tmnam20 commited on
Commit
cc90dad
1 Parent(s): 0c414c0

Training in progress, step 101000, checkpoint

Browse files
checkpoint-101000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-101000/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": 4,
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-101000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-101000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:532e691914a2e84f3ccb8628ec6da59731e164491e7bdc243901aebcdabba9a8
3
+ size 478721285
checkpoint-101000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63da3d43447c8241c7ca4ae5e06308f5e06362da1e6bde6f88147daf903e59ff
3
+ size 239361994
checkpoint-101000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:336bac01c841acdddb138777ff1c3166d526d8ad5608c5ccca795a28afe6d242
3
+ size 14575
checkpoint-101000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b34f84ad91c1739e199d0e8a8217e680da6d28d0f1ae145048f3ce603d177938
3
+ size 627
checkpoint-101000/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-101000/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-101000/trainer_state.json ADDED
@@ -0,0 +1,1540 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9806137681007385,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-96000",
4
+ "epoch": 4.115058670143416,
5
+ "eval_steps": 1000,
6
+ "global_step": 101000,
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": 1.2184,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1394,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.3327345309381238,
26
+ "eval_loss": 1.1149410009384155,
27
+ "eval_runtime": 20.6803,
28
+ "eval_samples_per_second": 242.26,
29
+ "eval_steps_per_second": 30.319,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1141,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.3401197604790419,
41
+ "eval_loss": 1.104099988937378,
42
+ "eval_runtime": 20.8477,
43
+ "eval_samples_per_second": 240.314,
44
+ "eval_steps_per_second": 30.075,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.330555555555556e-08,
50
+ "loss": 1.116,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3407185628742515,
56
+ "eval_loss": 1.1040862798690796,
57
+ "eval_runtime": 20.6818,
58
+ "eval_samples_per_second": 242.242,
59
+ "eval_steps_per_second": 30.317,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1108333333333333e-07,
65
+ "loss": 1.1158,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.32894211576846305,
71
+ "eval_loss": 1.1020556688308716,
72
+ "eval_runtime": 20.8541,
73
+ "eval_samples_per_second": 240.241,
74
+ "eval_steps_per_second": 30.066,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3883333333333335e-07,
80
+ "loss": 1.1135,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.34271457085828344,
86
+ "eval_loss": 1.1008552312850952,
87
+ "eval_runtime": 20.8055,
88
+ "eval_samples_per_second": 240.802,
89
+ "eval_steps_per_second": 30.136,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1121,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3395209580838323,
101
+ "eval_loss": 1.1004050970077515,
102
+ "eval_runtime": 20.8985,
103
+ "eval_samples_per_second": 239.731,
104
+ "eval_steps_per_second": 30.002,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9436111111111112e-07,
110
+ "loss": 1.1089,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.35788423153692617,
116
+ "eval_loss": 1.0985721349716187,
117
+ "eval_runtime": 20.84,
118
+ "eval_samples_per_second": 240.403,
119
+ "eval_steps_per_second": 30.086,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2213888888888891e-07,
125
+ "loss": 1.1079,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.3331337325349301,
131
+ "eval_loss": 1.098374843597412,
132
+ "eval_runtime": 20.7886,
133
+ "eval_samples_per_second": 240.998,
134
+ "eval_steps_per_second": 30.161,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4988888888888893e-07,
140
+ "loss": 1.1087,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.34510978043912177,
146
+ "eval_loss": 1.0993521213531494,
147
+ "eval_runtime": 20.782,
148
+ "eval_samples_per_second": 241.074,
149
+ "eval_steps_per_second": 30.17,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776666666666667e-07,
155
+ "loss": 1.109,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.3475049900199601,
161
+ "eval_loss": 1.0967597961425781,
162
+ "eval_runtime": 20.6798,
163
+ "eval_samples_per_second": 242.265,
164
+ "eval_steps_per_second": 30.319,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.054444444444444e-07,
170
+ "loss": 1.1052,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.37544910179640717,
176
+ "eval_loss": 1.0941349267959595,
177
+ "eval_runtime": 20.8641,
178
+ "eval_samples_per_second": 240.126,
179
+ "eval_steps_per_second": 30.052,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3322222222222225e-07,
185
+ "loss": 1.105,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.3834331337325349,
191
+ "eval_loss": 1.0927647352218628,
192
+ "eval_runtime": 20.6541,
193
+ "eval_samples_per_second": 242.567,
194
+ "eval_steps_per_second": 30.357,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.609722222222222e-07,
200
+ "loss": 1.1016,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.3457085828343313,
206
+ "eval_loss": 1.0942081212997437,
207
+ "eval_runtime": 21.0733,
208
+ "eval_samples_per_second": 237.742,
209
+ "eval_steps_per_second": 29.753,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8875e-07,
215
+ "loss": 1.1031,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.37005988023952097,
221
+ "eval_loss": 1.0918152332305908,
222
+ "eval_runtime": 20.9151,
223
+ "eval_samples_per_second": 239.54,
224
+ "eval_steps_per_second": 29.978,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1652777777777786e-07,
230
+ "loss": 1.1026,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.3790419161676647,
236
+ "eval_loss": 1.0895211696624756,
237
+ "eval_runtime": 21.0591,
238
+ "eval_samples_per_second": 237.902,
239
+ "eval_steps_per_second": 29.773,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4427777777777783e-07,
245
+ "loss": 1.0988,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.4101796407185629,
251
+ "eval_loss": 1.0852997303009033,
252
+ "eval_runtime": 20.9509,
253
+ "eval_samples_per_second": 239.131,
254
+ "eval_steps_per_second": 29.927,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.720555555555556e-07,
260
+ "loss": 1.0974,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.43213572854291415,
266
+ "eval_loss": 1.0791982412338257,
267
+ "eval_runtime": 20.7526,
268
+ "eval_samples_per_second": 241.415,
269
+ "eval_steps_per_second": 30.213,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.998055555555556e-07,
275
+ "loss": 1.0932,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4275449101796407,
281
+ "eval_loss": 1.072191596031189,
282
+ "eval_runtime": 21.2435,
283
+ "eval_samples_per_second": 235.837,
284
+ "eval_steps_per_second": 29.515,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275833333333334e-07,
290
+ "loss": 1.0833,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.43233532934131735,
296
+ "eval_loss": 1.06425940990448,
297
+ "eval_runtime": 20.7923,
298
+ "eval_samples_per_second": 240.955,
299
+ "eval_steps_per_second": 30.155,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553333333333334e-07,
305
+ "loss": 1.0787,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4295409181636727,
311
+ "eval_loss": 1.0638529062271118,
312
+ "eval_runtime": 21.0018,
313
+ "eval_samples_per_second": 238.551,
314
+ "eval_steps_per_second": 29.855,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.831111111111111e-07,
320
+ "loss": 1.0779,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.4243512974051896,
326
+ "eval_loss": 1.0603673458099365,
327
+ "eval_runtime": 20.9689,
328
+ "eval_samples_per_second": 238.926,
329
+ "eval_steps_per_second": 29.901,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108888888888888e-07,
335
+ "loss": 1.0751,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.43233532934131735,
341
+ "eval_loss": 1.0603009462356567,
342
+ "eval_runtime": 20.8897,
343
+ "eval_samples_per_second": 239.831,
344
+ "eval_steps_per_second": 30.015,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.386388888888889e-07,
350
+ "loss": 1.0776,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.42734530938123755,
356
+ "eval_loss": 1.0591468811035156,
357
+ "eval_runtime": 20.964,
358
+ "eval_samples_per_second": 238.981,
359
+ "eval_steps_per_second": 29.908,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.664166666666667e-07,
365
+ "loss": 1.0754,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.4245508982035928,
371
+ "eval_loss": 1.0589721202850342,
372
+ "eval_runtime": 20.9053,
373
+ "eval_samples_per_second": 239.652,
374
+ "eval_steps_per_second": 29.992,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941666666666667e-07,
380
+ "loss": 1.0736,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.43213572854291415,
386
+ "eval_loss": 1.0583962202072144,
387
+ "eval_runtime": 21.3265,
388
+ "eval_samples_per_second": 234.919,
389
+ "eval_steps_per_second": 29.4,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.219444444444444e-07,
395
+ "loss": 1.0717,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.4305389221556886,
401
+ "eval_loss": 1.0561293363571167,
402
+ "eval_runtime": 21.3034,
403
+ "eval_samples_per_second": 235.174,
404
+ "eval_steps_per_second": 29.432,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496944444444444e-07,
410
+ "loss": 1.0709,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4281437125748503,
416
+ "eval_loss": 1.0555357933044434,
417
+ "eval_runtime": 21.2178,
418
+ "eval_samples_per_second": 236.123,
419
+ "eval_steps_per_second": 29.551,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774722222222223e-07,
425
+ "loss": 1.0701,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.4217564870259481,
431
+ "eval_loss": 1.054961085319519,
432
+ "eval_runtime": 21.1775,
433
+ "eval_samples_per_second": 236.571,
434
+ "eval_steps_per_second": 29.607,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.052222222222223e-07,
440
+ "loss": 1.0641,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.4291417165668663,
446
+ "eval_loss": 1.0518379211425781,
447
+ "eval_runtime": 21.0932,
448
+ "eval_samples_per_second": 237.517,
449
+ "eval_steps_per_second": 29.725,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.330000000000001e-07,
455
+ "loss": 1.064,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.43173652694610776,
461
+ "eval_loss": 1.0493717193603516,
462
+ "eval_runtime": 21.2843,
463
+ "eval_samples_per_second": 235.385,
464
+ "eval_steps_per_second": 29.458,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.607500000000001e-07,
470
+ "loss": 1.0693,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.4291417165668663,
476
+ "eval_loss": 1.0521764755249023,
477
+ "eval_runtime": 21.2278,
478
+ "eval_samples_per_second": 236.011,
479
+ "eval_steps_per_second": 29.537,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.885277777777779e-07,
485
+ "loss": 1.0649,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.42375249500998,
491
+ "eval_loss": 1.0528494119644165,
492
+ "eval_runtime": 21.1249,
493
+ "eval_samples_per_second": 237.161,
494
+ "eval_steps_per_second": 29.681,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.163055555555556e-07,
500
+ "loss": 1.0619,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.43293413173652695,
506
+ "eval_loss": 1.049193263053894,
507
+ "eval_runtime": 21.241,
508
+ "eval_samples_per_second": 235.865,
509
+ "eval_steps_per_second": 29.518,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440555555555557e-07,
515
+ "loss": 1.0582,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.4357285429141717,
521
+ "eval_loss": 1.04512619972229,
522
+ "eval_runtime": 21.2531,
523
+ "eval_samples_per_second": 235.73,
524
+ "eval_steps_per_second": 29.502,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.718333333333334e-07,
530
+ "loss": 1.0629,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.43253493013972055,
536
+ "eval_loss": 1.043523907661438,
537
+ "eval_runtime": 21.2176,
538
+ "eval_samples_per_second": 236.124,
539
+ "eval_steps_per_second": 29.551,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995833333333334e-07,
545
+ "loss": 1.0588,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.4307385229540918,
551
+ "eval_loss": 1.0413768291473389,
552
+ "eval_runtime": 21.2225,
553
+ "eval_samples_per_second": 236.07,
554
+ "eval_steps_per_second": 29.544,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.0273611111111112e-06,
560
+ "loss": 1.0552,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.4415169660678643,
566
+ "eval_loss": 1.0397700071334839,
567
+ "eval_runtime": 21.0651,
568
+ "eval_samples_per_second": 237.835,
569
+ "eval_steps_per_second": 29.765,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.055138888888889e-06,
575
+ "loss": 1.0567,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.4419161676646707,
581
+ "eval_loss": 1.0390877723693848,
582
+ "eval_runtime": 21.2858,
583
+ "eval_samples_per_second": 235.368,
584
+ "eval_steps_per_second": 29.456,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0829166666666667e-06,
590
+ "loss": 1.054,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.43952095808383235,
596
+ "eval_loss": 1.0404103994369507,
597
+ "eval_runtime": 21.2544,
598
+ "eval_samples_per_second": 235.715,
599
+ "eval_steps_per_second": 29.5,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1106666666666668e-06,
605
+ "loss": 1.049,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.4477045908183633,
611
+ "eval_loss": 1.0360453128814697,
612
+ "eval_runtime": 21.0993,
613
+ "eval_samples_per_second": 237.448,
614
+ "eval_steps_per_second": 29.717,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.1384444444444446e-06,
620
+ "loss": 1.0522,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.443313373253493,
626
+ "eval_loss": 1.0359567403793335,
627
+ "eval_runtime": 21.2828,
628
+ "eval_samples_per_second": 235.401,
629
+ "eval_steps_per_second": 29.46,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.1661944444444447e-06,
635
+ "loss": 1.0459,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.437125748502994,
641
+ "eval_loss": 1.0328505039215088,
642
+ "eval_runtime": 21.2626,
643
+ "eval_samples_per_second": 235.625,
644
+ "eval_steps_per_second": 29.488,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.1939722222222222e-06,
650
+ "loss": 1.0488,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.4379241516966068,
656
+ "eval_loss": 1.0460196733474731,
657
+ "eval_runtime": 21.2023,
658
+ "eval_samples_per_second": 236.295,
659
+ "eval_steps_per_second": 29.572,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.2217222222222223e-06,
665
+ "loss": 1.0504,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.4421157684630739,
671
+ "eval_loss": 1.0345144271850586,
672
+ "eval_runtime": 21.2684,
673
+ "eval_samples_per_second": 235.561,
674
+ "eval_steps_per_second": 29.48,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2494722222222224e-06,
680
+ "loss": 1.0482,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.46467065868263474,
686
+ "eval_loss": 1.0280784368515015,
687
+ "eval_runtime": 21.1069,
688
+ "eval_samples_per_second": 237.363,
689
+ "eval_steps_per_second": 29.706,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2772500000000001e-06,
695
+ "loss": 1.046,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.4554890219560878,
701
+ "eval_loss": 1.031961441040039,
702
+ "eval_runtime": 21.2098,
703
+ "eval_samples_per_second": 236.211,
704
+ "eval_steps_per_second": 29.562,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3050277777777777e-06,
710
+ "loss": 1.0391,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.46706586826347307,
716
+ "eval_loss": 1.023941159248352,
717
+ "eval_runtime": 21.2269,
718
+ "eval_samples_per_second": 236.021,
719
+ "eval_steps_per_second": 29.538,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.3328055555555555e-06,
725
+ "loss": 1.0444,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.4491017964071856,
731
+ "eval_loss": 1.036636471748352,
732
+ "eval_runtime": 21.1179,
733
+ "eval_samples_per_second": 237.239,
734
+ "eval_steps_per_second": 29.69,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3605555555555555e-06,
740
+ "loss": 1.0386,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.47305389221556887,
746
+ "eval_loss": 1.0199605226516724,
747
+ "eval_runtime": 21.2669,
748
+ "eval_samples_per_second": 235.577,
749
+ "eval_steps_per_second": 29.482,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.3883333333333333e-06,
755
+ "loss": 1.0379,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.4738522954091816,
761
+ "eval_loss": 1.0184080600738525,
762
+ "eval_runtime": 21.2806,
763
+ "eval_samples_per_second": 235.426,
764
+ "eval_steps_per_second": 29.463,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.416111111111111e-06,
770
+ "loss": 1.0365,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.474251497005988,
776
+ "eval_loss": 1.0143232345581055,
777
+ "eval_runtime": 21.1003,
778
+ "eval_samples_per_second": 237.437,
779
+ "eval_steps_per_second": 29.715,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4438888888888889e-06,
785
+ "loss": 1.0356,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.4782435129740519,
791
+ "eval_loss": 1.0179694890975952,
792
+ "eval_runtime": 21.2847,
793
+ "eval_samples_per_second": 235.38,
794
+ "eval_steps_per_second": 29.458,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.471611111111111e-06,
800
+ "loss": 1.0401,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.4864271457085828,
806
+ "eval_loss": 1.008955955505371,
807
+ "eval_runtime": 21.0808,
808
+ "eval_samples_per_second": 237.657,
809
+ "eval_steps_per_second": 29.743,
810
+ "step": 53000
811
+ },
812
+ {
813
+ "epoch": 2.2,
814
+ "learning_rate": 1.4993888888888888e-06,
815
+ "loss": 1.0338,
816
+ "step": 54000
817
+ },
818
+ {
819
+ "epoch": 2.2,
820
+ "eval_accuracy": 0.4870259481037924,
821
+ "eval_loss": 1.0100921392440796,
822
+ "eval_runtime": 21.2581,
823
+ "eval_samples_per_second": 235.674,
824
+ "eval_steps_per_second": 29.495,
825
+ "step": 54000
826
+ },
827
+ {
828
+ "epoch": 2.24,
829
+ "learning_rate": 1.5271388888888889e-06,
830
+ "loss": 1.0322,
831
+ "step": 55000
832
+ },
833
+ {
834
+ "epoch": 2.24,
835
+ "eval_accuracy": 0.48403193612774453,
836
+ "eval_loss": 1.0082924365997314,
837
+ "eval_runtime": 21.2835,
838
+ "eval_samples_per_second": 235.394,
839
+ "eval_steps_per_second": 29.459,
840
+ "step": 55000
841
+ },
842
+ {
843
+ "epoch": 2.28,
844
+ "learning_rate": 1.5549166666666666e-06,
845
+ "loss": 1.0287,
846
+ "step": 56000
847
+ },
848
+ {
849
+ "epoch": 2.28,
850
+ "eval_accuracy": 0.47065868263473054,
851
+ "eval_loss": 1.0229328870773315,
852
+ "eval_runtime": 21.1356,
853
+ "eval_samples_per_second": 237.04,
854
+ "eval_steps_per_second": 29.666,
855
+ "step": 56000
856
+ },
857
+ {
858
+ "epoch": 2.32,
859
+ "learning_rate": 1.5826944444444446e-06,
860
+ "loss": 1.0291,
861
+ "step": 57000
862
+ },
863
+ {
864
+ "epoch": 2.32,
865
+ "eval_accuracy": 0.4844311377245509,
866
+ "eval_loss": 1.0074658393859863,
867
+ "eval_runtime": 21.2919,
868
+ "eval_samples_per_second": 235.301,
869
+ "eval_steps_per_second": 29.448,
870
+ "step": 57000
871
+ },
872
+ {
873
+ "epoch": 2.36,
874
+ "learning_rate": 1.6104722222222222e-06,
875
+ "loss": 1.0304,
876
+ "step": 58000
877
+ },
878
+ {
879
+ "epoch": 2.36,
880
+ "eval_accuracy": 0.4874251497005988,
881
+ "eval_loss": 1.007895588874817,
882
+ "eval_runtime": 21.1997,
883
+ "eval_samples_per_second": 236.324,
884
+ "eval_steps_per_second": 29.576,
885
+ "step": 58000
886
+ },
887
+ {
888
+ "epoch": 2.4,
889
+ "learning_rate": 1.6381944444444445e-06,
890
+ "loss": 1.0259,
891
+ "step": 59000
892
+ },
893
+ {
894
+ "epoch": 2.4,
895
+ "eval_accuracy": 0.492814371257485,
896
+ "eval_loss": 0.9993996620178223,
897
+ "eval_runtime": 21.0668,
898
+ "eval_samples_per_second": 237.815,
899
+ "eval_steps_per_second": 29.763,
900
+ "step": 59000
901
+ },
902
+ {
903
+ "epoch": 2.44,
904
+ "learning_rate": 1.6659722222222223e-06,
905
+ "loss": 1.0314,
906
+ "step": 60000
907
+ },
908
+ {
909
+ "epoch": 2.44,
910
+ "eval_accuracy": 0.48902195608782434,
911
+ "eval_loss": 1.0036838054656982,
912
+ "eval_runtime": 21.2359,
913
+ "eval_samples_per_second": 235.921,
914
+ "eval_steps_per_second": 29.525,
915
+ "step": 60000
916
+ },
917
+ {
918
+ "epoch": 2.49,
919
+ "learning_rate": 1.6937500000000003e-06,
920
+ "loss": 1.0276,
921
+ "step": 61000
922
+ },
923
+ {
924
+ "epoch": 2.49,
925
+ "eval_accuracy": 0.48502994011976047,
926
+ "eval_loss": 1.0065844058990479,
927
+ "eval_runtime": 21.07,
928
+ "eval_samples_per_second": 237.778,
929
+ "eval_steps_per_second": 29.758,
930
+ "step": 61000
931
+ },
932
+ {
933
+ "epoch": 2.53,
934
+ "learning_rate": 1.7215000000000002e-06,
935
+ "loss": 1.0298,
936
+ "step": 62000
937
+ },
938
+ {
939
+ "epoch": 2.53,
940
+ "eval_accuracy": 0.48263473053892214,
941
+ "eval_loss": 1.0085350275039673,
942
+ "eval_runtime": 21.4949,
943
+ "eval_samples_per_second": 233.079,
944
+ "eval_steps_per_second": 29.17,
945
+ "step": 62000
946
+ },
947
+ {
948
+ "epoch": 2.57,
949
+ "learning_rate": 1.7492777777777777e-06,
950
+ "loss": 1.0292,
951
+ "step": 63000
952
+ },
953
+ {
954
+ "epoch": 2.57,
955
+ "eval_accuracy": 0.49560878243512974,
956
+ "eval_loss": 0.9984678030014038,
957
+ "eval_runtime": 21.4688,
958
+ "eval_samples_per_second": 233.362,
959
+ "eval_steps_per_second": 29.205,
960
+ "step": 63000
961
+ },
962
+ {
963
+ "epoch": 2.61,
964
+ "learning_rate": 1.7770277777777778e-06,
965
+ "loss": 1.0273,
966
+ "step": 64000
967
+ },
968
+ {
969
+ "epoch": 2.61,
970
+ "eval_accuracy": 0.4942115768463074,
971
+ "eval_loss": 0.9960917830467224,
972
+ "eval_runtime": 21.307,
973
+ "eval_samples_per_second": 235.134,
974
+ "eval_steps_per_second": 29.427,
975
+ "step": 64000
976
+ },
977
+ {
978
+ "epoch": 2.65,
979
+ "learning_rate": 1.8048055555555558e-06,
980
+ "loss": 1.0256,
981
+ "step": 65000
982
+ },
983
+ {
984
+ "epoch": 2.65,
985
+ "eval_accuracy": 0.49840319361277446,
986
+ "eval_loss": 0.9994720816612244,
987
+ "eval_runtime": 21.5116,
988
+ "eval_samples_per_second": 232.897,
989
+ "eval_steps_per_second": 29.147,
990
+ "step": 65000
991
+ },
992
+ {
993
+ "epoch": 2.69,
994
+ "learning_rate": 1.8325833333333333e-06,
995
+ "loss": 1.0306,
996
+ "step": 66000
997
+ },
998
+ {
999
+ "epoch": 2.69,
1000
+ "eval_accuracy": 0.49660678642714573,
1001
+ "eval_loss": 0.9951530694961548,
1002
+ "eval_runtime": 21.5301,
1003
+ "eval_samples_per_second": 232.697,
1004
+ "eval_steps_per_second": 29.122,
1005
+ "step": 66000
1006
+ },
1007
+ {
1008
+ "epoch": 2.73,
1009
+ "learning_rate": 1.8603333333333334e-06,
1010
+ "loss": 1.0245,
1011
+ "step": 67000
1012
+ },
1013
+ {
1014
+ "epoch": 2.73,
1015
+ "eval_accuracy": 0.499001996007984,
1016
+ "eval_loss": 0.9978940486907959,
1017
+ "eval_runtime": 21.0729,
1018
+ "eval_samples_per_second": 237.746,
1019
+ "eval_steps_per_second": 29.754,
1020
+ "step": 67000
1021
+ },
1022
+ {
1023
+ "epoch": 2.77,
1024
+ "learning_rate": 1.8881111111111114e-06,
1025
+ "loss": 1.0247,
1026
+ "step": 68000
1027
+ },
1028
+ {
1029
+ "epoch": 2.77,
1030
+ "eval_accuracy": 0.5,
1031
+ "eval_loss": 0.9951012134552002,
1032
+ "eval_runtime": 21.1793,
1033
+ "eval_samples_per_second": 236.552,
1034
+ "eval_steps_per_second": 29.604,
1035
+ "step": 68000
1036
+ },
1037
+ {
1038
+ "epoch": 2.81,
1039
+ "learning_rate": 1.9158611111111115e-06,
1040
+ "loss": 1.0241,
1041
+ "step": 69000
1042
+ },
1043
+ {
1044
+ "epoch": 2.81,
1045
+ "eval_accuracy": 0.5021956087824351,
1046
+ "eval_loss": 0.9969767332077026,
1047
+ "eval_runtime": 21.2229,
1048
+ "eval_samples_per_second": 236.066,
1049
+ "eval_steps_per_second": 29.544,
1050
+ "step": 69000
1051
+ },
1052
+ {
1053
+ "epoch": 2.85,
1054
+ "learning_rate": 1.943638888888889e-06,
1055
+ "loss": 1.0242,
1056
+ "step": 70000
1057
+ },
1058
+ {
1059
+ "epoch": 2.85,
1060
+ "eval_accuracy": 0.49720558882235527,
1061
+ "eval_loss": 1.00163733959198,
1062
+ "eval_runtime": 21.0442,
1063
+ "eval_samples_per_second": 238.071,
1064
+ "eval_steps_per_second": 29.794,
1065
+ "step": 70000
1066
+ },
1067
+ {
1068
+ "epoch": 2.89,
1069
+ "learning_rate": 1.971416666666667e-06,
1070
+ "loss": 1.017,
1071
+ "step": 71000
1072
+ },
1073
+ {
1074
+ "epoch": 2.89,
1075
+ "eval_accuracy": 0.49600798403193613,
1076
+ "eval_loss": 1.0031572580337524,
1077
+ "eval_runtime": 21.1987,
1078
+ "eval_samples_per_second": 236.336,
1079
+ "eval_steps_per_second": 29.577,
1080
+ "step": 71000
1081
+ },
1082
+ {
1083
+ "epoch": 2.93,
1084
+ "learning_rate": 1.999166666666667e-06,
1085
+ "loss": 1.0237,
1086
+ "step": 72000
1087
+ },
1088
+ {
1089
+ "epoch": 2.93,
1090
+ "eval_accuracy": 0.49001996007984033,
1091
+ "eval_loss": 1.0067389011383057,
1092
+ "eval_runtime": 21.2825,
1093
+ "eval_samples_per_second": 235.405,
1094
+ "eval_steps_per_second": 29.461,
1095
+ "step": 72000
1096
+ },
1097
+ {
1098
+ "epoch": 2.97,
1099
+ "learning_rate": 2.0269444444444444e-06,
1100
+ "loss": 1.0208,
1101
+ "step": 73000
1102
+ },
1103
+ {
1104
+ "epoch": 2.97,
1105
+ "eval_accuracy": 0.49560878243512974,
1106
+ "eval_loss": 1.0011804103851318,
1107
+ "eval_runtime": 21.137,
1108
+ "eval_samples_per_second": 237.025,
1109
+ "eval_steps_per_second": 29.664,
1110
+ "step": 73000
1111
+ },
1112
+ {
1113
+ "epoch": 3.01,
1114
+ "learning_rate": 2.0546944444444447e-06,
1115
+ "loss": 1.021,
1116
+ "step": 74000
1117
+ },
1118
+ {
1119
+ "epoch": 3.01,
1120
+ "eval_accuracy": 0.49580838323353293,
1121
+ "eval_loss": 0.9941307902336121,
1122
+ "eval_runtime": 21.161,
1123
+ "eval_samples_per_second": 236.757,
1124
+ "eval_steps_per_second": 29.63,
1125
+ "step": 74000
1126
+ },
1127
+ {
1128
+ "epoch": 3.06,
1129
+ "learning_rate": 2.0824722222222223e-06,
1130
+ "loss": 1.0154,
1131
+ "step": 75000
1132
+ },
1133
+ {
1134
+ "epoch": 3.06,
1135
+ "eval_accuracy": 0.49580838323353293,
1136
+ "eval_loss": 0.9971462488174438,
1137
+ "eval_runtime": 21.0263,
1138
+ "eval_samples_per_second": 238.273,
1139
+ "eval_steps_per_second": 29.82,
1140
+ "step": 75000
1141
+ },
1142
+ {
1143
+ "epoch": 3.1,
1144
+ "learning_rate": 2.1102222222222226e-06,
1145
+ "loss": 1.0173,
1146
+ "step": 76000
1147
+ },
1148
+ {
1149
+ "epoch": 3.1,
1150
+ "eval_accuracy": 0.5011976047904192,
1151
+ "eval_loss": 0.9884433150291443,
1152
+ "eval_runtime": 21.2199,
1153
+ "eval_samples_per_second": 236.099,
1154
+ "eval_steps_per_second": 29.548,
1155
+ "step": 76000
1156
+ },
1157
+ {
1158
+ "epoch": 3.14,
1159
+ "learning_rate": 2.138e-06,
1160
+ "loss": 1.0181,
1161
+ "step": 77000
1162
+ },
1163
+ {
1164
+ "epoch": 3.14,
1165
+ "eval_accuracy": 0.4808383233532934,
1166
+ "eval_loss": 1.0162075757980347,
1167
+ "eval_runtime": 21.2108,
1168
+ "eval_samples_per_second": 236.201,
1169
+ "eval_steps_per_second": 29.56,
1170
+ "step": 77000
1171
+ },
1172
+ {
1173
+ "epoch": 3.18,
1174
+ "learning_rate": 2.16575e-06,
1175
+ "loss": 1.0112,
1176
+ "step": 78000
1177
+ },
1178
+ {
1179
+ "epoch": 3.18,
1180
+ "eval_accuracy": 0.49880239520958086,
1181
+ "eval_loss": 0.9902569651603699,
1182
+ "eval_runtime": 21.032,
1183
+ "eval_samples_per_second": 238.209,
1184
+ "eval_steps_per_second": 29.812,
1185
+ "step": 78000
1186
+ },
1187
+ {
1188
+ "epoch": 3.22,
1189
+ "learning_rate": 2.193527777777778e-06,
1190
+ "loss": 1.0166,
1191
+ "step": 79000
1192
+ },
1193
+ {
1194
+ "epoch": 3.22,
1195
+ "eval_accuracy": 0.49500998003992014,
1196
+ "eval_loss": 1.0056451559066772,
1197
+ "eval_runtime": 21.1866,
1198
+ "eval_samples_per_second": 236.47,
1199
+ "eval_steps_per_second": 29.594,
1200
+ "step": 79000
1201
+ },
1202
+ {
1203
+ "epoch": 3.26,
1204
+ "learning_rate": 2.221277777777778e-06,
1205
+ "loss": 1.0135,
1206
+ "step": 80000
1207
+ },
1208
+ {
1209
+ "epoch": 3.26,
1210
+ "eval_accuracy": 0.49161676646706587,
1211
+ "eval_loss": 1.004488229751587,
1212
+ "eval_runtime": 21.2075,
1213
+ "eval_samples_per_second": 236.237,
1214
+ "eval_steps_per_second": 29.565,
1215
+ "step": 80000
1216
+ },
1217
+ {
1218
+ "epoch": 3.3,
1219
+ "learning_rate": 2.249055555555556e-06,
1220
+ "loss": 1.0147,
1221
+ "step": 81000
1222
+ },
1223
+ {
1224
+ "epoch": 3.3,
1225
+ "eval_accuracy": 0.49101796407185627,
1226
+ "eval_loss": 1.0022324323654175,
1227
+ "eval_runtime": 21.1622,
1228
+ "eval_samples_per_second": 236.743,
1229
+ "eval_steps_per_second": 29.628,
1230
+ "step": 81000
1231
+ },
1232
+ {
1233
+ "epoch": 3.34,
1234
+ "learning_rate": 2.2768055555555557e-06,
1235
+ "loss": 1.0249,
1236
+ "step": 82000
1237
+ },
1238
+ {
1239
+ "epoch": 3.34,
1240
+ "eval_accuracy": 0.49540918163672654,
1241
+ "eval_loss": 0.9961836338043213,
1242
+ "eval_runtime": 21.2661,
1243
+ "eval_samples_per_second": 235.586,
1244
+ "eval_steps_per_second": 29.484,
1245
+ "step": 82000
1246
+ },
1247
+ {
1248
+ "epoch": 3.38,
1249
+ "learning_rate": 2.3045833333333336e-06,
1250
+ "loss": 1.012,
1251
+ "step": 83000
1252
+ },
1253
+ {
1254
+ "epoch": 3.38,
1255
+ "eval_accuracy": 0.49520958083832334,
1256
+ "eval_loss": 0.997968316078186,
1257
+ "eval_runtime": 21.1127,
1258
+ "eval_samples_per_second": 237.298,
1259
+ "eval_steps_per_second": 29.698,
1260
+ "step": 83000
1261
+ },
1262
+ {
1263
+ "epoch": 3.42,
1264
+ "learning_rate": 2.3323333333333335e-06,
1265
+ "loss": 1.0153,
1266
+ "step": 84000
1267
+ },
1268
+ {
1269
+ "epoch": 3.42,
1270
+ "eval_accuracy": 0.5063872255489021,
1271
+ "eval_loss": 0.9885319471359253,
1272
+ "eval_runtime": 21.2015,
1273
+ "eval_samples_per_second": 236.304,
1274
+ "eval_steps_per_second": 29.573,
1275
+ "step": 84000
1276
+ },
1277
+ {
1278
+ "epoch": 3.46,
1279
+ "learning_rate": 2.360111111111111e-06,
1280
+ "loss": 1.0113,
1281
+ "step": 85000
1282
+ },
1283
+ {
1284
+ "epoch": 3.46,
1285
+ "eval_accuracy": 0.49860279441117766,
1286
+ "eval_loss": 0.9927662014961243,
1287
+ "eval_runtime": 21.2447,
1288
+ "eval_samples_per_second": 235.823,
1289
+ "eval_steps_per_second": 29.513,
1290
+ "step": 85000
1291
+ },
1292
+ {
1293
+ "epoch": 3.5,
1294
+ "learning_rate": 2.3878611111111113e-06,
1295
+ "loss": 1.0222,
1296
+ "step": 86000
1297
+ },
1298
+ {
1299
+ "epoch": 3.5,
1300
+ "eval_accuracy": 0.5073852295409181,
1301
+ "eval_loss": 0.9855450987815857,
1302
+ "eval_runtime": 21.03,
1303
+ "eval_samples_per_second": 238.231,
1304
+ "eval_steps_per_second": 29.815,
1305
+ "step": 86000
1306
+ },
1307
+ {
1308
+ "epoch": 3.54,
1309
+ "learning_rate": 2.4156388888888893e-06,
1310
+ "loss": 1.0062,
1311
+ "step": 87000
1312
+ },
1313
+ {
1314
+ "epoch": 3.54,
1315
+ "eval_accuracy": 0.5075848303393213,
1316
+ "eval_loss": 0.9856505990028381,
1317
+ "eval_runtime": 21.1796,
1318
+ "eval_samples_per_second": 236.548,
1319
+ "eval_steps_per_second": 29.604,
1320
+ "step": 87000
1321
+ },
1322
+ {
1323
+ "epoch": 3.59,
1324
+ "learning_rate": 2.443388888888889e-06,
1325
+ "loss": 1.0157,
1326
+ "step": 88000
1327
+ },
1328
+ {
1329
+ "epoch": 3.59,
1330
+ "eval_accuracy": 0.5115768463073852,
1331
+ "eval_loss": 0.9844857454299927,
1332
+ "eval_runtime": 21.2024,
1333
+ "eval_samples_per_second": 236.294,
1334
+ "eval_steps_per_second": 29.572,
1335
+ "step": 88000
1336
+ },
1337
+ {
1338
+ "epoch": 3.63,
1339
+ "learning_rate": 2.4711666666666668e-06,
1340
+ "loss": 1.0069,
1341
+ "step": 89000
1342
+ },
1343
+ {
1344
+ "epoch": 3.63,
1345
+ "eval_accuracy": 0.499001996007984,
1346
+ "eval_loss": 0.9894497394561768,
1347
+ "eval_runtime": 21.2532,
1348
+ "eval_samples_per_second": 235.729,
1349
+ "eval_steps_per_second": 29.501,
1350
+ "step": 89000
1351
+ },
1352
+ {
1353
+ "epoch": 3.67,
1354
+ "learning_rate": 2.4989166666666666e-06,
1355
+ "loss": 1.0164,
1356
+ "step": 90000
1357
+ },
1358
+ {
1359
+ "epoch": 3.67,
1360
+ "eval_accuracy": 0.5101796407185629,
1361
+ "eval_loss": 0.9842925667762756,
1362
+ "eval_runtime": 21.1914,
1363
+ "eval_samples_per_second": 236.416,
1364
+ "eval_steps_per_second": 29.587,
1365
+ "step": 90000
1366
+ },
1367
+ {
1368
+ "epoch": 3.71,
1369
+ "learning_rate": 2.5266944444444446e-06,
1370
+ "loss": 1.0175,
1371
+ "step": 91000
1372
+ },
1373
+ {
1374
+ "epoch": 3.71,
1375
+ "eval_accuracy": 0.5055888223552895,
1376
+ "eval_loss": 0.9913986325263977,
1377
+ "eval_runtime": 21.0589,
1378
+ "eval_samples_per_second": 237.904,
1379
+ "eval_steps_per_second": 29.774,
1380
+ "step": 91000
1381
+ },
1382
+ {
1383
+ "epoch": 3.75,
1384
+ "learning_rate": 2.554444444444445e-06,
1385
+ "loss": 1.013,
1386
+ "step": 92000
1387
+ },
1388
+ {
1389
+ "epoch": 3.75,
1390
+ "eval_accuracy": 0.5057884231536927,
1391
+ "eval_loss": 0.9887382388114929,
1392
+ "eval_runtime": 21.2965,
1393
+ "eval_samples_per_second": 235.25,
1394
+ "eval_steps_per_second": 29.441,
1395
+ "step": 92000
1396
+ },
1397
+ {
1398
+ "epoch": 3.79,
1399
+ "learning_rate": 2.5822222222222224e-06,
1400
+ "loss": 1.0157,
1401
+ "step": 93000
1402
+ },
1403
+ {
1404
+ "epoch": 3.79,
1405
+ "eval_accuracy": 0.5127744510978044,
1406
+ "eval_loss": 0.9854277968406677,
1407
+ "eval_runtime": 21.2839,
1408
+ "eval_samples_per_second": 235.389,
1409
+ "eval_steps_per_second": 29.459,
1410
+ "step": 93000
1411
+ },
1412
+ {
1413
+ "epoch": 3.83,
1414
+ "learning_rate": 2.6099722222222223e-06,
1415
+ "loss": 1.0108,
1416
+ "step": 94000
1417
+ },
1418
+ {
1419
+ "epoch": 3.83,
1420
+ "eval_accuracy": 0.5039920159680639,
1421
+ "eval_loss": 0.9907957315444946,
1422
+ "eval_runtime": 21.0477,
1423
+ "eval_samples_per_second": 238.03,
1424
+ "eval_steps_per_second": 29.789,
1425
+ "step": 94000
1426
+ },
1427
+ {
1428
+ "epoch": 3.87,
1429
+ "learning_rate": 2.6377500000000003e-06,
1430
+ "loss": 1.0132,
1431
+ "step": 95000
1432
+ },
1433
+ {
1434
+ "epoch": 3.87,
1435
+ "eval_accuracy": 0.5161676646706587,
1436
+ "eval_loss": 0.9830310940742493,
1437
+ "eval_runtime": 21.2625,
1438
+ "eval_samples_per_second": 235.626,
1439
+ "eval_steps_per_second": 29.489,
1440
+ "step": 95000
1441
+ },
1442
+ {
1443
+ "epoch": 3.91,
1444
+ "learning_rate": 2.6655e-06,
1445
+ "loss": 1.0139,
1446
+ "step": 96000
1447
+ },
1448
+ {
1449
+ "epoch": 3.91,
1450
+ "eval_accuracy": 0.5125748502994012,
1451
+ "eval_loss": 0.9806137681007385,
1452
+ "eval_runtime": 21.1848,
1453
+ "eval_samples_per_second": 236.49,
1454
+ "eval_steps_per_second": 29.597,
1455
+ "step": 96000
1456
+ },
1457
+ {
1458
+ "epoch": 3.95,
1459
+ "learning_rate": 2.6932777777777777e-06,
1460
+ "loss": 1.0104,
1461
+ "step": 97000
1462
+ },
1463
+ {
1464
+ "epoch": 3.95,
1465
+ "eval_accuracy": 0.49640718562874253,
1466
+ "eval_loss": 0.9965940117835999,
1467
+ "eval_runtime": 21.1145,
1468
+ "eval_samples_per_second": 237.278,
1469
+ "eval_steps_per_second": 29.695,
1470
+ "step": 97000
1471
+ },
1472
+ {
1473
+ "epoch": 3.99,
1474
+ "learning_rate": 2.721027777777778e-06,
1475
+ "loss": 1.011,
1476
+ "step": 98000
1477
+ },
1478
+ {
1479
+ "epoch": 3.99,
1480
+ "eval_accuracy": 0.49800399201596807,
1481
+ "eval_loss": 0.9846762418746948,
1482
+ "eval_runtime": 21.2444,
1483
+ "eval_samples_per_second": 235.827,
1484
+ "eval_steps_per_second": 29.514,
1485
+ "step": 98000
1486
+ },
1487
+ {
1488
+ "epoch": 4.03,
1489
+ "learning_rate": 2.748805555555556e-06,
1490
+ "loss": 1.0063,
1491
+ "step": 99000
1492
+ },
1493
+ {
1494
+ "epoch": 4.03,
1495
+ "eval_accuracy": 0.5103792415169661,
1496
+ "eval_loss": 0.9853057265281677,
1497
+ "eval_runtime": 21.0515,
1498
+ "eval_samples_per_second": 237.988,
1499
+ "eval_steps_per_second": 29.784,
1500
+ "step": 99000
1501
+ },
1502
+ {
1503
+ "epoch": 4.07,
1504
+ "learning_rate": 2.776555555555556e-06,
1505
+ "loss": 1.007,
1506
+ "step": 100000
1507
+ },
1508
+ {
1509
+ "epoch": 4.07,
1510
+ "eval_accuracy": 0.5115768463073852,
1511
+ "eval_loss": 0.983065128326416,
1512
+ "eval_runtime": 21.2494,
1513
+ "eval_samples_per_second": 235.771,
1514
+ "eval_steps_per_second": 29.507,
1515
+ "step": 100000
1516
+ },
1517
+ {
1518
+ "epoch": 4.12,
1519
+ "learning_rate": 2.8043333333333334e-06,
1520
+ "loss": 1.0107,
1521
+ "step": 101000
1522
+ },
1523
+ {
1524
+ "epoch": 4.12,
1525
+ "eval_accuracy": 0.5129740518962076,
1526
+ "eval_loss": 0.9863881468772888,
1527
+ "eval_runtime": 21.2658,
1528
+ "eval_samples_per_second": 235.59,
1529
+ "eval_steps_per_second": 29.484,
1530
+ "step": 101000
1531
+ }
1532
+ ],
1533
+ "logging_steps": 1000,
1534
+ "max_steps": 10000000,
1535
+ "num_train_epochs": 408,
1536
+ "save_steps": 1000,
1537
+ "total_flos": 7.03787535999959e+16,
1538
+ "trial_name": null,
1539
+ "trial_params": null
1540
+ }
checkpoint-101000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d42437e01c6a8e691de275f2b2ef4e9ee68d60c829bca2508adb832bd55170f9
3
+ size 4027
checkpoint-101000/vocab.json ADDED
The diff for this file is too large to render. See raw diff