0295,
88
+ "eval_samples_per_second": 172.583,
89
+ "eval_steps_per_second": 21.599,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.118,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3401197604790419,
101
+ "eval_loss": 1.1030455827713013,
102
+ "eval_runtime": 28.9748,
103
+ "eval_samples_per_second": 172.909,
104
+ "eval_steps_per_second": 21.639,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9433333333333334e-07,
110
+ "loss": 1.1135,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.35568862275449104,
116
+ "eval_loss": 1.0998502969741821,
117
+ "eval_runtime": 28.9964,
118
+ "eval_samples_per_second": 172.78,
119
+ "eval_steps_per_second": 21.623,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2211111111111114e-07,
125
+ "loss": 1.1136,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.34510978043912177,
131
+ "eval_loss": 1.098555564880371,
132
+ "eval_runtime": 28.983,
133
+ "eval_samples_per_second": 172.86,
134
+ "eval_steps_per_second": 21.633,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4986111111111113e-07,
140
+ "loss": 1.11,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.3411177644710579,
146
+ "eval_loss": 1.1035741567611694,
147
+ "eval_runtime": 28.9758,
148
+ "eval_samples_per_second": 172.903,
149
+ "eval_steps_per_second": 21.639,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776388888888889e-07,
155
+ "loss": 1.1076,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.35788423153692617,
161
+ "eval_loss": 1.0970662832260132,
162
+ "eval_runtime": 29.0262,
163
+ "eval_samples_per_second": 172.603,
164
+ "eval_steps_per_second": 21.601,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.0541666666666667e-07,
170
+ "loss": 1.1067,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.37425149700598803,
176
+ "eval_loss": 1.094608187675476,
177
+ "eval_runtime": 28.8738,
178
+ "eval_samples_per_second": 173.513,
179
+ "eval_steps_per_second": 21.715,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3319444444444444e-07,
185
+ "loss": 1.1032,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.37924151696606784,
191
+ "eval_loss": 1.092067837715149,
192
+ "eval_runtime": 28.9514,
193
+ "eval_samples_per_second": 173.049,
194
+ "eval_steps_per_second": 21.657,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.6094444444444446e-07,
200
+ "loss": 1.1052,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3435129740518962,
206
+ "eval_loss": 1.0962508916854858,
207
+ "eval_runtime": 29.2203,
208
+ "eval_samples_per_second": 171.456,
209
+ "eval_steps_per_second": 21.458,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8872222222222223e-07,
215
+ "loss": 1.1015,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.36806387225548903,
221
+ "eval_loss": 1.0941067934036255,
222
+ "eval_runtime": 29.2088,
223
+ "eval_samples_per_second": 171.524,
224
+ "eval_steps_per_second": 21.466,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1650000000000006e-07,
230
+ "loss": 1.0986,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.3718562874251497,
236
+ "eval_loss": 1.0906143188476562,
237
+ "eval_runtime": 28.9463,
238
+ "eval_samples_per_second": 173.079,
239
+ "eval_steps_per_second": 21.661,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4425e-07,
245
+ "loss": 1.0984,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.38183632734530937,
251
+ "eval_loss": 1.0862348079681396,
252
+ "eval_runtime": 28.9909,
253
+ "eval_samples_per_second": 172.813,
254
+ "eval_steps_per_second": 21.627,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7202777777777785e-07,
260
+ "loss": 1.0959,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.40079840319361276,
266
+ "eval_loss": 1.0846956968307495,
267
+ "eval_runtime": 28.9606,
268
+ "eval_samples_per_second": 172.994,
269
+ "eval_steps_per_second": 21.65,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.998055555555556e-07,
275
+ "loss": 1.0965,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4043912175648703,
281
+ "eval_loss": 1.086153507232666,
282
+ "eval_runtime": 28.8411,
283
+ "eval_samples_per_second": 173.71,
284
+ "eval_steps_per_second": 21.74,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275555555555556e-07,
290
+ "loss": 1.0966,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.4041916167664671,
296
+ "eval_loss": 1.0824861526489258,
297
+ "eval_runtime": 28.9652,
298
+ "eval_samples_per_second": 172.966,
299
+ "eval_steps_per_second": 21.647,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553333333333334e-07,
305
+ "loss": 1.0943,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4103792415169661,
311
+ "eval_loss": 1.0817394256591797,
312
+ "eval_runtime": 28.9548,
313
+ "eval_samples_per_second": 173.028,
314
+ "eval_steps_per_second": 21.654,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830833333333334e-07,
320
+ "loss": 1.0898,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.41357285429141716,
326
+ "eval_loss": 1.0764833688735962,
327
+ "eval_runtime": 28.9932,
328
+ "eval_samples_per_second": 172.799,
329
+ "eval_steps_per_second": 21.626,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108611111111111e-07,
335
+ "loss": 1.0877,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.4293413173652695,
341
+ "eval_loss": 1.0701513290405273,
342
+ "eval_runtime": 28.9628,
343
+ "eval_samples_per_second": 172.98,
344
+ "eval_steps_per_second": 21.648,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386388888888889e-07,
350
+ "loss": 1.0826,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.4285429141716567,
356
+ "eval_loss": 1.0636069774627686,
357
+ "eval_runtime": 28.9704,
358
+ "eval_samples_per_second": 172.935,
359
+ "eval_steps_per_second": 21.643,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.664166666666667e-07,
365
+ "loss": 1.0781,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.43213572854291415,
371
+ "eval_loss": 1.059762954711914,
372
+ "eval_runtime": 29.0265,
373
+ "eval_samples_per_second": 172.601,
374
+ "eval_steps_per_second": 21.601,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941666666666667e-07,
380
+ "loss": 1.0728,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4373253493013972,
386
+ "eval_loss": 1.056991696357727,
387
+ "eval_runtime": 28.9761,
388
+ "eval_samples_per_second": 172.901,
389
+ "eval_steps_per_second": 21.639,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.219444444444444e-07,
395
+ "loss": 1.0719,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.43193612774451096,
401
+ "eval_loss": 1.0573900938034058,
402
+ "eval_runtime": 28.9853,
403
+ "eval_samples_per_second": 172.846,
404
+ "eval_steps_per_second": 21.632,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496944444444444e-07,
410
+ "loss": 1.0683,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4305389221556886,
416
+ "eval_loss": 1.0567399263381958,
417
+ "eval_runtime": 28.9611,
418
+ "eval_samples_per_second": 172.991,
419
+ "eval_steps_per_second": 21.65,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774722222222223e-07,
425
+ "loss": 1.0669,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.4297405189620758,
431
+ "eval_loss": 1.05367910861969,
432
+ "eval_runtime": 28.9913,
433
+ "eval_samples_per_second": 172.81,
434
+ "eval_steps_per_second": 21.627,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.052222222222223e-07,
440
+ "loss": 1.0629,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.4303393213572854,
446
+ "eval_loss": 1.0497413873672485,
447
+ "eval_runtime": 28.9605,
448
+ "eval_samples_per_second": 172.994,
449
+ "eval_steps_per_second": 21.65,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.330000000000001e-07,
455
+ "loss": 1.0614,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.43173652694610776,
461
+ "eval_loss": 1.046238899230957,
462
+ "eval_runtime": 28.9896,
463
+ "eval_samples_per_second": 172.821,
464
+ "eval_steps_per_second": 21.628,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.607500000000001e-07,
470
+ "loss": 1.0669,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.43233532934131735,
476
+ "eval_loss": 1.0541218519210815,
477
+ "eval_runtime": 28.9432,
478
+ "eval_samples_per_second": 173.098,
479
+ "eval_steps_per_second": 21.663,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.885277777777779e-07,
485
+ "loss": 1.0612,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.4285429141716567,
491
+ "eval_loss": 1.0527743101119995,
492
+ "eval_runtime": 29.0895,
493
+ "eval_samples_per_second": 172.227,
494
+ "eval_steps_per_second": 21.554,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.163055555555556e-07,
500
+ "loss": 1.0622,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.4339321357285429,
506
+ "eval_loss": 1.0462977886199951,
507
+ "eval_runtime": 29.1293,
508
+ "eval_samples_per_second": 171.992,
509
+ "eval_steps_per_second": 21.525,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440555555555557e-07,
515
+ "loss": 1.0542,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.43313373253493015,
521
+ "eval_loss": 1.044585943222046,
522
+ "eval_runtime": 28.9734,
523
+ "eval_samples_per_second": 172.917,
524
+ "eval_steps_per_second": 21.641,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.718333333333334e-07,
530
+ "loss": 1.0588,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.4281437125748503,
536
+ "eval_loss": 1.0411622524261475,
537
+ "eval_runtime": 28.9957,
538
+ "eval_samples_per_second": 172.784,
539
+ "eval_steps_per_second": 21.624,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995833333333334e-07,
545
+ "loss": 1.0556,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.44471057884231535,
551
+ "eval_loss": 1.037837266921997,
552
+ "eval_runtime": 29.0336,
553
+ "eval_samples_per_second": 172.559,
554
+ "eval_steps_per_second": 21.596,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.0273611111111112e-06,
560
+ "loss": 1.0548,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.4502994011976048,
566
+ "eval_loss": 1.035124659538269,
567
+ "eval_runtime": 29.0914,
568
+ "eval_samples_per_second": 172.216,
569
+ "eval_steps_per_second": 21.553,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.055138888888889e-06,
575
+ "loss": 1.0566,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.4497005988023952,
581
+ "eval_loss": 1.032871127128601,
582
+ "eval_runtime": 29.0304,
583
+ "eval_samples_per_second": 172.578,
584
+ "eval_steps_per_second": 21.598,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0829166666666667e-06,
590
+ "loss": 1.0513,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.4477045908183633,
596
+ "eval_loss": 1.0365269184112549,
597
+ "eval_runtime": 29.0199,
598
+ "eval_samples_per_second": 172.64,
599
+ "eval_steps_per_second": 21.606,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1106666666666668e-06,
605
+ "loss": 1.0453,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.45708582834331335,
611
+ "eval_loss": 1.0288387537002563,
612
+ "eval_runtime": 29.1526,
613
+ "eval_samples_per_second": 171.854,
614
+ "eval_steps_per_second": 21.508,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.1384166666666669e-06,
620
+ "loss": 1.048,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.44530938123752495,
626
+ "eval_loss": 1.0338605642318726,
627
+ "eval_runtime": 29.1871,
628
+ "eval_samples_per_second": 171.651,
629
+ "eval_steps_per_second": 21.482,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.1661944444444447e-06,
635
+ "loss": 1.042,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.45109780439121755,
641
+ "eval_loss": 1.0249321460723877,
642
+ "eval_runtime": 29.0275,
643
+ "eval_samples_per_second": 172.595,
644
+ "eval_steps_per_second": 21.6,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.1939444444444445e-06,
650
+ "loss": 1.0437,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.44510978043912175,
656
+ "eval_loss": 1.0416841506958008,
657
+ "eval_runtime": 29.0578,
658
+ "eval_samples_per_second": 172.415,
659
+ "eval_steps_per_second": 21.578,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.2217222222222223e-06,
665
+ "loss": 1.0454,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4630738522954092,
671
+ "eval_loss": 1.020729422569275,
672
+ "eval_runtime": 29.0874,
673
+ "eval_samples_per_second": 172.24,
674
+ "eval_steps_per_second": 21.556,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2495e-06,
680
+ "loss": 1.0439,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.4746506986027944,
686
+ "eval_loss": 1.0175539255142212,
687
+ "eval_runtime": 29.093,
688
+ "eval_samples_per_second": 172.206,
689
+ "eval_steps_per_second": 21.552,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2772500000000001e-06,
695
+ "loss": 1.0412,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.4626746506986028,
701
+ "eval_loss": 1.0218861103057861,
702
+ "eval_runtime": 28.9933,
703
+ "eval_samples_per_second": 172.798,
704
+ "eval_steps_per_second": 21.626,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3050277777777777e-06,
710
+ "loss": 1.0362,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.47944111776447107,
716
+ "eval_loss": 1.0074430704116821,
717
+ "eval_runtime": 29.1004,
718
+ "eval_samples_per_second": 172.163,
719
+ "eval_steps_per_second": 21.546,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.3328055555555555e-06,
725
+ "loss": 1.0391,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.4614770459081836,
731
+ "eval_loss": 1.0313252210617065,
732
+ "eval_runtime": 29.0173,
733
+ "eval_samples_per_second": 172.655,
734
+ "eval_steps_per_second": 21.608,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3605555555555555e-06,
740
+ "loss": 1.0331,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.48303393213572854,
746
+ "eval_loss": 1.0042223930358887,
747
+ "eval_runtime": 29.0485,
748
+ "eval_samples_per_second": 172.47,
749
+ "eval_steps_per_second": 21.585,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.3883333333333333e-06,
755
+ "loss": 1.0328,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.4810379241516966,
761
+ "eval_loss": 1.0079505443572998,
762
+ "eval_runtime": 29.4084,
763
+ "eval_samples_per_second": 170.359,
764
+ "eval_steps_per_second": 21.32,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.4160833333333334e-06,
770
+ "loss": 1.0325,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.48403193612774453,
776
+ "eval_loss": 1.0023722648620605,
777
+ "eval_runtime": 29.4825,
778
+ "eval_samples_per_second": 169.932,
779
+ "eval_steps_per_second": 21.267,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4438611111111112e-06,
785
+ "loss": 1.029,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.4786427145708583,
791
+ "eval_loss": 1.015783667564392,
792
+ "eval_runtime": 29.182,
793
+ "eval_samples_per_second": 171.681,
794
+ "eval_steps_per_second": 21.486,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.471638888888889e-06,
800
+ "loss": 1.0317,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.49001996007984033,
806
+ "eval_loss": 0.9999898672103882,
807
+ "eval_runtime": 29.4591,
808
+ "eval_samples_per_second": 170.067,
809
+ "eval_steps_per_second": 21.284,
810
+ "step": 53000
811
+ },
812
+ {
813
+ "epoch": 2.2,
814
+ "learning_rate": 1.4993888888888888e-06,
815
+ "loss": 1.0319,
816
+ "step": 54000
817
+ },
818
+ {
819
+ "epoch": 2.2,
820
+ "eval_accuracy": 0.48882235528942114,
821
+ "eval_loss": 1.0013622045516968,
822
+ "eval_runtime": 29.3854,
823
+ "eval_samples_per_second": 170.493,
824
+ "eval_steps_per_second": 21.337,
825
+ "step": 54000
826
+ },
827
+ {
828
+ "epoch": 2.24,
829
+ "learning_rate": 1.5271666666666668e-06,
830
+ "loss": 1.0269,
831
+ "step": 55000
832
+ },
833
+ {
834
+ "epoch": 2.24,
835
+ "eval_accuracy": 0.4878243512974052,
836
+ "eval_loss": 0.9981361627578735,
837
+ "eval_runtime": 29.0692,
838
+ "eval_samples_per_second": 172.347,
839
+ "eval_steps_per_second": 21.569,
840
+ "step": 55000
841
+ },
842
+ {
843
+ "epoch": 2.28,
844
+ "learning_rate": 1.5549166666666666e-06,
845
+ "loss": 1.0236,
846
+ "step": 56000
847
+ },
848
+ {
849
+ "epoch": 2.28,
850
+ "eval_accuracy": 0.474251497005988,
851
+ "eval_loss": 1.0223956108093262,
852
+ "eval_runtime": 29.0989,
853
+ "eval_samples_per_second": 172.172,
854
+ "eval_steps_per_second": 21.547,
855
+ "step": 56000
856
+ },
857
+ {
858
+ "epoch": 2.32,
859
+ "learning_rate": 1.5826944444444446e-06,
860
+ "loss": 1.022,
861
+ "step": 57000
862
+ },
863
+ {
864
+ "epoch": 2.32,
865
+ "eval_accuracy": 0.48962075848303394,
866
+ "eval_loss": 0.999485969543457,
867
+ "eval_runtime": 29.0951,
868
+ "eval_samples_per_second": 172.194,
869
+ "eval_steps_per_second": 21.55,
870
+ "step": 57000
871
+ },
872
+ {
873
+ "epoch": 2.36,
874
+ "learning_rate": 1.6104444444444447e-06,
875
+ "loss": 1.0259,
876
+ "step": 58000
877
+ },
878
+ {
879
+ "epoch": 2.36,
880
+ "eval_accuracy": 0.49161676646706587,
881
+ "eval_loss": 1.00543212890625,
882
+ "eval_runtime": 29.0186,
883
+ "eval_samples_per_second": 172.648,
884
+ "eval_steps_per_second": 21.607,
885
+ "step": 58000
886
+ },
887
+ {
888
+ "epoch": 2.4,
889
+ "learning_rate": 1.6382222222222222e-06,
890
+ "loss": 1.0205,
891
+ "step": 59000
892
+ },
893
+ {
894
+ "epoch": 2.4,
895
+ "eval_accuracy": 0.49620758483033933,
896
+ "eval_loss": 0.9923607110977173,
897
+ "eval_runtime": 29.1145,
898
+ "eval_samples_per_second": 172.079,
899
+ "eval_steps_per_second": 21.536,
900
+ "step": 59000
901
+ },
902
+ {
903
+ "epoch": 2.44,
904
+ "learning_rate": 1.6659722222222223e-06,
905
+ "loss": 1.0272,
906
+ "step": 60000
907
+ },
908
+ {
909
+ "epoch": 2.44,
910
+ "eval_accuracy": 0.49620758483033933,
911
+ "eval_loss": 0.9994486570358276,
912
+ "eval_runtime": 29.0247,
913
+ "eval_samples_per_second": 172.612,
914
+ "eval_steps_per_second": 21.602,
915
+ "step": 60000
916
+ },
917
+ {
918
+ "epoch": 2.49,
919
+ "learning_rate": 1.6937500000000003e-06,
920
+ "loss": 1.023,
921
+ "step": 61000
922
+ },
923
+ {
924
+ "epoch": 2.49,
925
+ "eval_accuracy": 0.49500998003992014,
926
+ "eval_loss": 0.9992738366127014,
927
+ "eval_runtime": 29.024,
928
+ "eval_samples_per_second": 172.616,
929
+ "eval_steps_per_second": 21.603,
930
+ "step": 61000
931
+ },
932
+ {
933
+ "epoch": 2.53,
934
+ "learning_rate": 1.7215000000000002e-06,
935
+ "loss": 1.0257,
936
+ "step": 62000
937
+ },
938
+ {
939
+ "epoch": 2.53,
940
+ "eval_accuracy": 0.481437125748503,
941
+ "eval_loss": 1.0081721544265747,
942
+ "eval_runtime": 29.125,
943
+ "eval_samples_per_second": 172.017,
944
+ "eval_steps_per_second": 21.528,
945
+ "step": 62000
946
+ },
947
+ {
948
+ "epoch": 2.57,
949
+ "learning_rate": 1.7492777777777777e-06,
950
+ "loss": 1.0233,
951
+ "step": 63000
952
+ },
953
+ {
954
+ "epoch": 2.57,
955
+ "eval_accuracy": 0.49660678642714573,
956
+ "eval_loss": 0.9943819642066956,
957
+ "eval_runtime": 29.3033,
958
+ "eval_samples_per_second": 170.97,
959
+ "eval_steps_per_second": 21.397,
960
+ "step": 63000
961
+ },
962
+ {
963
+ "epoch": 2.61,
964
+ "learning_rate": 1.7770277777777778e-06,
965
+ "loss": 1.0207,
966
+ "step": 64000
967
+ },
968
+ {
969
+ "epoch": 2.61,
970
+ "eval_accuracy": 0.500998003992016,
971
+ "eval_loss": 0.9942530989646912,
972
+ "eval_runtime": 29.3586,
973
+ "eval_samples_per_second": 170.648,
974
+ "eval_steps_per_second": 21.357,
975
+ "step": 64000
976
+ },
977
+ {
978
+ "epoch": 2.65,
979
+ "learning_rate": 1.8048055555555558e-06,
980
+ "loss": 1.0212,
981
+ "step": 65000
982
+ },
983
+ {
984
+ "epoch": 2.65,
985
+ "eval_accuracy": 0.5,
986
+ "eval_loss": 0.996075451374054,
987
+ "eval_runtime": 29.5902,
988
+ "eval_samples_per_second": 169.313,
989
+ "eval_steps_per_second": 21.189,
990
+ "step": 65000
991
+ },
992
+ {
993
+ "epoch": 2.69,
994
+ "learning_rate": 1.8325833333333333e-06,
995
+ "loss": 1.0259,
996
+ "step": 66000
997
+ },
998
+ {
999
+ "epoch": 2.69,
1000
+ "eval_accuracy": 0.5043912175648703,
1001
+ "eval_loss": 0.991050124168396,
1002
+ "eval_runtime": 29.5206,
1003
+ "eval_samples_per_second": 169.712,
1004
+ "eval_steps_per_second": 21.239,
1005
+ "step": 66000
1006
+ },
1007
+ {
1008
+ "epoch": 2.73,
1009
+ "learning_rate": 1.8603333333333334e-06,
1010
+ "loss": 1.0199,
1011
+ "step": 67000
1012
+ },
1013
+ {
1014
+ "epoch": 2.73,
1015
+ "eval_accuracy": 0.49181636726546907,
1016
+ "eval_loss": 0.9968613386154175,
1017
+ "eval_runtime": 29.1614,
1018
+ "eval_samples_per_second": 171.802,
1019
+ "eval_steps_per_second": 21.501,
1020
+ "step": 67000
1021
+ },
1022
+ {
1023
+ "epoch": 2.77,
1024
+ "learning_rate": 1.8881111111111114e-06,
1025
+ "loss": 1.0215,
1026
+ "step": 68000
1027
+ },
1028
+ {
1029
+ "epoch": 2.77,
1030
+ "eval_accuracy": 0.5023952095808383,
1031
+ "eval_loss": 0.9903049468994141,
1032
+ "eval_runtime": 29.1422,
1033
+ "eval_samples_per_second": 171.916,
1034
+ "eval_steps_per_second": 21.515,
1035
+ "step": 68000
1036
+ },
1037
+ {
1038
+ "epoch": 2.81,
1039
+ "learning_rate": 1.9158611111111115e-06,
1040
+ "loss": 1.0189,
1041
+ "step": 69000
1042
+ },
1043
+ {
1044
+ "epoch": 2.81,
1045
+ "eval_accuracy": 0.5083832335329341,
1046
+ "eval_loss": 0.9923155903816223,
1047
+ "eval_runtime": 29.1108,
1048
+ "eval_samples_per_second": 172.101,
1049
+ "eval_steps_per_second": 21.538,
1050
+ "step": 69000
1051
+ },
1052
+ {
1053
+ "epoch": 2.85,
1054
+ "learning_rate": 1.943638888888889e-06,
1055
+ "loss": 1.02,
1056
+ "step": 70000
1057
+ },
1058
+ {
1059
+ "epoch": 2.85,
1060
+ "eval_accuracy": 0.4992015968063872,
1061
+ "eval_loss": 0.99754798412323,
1062
+ "eval_runtime": 29.0673,
1063
+ "eval_samples_per_second": 172.359,
1064
+ "eval_steps_per_second": 21.571,
1065
+ "step": 70000
1066
+ },
1067
+ {
1068
+ "epoch": 2.89,
1069
+ "learning_rate": 1.971388888888889e-06,
1070
+ "loss": 1.0124,
1071
+ "step": 71000
1072
+ },
1073
+ {
1074
+ "epoch": 2.89,
1075
+ "eval_accuracy": 0.5003992015968064,
1076
+ "eval_loss": 0.9986121654510498,
1077
+ "eval_runtime": 29.1291,
1078
+ "eval_samples_per_second": 171.993,
1079
+ "eval_steps_per_second": 21.525,
1080
+ "step": 71000
1081
+ },
1082
+ {
1083
+ "epoch": 2.93,
1084
+ "learning_rate": 1.999166666666667e-06,
1085
+ "loss": 1.0212,
1086
+ "step": 72000
1087
+ },
1088
+ {
1089
+ "epoch": 2.93,
1090
+ "eval_accuracy": 0.48862275449101794,
1091
+ "eval_loss": 1.0082526206970215,
1092
+ "eval_runtime": 29.139,
1093
+ "eval_samples_per_second": 171.935,
1094
+ "eval_steps_per_second": 21.518,
1095
+ "step": 72000
1096
+ },
1097
+ {
1098
+ "epoch": 2.97,
1099
+ "learning_rate": 2.0269166666666667e-06,
1100
+ "loss": 1.019,
1101
+ "step": 73000
1102
+ },
1103
+ {
1104
+ "epoch": 2.97,
1105
+ "eval_accuracy": 0.5025948103792415,
1106
+ "eval_loss": 0.9981805682182312,
1107
+ "eval_runtime": 29.4021,
1108
+ "eval_samples_per_second": 170.396,
1109
+ "eval_steps_per_second": 21.325,
1110
+ "step": 73000
1111
+ },
1112
+ {
1113
+ "epoch": 3.01,
1114
+ "learning_rate": 2.0546944444444447e-06,
1115
+ "loss": 1.0205,
1116
+ "step": 74000
1117
+ },
1118
+ {
1119
+ "epoch": 3.01,
1120
+ "eval_accuracy": 0.49800399201596807,
1121
+ "eval_loss": 0.9870504140853882,
1122
+ "eval_runtime": 29.1143,
1123
+ "eval_samples_per_second": 172.081,
1124
+ "eval_steps_per_second": 21.536,
1125
+ "step": 74000
1126
+ },
1127
+ {
1128
+ "epoch": 3.06,
1129
+ "learning_rate": 2.0824444444444446e-06,
1130
+ "loss": 1.0085,
1131
+ "step": 75000
1132
+ },
1133
+ {
1134
+ "epoch": 3.06,
1135
+ "eval_accuracy": 0.5017964071856288,
1136
+ "eval_loss": 0.9926252961158752,
1137
+ "eval_runtime": 29.2368,
1138
+ "eval_samples_per_second": 171.359,
1139
+ "eval_steps_per_second": 21.446,
1140
+ "step": 75000
1141
+ },
1142
+ {
1143
+ "epoch": 3.1,
1144
+ "learning_rate": 2.1102222222222226e-06,
1145
+ "loss": 1.0116,
1146
+ "step": 76000
1147
+ },
1148
+ {
1149
+ "epoch": 3.1,
1150
+ "eval_accuracy": 0.5039920159680639,
1151
+ "eval_loss": 0.9843892455101013,
1152
+ "eval_runtime": 29.4161,
1153
+ "eval_samples_per_second": 170.315,
1154
+ "eval_steps_per_second": 21.315,
1155
+ "step": 76000
1156
+ },
1157
+ {
1158
+ "epoch": 3.14,
1159
+ "learning_rate": 2.1379722222222224e-06,
1160
+ "loss": 1.0128,
1161
+ "step": 77000
1162
+ },
1163
+ {
1164
+ "epoch": 3.14,
1165
+ "eval_accuracy": 0.4746506986027944,
1166
+ "eval_loss": 1.0196081399917603,
1167
+ "eval_runtime": 29.2722,
1168
+ "eval_samples_per_second": 171.152,
1169
+ "eval_steps_per_second": 21.42,
1170
+ "step": 77000
1171
+ },
1172
+ {
1173
+ "epoch": 3.18,
1174
+ "learning_rate": 2.16575e-06,
1175
+ "loss": 1.0051,
1176
+ "step": 78000
1177
+ },
1178
+ {
1179
+ "epoch": 3.18,
1180
+ "eval_accuracy": 0.5039920159680639,
1181
+ "eval_loss": 0.9861226677894592,
1182
+ "eval_runtime": 29.0739,
1183
+ "eval_samples_per_second": 172.32,
1184
+ "eval_steps_per_second": 21.566,
1185
+ "step": 78000
1186
+ },
1187
+ {
1188
+ "epoch": 3.22,
1189
+ "learning_rate": 2.1935000000000003e-06,
1190
+ "loss": 1.0106,
1191
+ "step": 79000
1192
+ },
1193
+ {
1194
+ "epoch": 3.22,
1195
+ "eval_accuracy": 0.49241516966067866,
1196
+ "eval_loss": 1.0106168985366821,
1197
+ "eval_runtime": 29.2182,
1198
+ "eval_samples_per_second": 171.469,
1199
+ "eval_steps_per_second": 21.459,
1200
+ "step": 79000
1201
+ },
1202
+ {
1203
+ "epoch": 3.26,
1204
+ "learning_rate": 2.221277777777778e-06,
1205
+ "loss": 1.0086,
1206
+ "step": 80000
1207
+ },
1208
+ {
1209
+ "epoch": 3.26,
1210
+ "eval_accuracy": 0.4942115768463074,
1211
+ "eval_loss": 1.0031169652938843,
1212
+ "eval_runtime": 29.1384,
1213
+ "eval_samples_per_second": 171.938,
1214
+ "eval_steps_per_second": 21.518,
1215
+ "step": 80000
1216
+ },
1217
+ {
1218
+ "epoch": 3.3,
1219
+ "learning_rate": 2.249027777777778e-06,
1220
+ "loss": 1.0097,
1221
+ "step": 81000
1222
+ },
1223
+ {
1224
+ "epoch": 3.3,
1225
+ "eval_accuracy": 0.4932135728542914,
1226
+ "eval_loss": 0.9960971474647522,
1227
+ "eval_runtime": 29.1317,
1228
+ "eval_samples_per_second": 171.977,
1229
+ "eval_steps_per_second": 21.523,
1230
+ "step": 81000
1231
+ },
1232
+ {
1233
+ "epoch": 3.34,
1234
+ "learning_rate": 2.2768055555555557e-06,
1235
+ "loss": 1.0192,
1236
+ "step": 82000
1237
+ },
1238
+ {
1239
+ "epoch": 3.34,
1240
+ "eval_accuracy": 0.49620758483033933,
1241
+ "eval_loss": 0.9951682686805725,
1242
+ "eval_runtime": 29.1358,
1243
+ "eval_samples_per_second": 171.954,
1244
+ "eval_steps_per_second": 21.52,
1245
+ "step": 82000
1246
+ },
1247
+ {
1248
+ "epoch": 3.38,
1249
+ "learning_rate": 2.3045555555555555e-06,
1250
+ "loss": 1.006,
1251
+ "step": 83000
1252
+ },
1253
+ {
1254
+ "epoch": 3.38,
1255
+ "eval_accuracy": 0.49840319361277446,
1256
+ "eval_loss": 0.9937891960144043,
1257
+ "eval_runtime": 29.1682,
1258
+ "eval_samples_per_second": 171.762,
1259
+ "eval_steps_per_second": 21.496,
1260
+ "step": 83000
1261
+ },
1262
+ {
1263
+ "epoch": 3.42,
1264
+ "learning_rate": 2.3323333333333335e-06,
1265
+ "loss": 1.0118,
1266
+ "step": 84000
1267
+ },
1268
+ {
1269
+ "epoch": 3.42,
1270
+ "eval_accuracy": 0.5007984031936128,
1271
+ "eval_loss": 0.9840214252471924,
1272
+ "eval_runtime": 29.1175,
1273
+ "eval_samples_per_second": 172.061,
1274
+ "eval_steps_per_second": 21.533,
1275
+ "step": 84000
1276
+ },
1277
+ {
1278
+ "epoch": 3.46,
1279
+ "learning_rate": 2.360083333333334e-06,
1280
+ "loss": 1.006,
1281
+ "step": 85000
1282
+ },
1283
+ {
1284
+ "epoch": 3.46,
1285
+ "eval_accuracy": 0.49620758483033933,
1286
+ "eval_loss": 0.9880582690238953,
1287
+ "eval_runtime": 29.1304,
1288
+ "eval_samples_per_second": 171.985,
1289
+ "eval_steps_per_second": 21.524,
1290
+ "step": 85000
1291
+ },
1292
+ {
1293
+ "epoch": 3.5,
1294
+ "learning_rate": 2.3878611111111113e-06,
1295
+ "loss": 1.0151,
1296
+ "step": 86000
1297
+ },
1298
+ {
1299
+ "epoch": 3.5,
1300
+ "eval_accuracy": 0.5053892215568863,
1301
+ "eval_loss": 0.9828479886054993,
1302
+ "eval_runtime": 29.1321,
1303
+ "eval_samples_per_second": 171.975,
1304
+ "eval_steps_per_second": 21.523,
1305
+ "step": 86000
1306
+ },
1307
+ {
1308
+ "epoch": 3.54,
1309
+ "learning_rate": 2.4156388888888893e-06,
1310
+ "loss": 1.0028,
1311
+ "step": 87000
1312
+ },
1313
+ {
1314
+ "epoch": 3.54,
1315
+ "eval_accuracy": 0.5073852295409181,
1316
+ "eval_loss": 0.9821121692657471,
1317
+ "eval_runtime": 29.1459,
1318
+ "eval_samples_per_second": 171.894,
1319
+ "eval_steps_per_second": 21.512,
1320
+ "step": 87000
1321
+ },
1322
+ {
1323
+ "epoch": 3.59,
1324
+ "learning_rate": 2.443388888888889e-06,
1325
+ "loss": 1.0092,
1326
+ "step": 88000
1327
+ },
1328
+ {
1329
+ "epoch": 3.59,
1330
+ "eval_accuracy": 0.5105788423153692,
1331
+ "eval_loss": 0.9791818857192993,
1332
+ "eval_runtime": 29.1206,
1333
+ "eval_samples_per_second": 172.043,
1334
+ "eval_steps_per_second": 21.531,
1335
+ "step": 88000
1336
+ },
1337
+ {
1338
+ "epoch": 3.63,
1339
+ "learning_rate": 2.4711666666666668e-06,
1340
+ "loss": 1.0024,
1341
+ "step": 89000
1342
+ },
1343
+ {
1344
+ "epoch": 3.63,
1345
+ "eval_accuracy": 0.4940119760479042,
1346
+ "eval_loss": 0.986344039440155,
1347
+ "eval_runtime": 29.1948,
1348
+ "eval_samples_per_second": 171.606,
1349
+ "eval_steps_per_second": 21.476,
1350
+ "step": 89000
1351
+ },
1352
+ {
1353
+ "epoch": 3.67,
1354
+ "learning_rate": 2.4989166666666666e-06,
1355
+ "loss": 1.0112,
1356
+ "step": 90000
1357
+ },
1358
+ {
1359
+ "epoch": 3.67,
1360
+ "eval_accuracy": 0.5101796407185629,
1361
+ "eval_loss": 0.9789302349090576,
1362
+ "eval_runtime": 29.2661,
1363
+ "eval_samples_per_second": 171.188,
1364
+ "eval_steps_per_second": 21.424,
1365
+ "step": 90000
1366
+ },
1367
+ {
1368
+ "epoch": 3.71,
1369
+ "learning_rate": 2.5266944444444446e-06,
1370
+ "loss": 1.0137,
1371
+ "step": 91000
1372
+ },
1373
+ {
1374
+ "epoch": 3.71,
1375
+ "eval_accuracy": 0.4936127744510978,
1376
+ "eval_loss": 0.9870284795761108,
1377
+ "eval_runtime": 29.4012,
1378
+ "eval_samples_per_second": 170.401,
1379
+ "eval_steps_per_second": 21.326,
1380
+ "step": 91000
1381
+ },
1382
+ {
1383
+ "epoch": 3.75,
1384
+ "learning_rate": 2.554444444444445e-06,
1385
+ "loss": 1.0076,
1386
+ "step": 92000
1387
+ },
1388
+ {
1389
+ "epoch": 3.75,
1390
+ "eval_accuracy": 0.49880239520958086,
1391
+ "eval_loss": 0.9906250238418579,
1392
+ "eval_runtime": 29.2222,
1393
+ "eval_samples_per_second": 171.445,
1394
+ "eval_steps_per_second": 21.456,
1395
+ "step": 92000
1396
+ },
1397
+ {
1398
+ "epoch": 3.79,
1399
+ "learning_rate": 2.5822222222222224e-06,
1400
+ "loss": 1.0093,
1401
+ "step": 93000
1402
+ },
1403
+ {
1404
+ "epoch": 3.79,
1405
+ "eval_accuracy": 0.5121756487025948,
1406
+ "eval_loss": 0.9766275882720947,
1407
+ "eval_runtime": 29.2374,
1408
+ "eval_samples_per_second": 171.356,
1409
+ "eval_steps_per_second": 21.445,
1410
+ "step": 93000
1411
+ },
1412
+ {
1413
+ "epoch": 3.83,
1414
+ "learning_rate": 2.6099722222222223e-06,
1415
+ "loss": 1.0015,
1416
+ "step": 94000
1417
+ },
1418
+ {
1419
+ "epoch": 3.83,
1420
+ "eval_accuracy": 0.4942115768463074,
1421
+ "eval_loss": 0.991917073726654,
1422
+ "eval_runtime": 29.3042,
1423
+ "eval_samples_per_second": 170.965,
1424
+ "eval_steps_per_second": 21.396,
1425
+ "step": 94000
1426
+ },
1427
+ {
1428
+ "epoch": 3.87,
1429
+ "learning_rate": 2.6377500000000003e-06,
1430
+ "loss": 1.0066,
1431
+ "step": 95000
1432
+ },
1433
+ {
1434
+ "epoch": 3.87,
1435
+ "eval_accuracy": 0.5101796407185629,
1436
+ "eval_loss": 0.9762089252471924,
1437
+ "eval_runtime": 29.2686,
1438
+ "eval_samples_per_second": 171.173,
1439
+ "eval_steps_per_second": 21.422,
1440
+ "step": 95000
1441
+ },
1442
+ {
1443
+ "epoch": 3.91,
1444
+ "learning_rate": 2.6655e-06,
1445
+ "loss": 1.0033,
1446
+ "step": 96000
1447
+ },
1448
+ {
1449
+ "epoch": 3.91,
1450
+ "eval_accuracy": 0.5071856287425149,
1451
+ "eval_loss": 0.9739634990692139,
1452
+ "eval_runtime": 29.2989,
1453
+ "eval_samples_per_second": 170.996,
1454
+ "eval_steps_per_second": 21.4,
1455
+ "step": 96000
1456
+ },
1457
+ {
1458
+ "epoch": 3.95,
1459
+ "learning_rate": 2.6932777777777777e-06,
1460
+ "loss": 1.0054,
1461
+ "step": 97000
1462
+ },
1463
+ {
1464
+ "epoch": 3.95,
1465
+ "eval_accuracy": 0.5033932135728543,
1466
+ "eval_loss": 0.9856627583503723,
1467
+ "eval_runtime": 29.2684,
1468
+ "eval_samples_per_second": 171.174,
1469
+ "eval_steps_per_second": 21.422,
1470
+ "step": 97000
1471
+ },
1472
+ {
1473
+ "epoch": 3.99,
1474
+ "learning_rate": 2.721027777777778e-06,
1475
+ "loss": 1.0034,
1476
+ "step": 98000
1477
+ },
1478
+ {
1479
+ "epoch": 3.99,
1480
+ "eval_accuracy": 0.5055888223552895,
1481
+ "eval_loss": 0.9802790880203247,
1482
+ "eval_runtime": 29.2896,
1483
+ "eval_samples_per_second": 171.051,
1484
+ "eval_steps_per_second": 21.407,
1485
+ "step": 98000
1486
+ },
1487
+ {
1488
+ "epoch": 4.03,
1489
+ "learning_rate": 2.748805555555556e-06,
1490
+ "loss": 1.0003,
1491
+ "step": 99000
1492
+ },
1493
+ {
1494
+ "epoch": 4.03,
1495
+ "eval_accuracy": 0.5183632734530939,
1496
+ "eval_loss": 0.9739055037498474,
1497
+ "eval_runtime": 29.3152,
1498
+ "eval_samples_per_second": 170.901,
1499
+ "eval_steps_per_second": 21.388,
1500
+ "step": 99000
1501
+ },
1502
+ {
1503
+ "epoch": 4.07,
1504
+ "learning_rate": 2.776555555555556e-06,
1505
+ "loss": 0.9978,
1506
+ "step": 100000
1507
+ },
1508
+ {
1509
+ "epoch": 4.07,
1510
+ "eval_accuracy": 0.5153692614770459,
1511
+ "eval_loss": 0.9735556244850159,
1512
+ "eval_runtime": 29.6841,
1513
+ "eval_samples_per_second": 168.777,
1514
+ "eval_steps_per_second": 21.122,
1515
+ "step": 100000
1516
+ }
1517
+ ],
1518
+ "logging_steps": 1000,
1519
+ "max_steps": 10000000,
1520
+ "num_train_epochs": 408,
1521
+ "save_steps": 1000,
1522
+ "total_flos": 1.0452006507066163e+17,
1523
+ "trial_name": null,
1524
+ "trial_params": null
1525
+ }
checkpoint-100000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35131e4915b0e8d5c12cfe76344eb3f82740d8fb21233ad6a425c6f5e63a567e
3
+ size 4027
checkpoint-100000/vocab.json ADDED
The diff for this file is too large to render. See raw diff