tmnam20 commited on
Commit
b74de91
1 Parent(s): f5dd4e5

Training in progress, step 106000, checkpoint

Browse files
checkpoint-106000/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<bos>": 40478,
3
+ "<eos>": 40479,
4
+ "<pad>": 40480
5
+ }
checkpoint-106000/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-106000/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-106000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:520e7c4223f0c3e08842a8c3cbbd505d0bf0b8b9bcd2b669eb01ecb122d7d0ec
3
+ size 592141381
checkpoint-106000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2107347721afc5012422e30cf3a0338ca64c0f1638e63499d4fe39c6a8b0cd02
3
+ size 296072722
checkpoint-106000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5acbd4e560909ffdd187c18120ce27d61e2e51c71f2939f962a84a33371e628f
3
+ size 14575
checkpoint-106000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4004b29e948e3d51c2adbe68fc924aee1ba6bf53b0b0f23aed4eb679367088e6
3
+ size 627
checkpoint-106000/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-106000/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-106000/trainer_state.json ADDED
@@ -0,0 +1,1615 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.965115487575531,
3
+ "best_model_checkpoint": "/kaggle/output/checkpoint-103000",
4
+ "epoch": 4.31877444589309,
5
+ "eval_steps": 1000,
6
+ "global_step": 106000,
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": 1.05,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 2.7750000000000004e-08,
20
+ "loss": 1.1403,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 0.04,
25
+ "eval_accuracy": 0.33712574850299404,
26
+ "eval_loss": 1.1030837297439575,
27
+ "eval_runtime": 29.7882,
28
+ "eval_samples_per_second": 168.187,
29
+ "eval_steps_per_second": 21.049,
30
+ "step": 1000
31
+ },
32
+ {
33
+ "epoch": 0.08,
34
+ "learning_rate": 5.5527777777777784e-08,
35
+ "loss": 1.1194,
36
+ "step": 2000
37
+ },
38
+ {
39
+ "epoch": 0.08,
40
+ "eval_accuracy": 0.34331337325349304,
41
+ "eval_loss": 1.1002165079116821,
42
+ "eval_runtime": 29.8685,
43
+ "eval_samples_per_second": 167.735,
44
+ "eval_steps_per_second": 20.992,
45
+ "step": 2000
46
+ },
47
+ {
48
+ "epoch": 0.12,
49
+ "learning_rate": 8.327777777777778e-08,
50
+ "loss": 1.1194,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "eval_accuracy": 0.3471057884231537,
56
+ "eval_loss": 1.101266622543335,
57
+ "eval_runtime": 29.709,
58
+ "eval_samples_per_second": 168.636,
59
+ "eval_steps_per_second": 21.105,
60
+ "step": 3000
61
+ },
62
+ {
63
+ "epoch": 0.16,
64
+ "learning_rate": 1.1105555555555557e-07,
65
+ "loss": 1.1166,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 0.16,
70
+ "eval_accuracy": 0.34311377245508984,
71
+ "eval_loss": 1.0984796285629272,
72
+ "eval_runtime": 29.9401,
73
+ "eval_samples_per_second": 167.334,
74
+ "eval_steps_per_second": 20.942,
75
+ "step": 4000
76
+ },
77
+ {
78
+ "epoch": 0.2,
79
+ "learning_rate": 1.3883333333333335e-07,
80
+ "loss": 1.1137,
81
+ "step": 5000
82
+ },
83
+ {
84
+ "epoch": 0.2,
85
+ "eval_accuracy": 0.3403193612774451,
86
+ "eval_loss": 1.0974279642105103,
87
+ "eval_runtime": 29.8185,
88
+ "eval_samples_per_second": 168.016,
89
+ "eval_steps_per_second": 21.027,
90
+ "step": 5000
91
+ },
92
+ {
93
+ "epoch": 0.24,
94
+ "learning_rate": 1.6658333333333335e-07,
95
+ "loss": 1.1109,
96
+ "step": 6000
97
+ },
98
+ {
99
+ "epoch": 0.24,
100
+ "eval_accuracy": 0.3592814371257485,
101
+ "eval_loss": 1.0984892845153809,
102
+ "eval_runtime": 29.893,
103
+ "eval_samples_per_second": 167.598,
104
+ "eval_steps_per_second": 20.975,
105
+ "step": 6000
106
+ },
107
+ {
108
+ "epoch": 0.29,
109
+ "learning_rate": 1.9436111111111112e-07,
110
+ "loss": 1.1122,
111
+ "step": 7000
112
+ },
113
+ {
114
+ "epoch": 0.29,
115
+ "eval_accuracy": 0.36327345309381237,
116
+ "eval_loss": 1.0960686206817627,
117
+ "eval_runtime": 29.8315,
118
+ "eval_samples_per_second": 167.943,
119
+ "eval_steps_per_second": 21.018,
120
+ "step": 7000
121
+ },
122
+ {
123
+ "epoch": 0.33,
124
+ "learning_rate": 2.2213888888888891e-07,
125
+ "loss": 1.1091,
126
+ "step": 8000
127
+ },
128
+ {
129
+ "epoch": 0.33,
130
+ "eval_accuracy": 0.35728542914171657,
131
+ "eval_loss": 1.0958523750305176,
132
+ "eval_runtime": 29.9494,
133
+ "eval_samples_per_second": 167.282,
134
+ "eval_steps_per_second": 20.935,
135
+ "step": 8000
136
+ },
137
+ {
138
+ "epoch": 0.37,
139
+ "learning_rate": 2.4988888888888893e-07,
140
+ "loss": 1.1077,
141
+ "step": 9000
142
+ },
143
+ {
144
+ "epoch": 0.37,
145
+ "eval_accuracy": 0.35528942115768464,
146
+ "eval_loss": 1.1003851890563965,
147
+ "eval_runtime": 29.9762,
148
+ "eval_samples_per_second": 167.133,
149
+ "eval_steps_per_second": 20.917,
150
+ "step": 9000
151
+ },
152
+ {
153
+ "epoch": 0.41,
154
+ "learning_rate": 2.776666666666667e-07,
155
+ "loss": 1.1091,
156
+ "step": 10000
157
+ },
158
+ {
159
+ "epoch": 0.41,
160
+ "eval_accuracy": 0.3790419161676647,
161
+ "eval_loss": 1.0951762199401855,
162
+ "eval_runtime": 29.8593,
163
+ "eval_samples_per_second": 167.787,
164
+ "eval_steps_per_second": 20.998,
165
+ "step": 10000
166
+ },
167
+ {
168
+ "epoch": 0.45,
169
+ "learning_rate": 3.0541666666666667e-07,
170
+ "loss": 1.1047,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.45,
175
+ "eval_accuracy": 0.3780439121756487,
176
+ "eval_loss": 1.0928910970687866,
177
+ "eval_runtime": 29.8768,
178
+ "eval_samples_per_second": 167.689,
179
+ "eval_steps_per_second": 20.986,
180
+ "step": 11000
181
+ },
182
+ {
183
+ "epoch": 0.49,
184
+ "learning_rate": 3.3319444444444444e-07,
185
+ "loss": 1.1053,
186
+ "step": 12000
187
+ },
188
+ {
189
+ "epoch": 0.49,
190
+ "eval_accuracy": 0.38323353293413176,
191
+ "eval_loss": 1.0910215377807617,
192
+ "eval_runtime": 29.8703,
193
+ "eval_samples_per_second": 167.725,
194
+ "eval_steps_per_second": 20.991,
195
+ "step": 12000
196
+ },
197
+ {
198
+ "epoch": 0.53,
199
+ "learning_rate": 3.6094444444444446e-07,
200
+ "loss": 1.1035,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 0.53,
205
+ "eval_accuracy": 0.34850299401197604,
206
+ "eval_loss": 1.0930161476135254,
207
+ "eval_runtime": 29.8865,
208
+ "eval_samples_per_second": 167.634,
209
+ "eval_steps_per_second": 20.979,
210
+ "step": 13000
211
+ },
212
+ {
213
+ "epoch": 0.57,
214
+ "learning_rate": 3.8872222222222223e-07,
215
+ "loss": 1.1002,
216
+ "step": 14000
217
+ },
218
+ {
219
+ "epoch": 0.57,
220
+ "eval_accuracy": 0.3646706586826347,
221
+ "eval_loss": 1.093497633934021,
222
+ "eval_runtime": 30.0222,
223
+ "eval_samples_per_second": 166.877,
224
+ "eval_steps_per_second": 20.885,
225
+ "step": 14000
226
+ },
227
+ {
228
+ "epoch": 0.61,
229
+ "learning_rate": 4.1647222222222225e-07,
230
+ "loss": 1.1025,
231
+ "step": 15000
232
+ },
233
+ {
234
+ "epoch": 0.61,
235
+ "eval_accuracy": 0.37924151696606784,
236
+ "eval_loss": 1.0900779962539673,
237
+ "eval_runtime": 30.0049,
238
+ "eval_samples_per_second": 166.973,
239
+ "eval_steps_per_second": 20.897,
240
+ "step": 15000
241
+ },
242
+ {
243
+ "epoch": 0.65,
244
+ "learning_rate": 4.4425e-07,
245
+ "loss": 1.0992,
246
+ "step": 16000
247
+ },
248
+ {
249
+ "epoch": 0.65,
250
+ "eval_accuracy": 0.40119760479041916,
251
+ "eval_loss": 1.0855979919433594,
252
+ "eval_runtime": 30.1095,
253
+ "eval_samples_per_second": 166.393,
254
+ "eval_steps_per_second": 20.824,
255
+ "step": 16000
256
+ },
257
+ {
258
+ "epoch": 0.69,
259
+ "learning_rate": 4.7200000000000004e-07,
260
+ "loss": 1.0965,
261
+ "step": 17000
262
+ },
263
+ {
264
+ "epoch": 0.69,
265
+ "eval_accuracy": 0.4171656686626746,
266
+ "eval_loss": 1.0845381021499634,
267
+ "eval_runtime": 30.0586,
268
+ "eval_samples_per_second": 166.674,
269
+ "eval_steps_per_second": 20.859,
270
+ "step": 17000
271
+ },
272
+ {
273
+ "epoch": 0.73,
274
+ "learning_rate": 4.997777777777779e-07,
275
+ "loss": 1.0962,
276
+ "step": 18000
277
+ },
278
+ {
279
+ "epoch": 0.73,
280
+ "eval_accuracy": 0.4087824351297405,
281
+ "eval_loss": 1.0860552787780762,
282
+ "eval_runtime": 30.2343,
283
+ "eval_samples_per_second": 165.706,
284
+ "eval_steps_per_second": 20.738,
285
+ "step": 18000
286
+ },
287
+ {
288
+ "epoch": 0.77,
289
+ "learning_rate": 5.275277777777778e-07,
290
+ "loss": 1.0936,
291
+ "step": 19000
292
+ },
293
+ {
294
+ "epoch": 0.77,
295
+ "eval_accuracy": 0.4039920159680639,
296
+ "eval_loss": 1.0822259187698364,
297
+ "eval_runtime": 30.216,
298
+ "eval_samples_per_second": 165.806,
299
+ "eval_steps_per_second": 20.751,
300
+ "step": 19000
301
+ },
302
+ {
303
+ "epoch": 0.81,
304
+ "learning_rate": 5.553055555555556e-07,
305
+ "loss": 1.0923,
306
+ "step": 20000
307
+ },
308
+ {
309
+ "epoch": 0.81,
310
+ "eval_accuracy": 0.4095808383233533,
311
+ "eval_loss": 1.0820951461791992,
312
+ "eval_runtime": 29.968,
313
+ "eval_samples_per_second": 167.178,
314
+ "eval_steps_per_second": 20.922,
315
+ "step": 20000
316
+ },
317
+ {
318
+ "epoch": 0.86,
319
+ "learning_rate": 5.830555555555556e-07,
320
+ "loss": 1.0897,
321
+ "step": 21000
322
+ },
323
+ {
324
+ "epoch": 0.86,
325
+ "eval_accuracy": 0.4149700598802395,
326
+ "eval_loss": 1.0763800144195557,
327
+ "eval_runtime": 30.021,
328
+ "eval_samples_per_second": 166.883,
329
+ "eval_steps_per_second": 20.885,
330
+ "step": 21000
331
+ },
332
+ {
333
+ "epoch": 0.9,
334
+ "learning_rate": 6.108333333333333e-07,
335
+ "loss": 1.0888,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 0.9,
340
+ "eval_accuracy": 0.43193612774451096,
341
+ "eval_loss": 1.0687414407730103,
342
+ "eval_runtime": 30.1577,
343
+ "eval_samples_per_second": 166.127,
344
+ "eval_steps_per_second": 20.791,
345
+ "step": 22000
346
+ },
347
+ {
348
+ "epoch": 0.94,
349
+ "learning_rate": 6.385833333333334e-07,
350
+ "loss": 1.0823,
351
+ "step": 23000
352
+ },
353
+ {
354
+ "epoch": 0.94,
355
+ "eval_accuracy": 0.4285429141716567,
356
+ "eval_loss": 1.0642070770263672,
357
+ "eval_runtime": 30.0701,
358
+ "eval_samples_per_second": 166.611,
359
+ "eval_steps_per_second": 20.851,
360
+ "step": 23000
361
+ },
362
+ {
363
+ "epoch": 0.98,
364
+ "learning_rate": 6.663611111111112e-07,
365
+ "loss": 1.0786,
366
+ "step": 24000
367
+ },
368
+ {
369
+ "epoch": 0.98,
370
+ "eval_accuracy": 0.4363273453093812,
371
+ "eval_loss": 1.0598841905593872,
372
+ "eval_runtime": 30.0689,
373
+ "eval_samples_per_second": 166.617,
374
+ "eval_steps_per_second": 20.852,
375
+ "step": 24000
376
+ },
377
+ {
378
+ "epoch": 1.02,
379
+ "learning_rate": 6.941111111111112e-07,
380
+ "loss": 1.0708,
381
+ "step": 25000
382
+ },
383
+ {
384
+ "epoch": 1.02,
385
+ "eval_accuracy": 0.4429141716566866,
386
+ "eval_loss": 1.0561795234680176,
387
+ "eval_runtime": 30.1301,
388
+ "eval_samples_per_second": 166.279,
389
+ "eval_steps_per_second": 20.81,
390
+ "step": 25000
391
+ },
392
+ {
393
+ "epoch": 1.06,
394
+ "learning_rate": 7.218888888888889e-07,
395
+ "loss": 1.072,
396
+ "step": 26000
397
+ },
398
+ {
399
+ "epoch": 1.06,
400
+ "eval_accuracy": 0.4303393213572854,
401
+ "eval_loss": 1.0574887990951538,
402
+ "eval_runtime": 30.0473,
403
+ "eval_samples_per_second": 166.737,
404
+ "eval_steps_per_second": 20.867,
405
+ "step": 26000
406
+ },
407
+ {
408
+ "epoch": 1.1,
409
+ "learning_rate": 7.496666666666667e-07,
410
+ "loss": 1.0701,
411
+ "step": 27000
412
+ },
413
+ {
414
+ "epoch": 1.1,
415
+ "eval_accuracy": 0.4421157684630739,
416
+ "eval_loss": 1.055459976196289,
417
+ "eval_runtime": 30.1049,
418
+ "eval_samples_per_second": 166.418,
419
+ "eval_steps_per_second": 20.827,
420
+ "step": 27000
421
+ },
422
+ {
423
+ "epoch": 1.14,
424
+ "learning_rate": 7.774166666666668e-07,
425
+ "loss": 1.0677,
426
+ "step": 28000
427
+ },
428
+ {
429
+ "epoch": 1.14,
430
+ "eval_accuracy": 0.4285429141716567,
431
+ "eval_loss": 1.0527067184448242,
432
+ "eval_runtime": 30.1476,
433
+ "eval_samples_per_second": 166.182,
434
+ "eval_steps_per_second": 20.798,
435
+ "step": 28000
436
+ },
437
+ {
438
+ "epoch": 1.18,
439
+ "learning_rate": 8.051944444444445e-07,
440
+ "loss": 1.065,
441
+ "step": 29000
442
+ },
443
+ {
444
+ "epoch": 1.18,
445
+ "eval_accuracy": 0.43852295409181635,
446
+ "eval_loss": 1.0494154691696167,
447
+ "eval_runtime": 30.036,
448
+ "eval_samples_per_second": 166.8,
449
+ "eval_steps_per_second": 20.875,
450
+ "step": 29000
451
+ },
452
+ {
453
+ "epoch": 1.22,
454
+ "learning_rate": 8.329444444444445e-07,
455
+ "loss": 1.0628,
456
+ "step": 30000
457
+ },
458
+ {
459
+ "epoch": 1.22,
460
+ "eval_accuracy": 0.4377245508982036,
461
+ "eval_loss": 1.0461777448654175,
462
+ "eval_runtime": 30.1005,
463
+ "eval_samples_per_second": 166.442,
464
+ "eval_steps_per_second": 20.83,
465
+ "step": 30000
466
+ },
467
+ {
468
+ "epoch": 1.26,
469
+ "learning_rate": 8.607222222222223e-07,
470
+ "loss": 1.0672,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 1.26,
475
+ "eval_accuracy": 0.4435129740518962,
476
+ "eval_loss": 1.0531694889068604,
477
+ "eval_runtime": 30.1733,
478
+ "eval_samples_per_second": 166.041,
479
+ "eval_steps_per_second": 20.78,
480
+ "step": 31000
481
+ },
482
+ {
483
+ "epoch": 1.3,
484
+ "learning_rate": 8.884722222222224e-07,
485
+ "loss": 1.0638,
486
+ "step": 32000
487
+ },
488
+ {
489
+ "epoch": 1.3,
490
+ "eval_accuracy": 0.4297405189620758,
491
+ "eval_loss": 1.0501775741577148,
492
+ "eval_runtime": 30.058,
493
+ "eval_samples_per_second": 166.678,
494
+ "eval_steps_per_second": 20.86,
495
+ "step": 32000
496
+ },
497
+ {
498
+ "epoch": 1.34,
499
+ "learning_rate": 9.162500000000001e-07,
500
+ "loss": 1.0637,
501
+ "step": 33000
502
+ },
503
+ {
504
+ "epoch": 1.34,
505
+ "eval_accuracy": 0.44491017964071855,
506
+ "eval_loss": 1.0452704429626465,
507
+ "eval_runtime": 30.3623,
508
+ "eval_samples_per_second": 165.007,
509
+ "eval_steps_per_second": 20.651,
510
+ "step": 33000
511
+ },
512
+ {
513
+ "epoch": 1.39,
514
+ "learning_rate": 9.440000000000001e-07,
515
+ "loss": 1.0558,
516
+ "step": 34000
517
+ },
518
+ {
519
+ "epoch": 1.39,
520
+ "eval_accuracy": 0.44510978043912175,
521
+ "eval_loss": 1.0434744358062744,
522
+ "eval_runtime": 30.2239,
523
+ "eval_samples_per_second": 165.763,
524
+ "eval_steps_per_second": 20.745,
525
+ "step": 34000
526
+ },
527
+ {
528
+ "epoch": 1.43,
529
+ "learning_rate": 9.71777777777778e-07,
530
+ "loss": 1.0598,
531
+ "step": 35000
532
+ },
533
+ {
534
+ "epoch": 1.43,
535
+ "eval_accuracy": 0.4473053892215569,
536
+ "eval_loss": 1.039801001548767,
537
+ "eval_runtime": 30.107,
538
+ "eval_samples_per_second": 166.407,
539
+ "eval_steps_per_second": 20.826,
540
+ "step": 35000
541
+ },
542
+ {
543
+ "epoch": 1.47,
544
+ "learning_rate": 9.995277777777778e-07,
545
+ "loss": 1.0601,
546
+ "step": 36000
547
+ },
548
+ {
549
+ "epoch": 1.47,
550
+ "eval_accuracy": 0.4469061876247505,
551
+ "eval_loss": 1.037074327468872,
552
+ "eval_runtime": 30.1954,
553
+ "eval_samples_per_second": 165.919,
554
+ "eval_steps_per_second": 20.765,
555
+ "step": 36000
556
+ },
557
+ {
558
+ "epoch": 1.51,
559
+ "learning_rate": 1.0273055555555556e-06,
560
+ "loss": 1.0529,
561
+ "step": 37000
562
+ },
563
+ {
564
+ "epoch": 1.51,
565
+ "eval_accuracy": 0.45229540918163674,
566
+ "eval_loss": 1.0346434116363525,
567
+ "eval_runtime": 30.2856,
568
+ "eval_samples_per_second": 165.425,
569
+ "eval_steps_per_second": 20.703,
570
+ "step": 37000
571
+ },
572
+ {
573
+ "epoch": 1.55,
574
+ "learning_rate": 1.0550555555555557e-06,
575
+ "loss": 1.0544,
576
+ "step": 38000
577
+ },
578
+ {
579
+ "epoch": 1.55,
580
+ "eval_accuracy": 0.45129740518962075,
581
+ "eval_loss": 1.0338674783706665,
582
+ "eval_runtime": 30.1511,
583
+ "eval_samples_per_second": 166.163,
584
+ "eval_steps_per_second": 20.795,
585
+ "step": 38000
586
+ },
587
+ {
588
+ "epoch": 1.59,
589
+ "learning_rate": 1.0828333333333334e-06,
590
+ "loss": 1.0528,
591
+ "step": 39000
592
+ },
593
+ {
594
+ "epoch": 1.59,
595
+ "eval_accuracy": 0.44550898203592815,
596
+ "eval_loss": 1.0376836061477661,
597
+ "eval_runtime": 30.304,
598
+ "eval_samples_per_second": 165.324,
599
+ "eval_steps_per_second": 20.69,
600
+ "step": 39000
601
+ },
602
+ {
603
+ "epoch": 1.63,
604
+ "learning_rate": 1.1106111111111112e-06,
605
+ "loss": 1.0482,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 1.63,
610
+ "eval_accuracy": 0.4540918163672655,
611
+ "eval_loss": 1.031366229057312,
612
+ "eval_runtime": 30.2062,
613
+ "eval_samples_per_second": 165.86,
614
+ "eval_steps_per_second": 20.757,
615
+ "step": 40000
616
+ },
617
+ {
618
+ "epoch": 1.67,
619
+ "learning_rate": 1.1383611111111113e-06,
620
+ "loss": 1.0494,
621
+ "step": 41000
622
+ },
623
+ {
624
+ "epoch": 1.67,
625
+ "eval_accuracy": 0.4552894211576846,
626
+ "eval_loss": 1.032992959022522,
627
+ "eval_runtime": 30.1667,
628
+ "eval_samples_per_second": 166.077,
629
+ "eval_steps_per_second": 20.784,
630
+ "step": 41000
631
+ },
632
+ {
633
+ "epoch": 1.71,
634
+ "learning_rate": 1.166138888888889e-06,
635
+ "loss": 1.0474,
636
+ "step": 42000
637
+ },
638
+ {
639
+ "epoch": 1.71,
640
+ "eval_accuracy": 0.45129740518962075,
641
+ "eval_loss": 1.0255250930786133,
642
+ "eval_runtime": 30.2122,
643
+ "eval_samples_per_second": 165.827,
644
+ "eval_steps_per_second": 20.753,
645
+ "step": 42000
646
+ },
647
+ {
648
+ "epoch": 1.75,
649
+ "learning_rate": 1.193888888888889e-06,
650
+ "loss": 1.0448,
651
+ "step": 43000
652
+ },
653
+ {
654
+ "epoch": 1.75,
655
+ "eval_accuracy": 0.4483033932135729,
656
+ "eval_loss": 1.0412585735321045,
657
+ "eval_runtime": 30.2083,
658
+ "eval_samples_per_second": 165.849,
659
+ "eval_steps_per_second": 20.756,
660
+ "step": 43000
661
+ },
662
+ {
663
+ "epoch": 1.79,
664
+ "learning_rate": 1.2216666666666667e-06,
665
+ "loss": 1.0458,
666
+ "step": 44000
667
+ },
668
+ {
669
+ "epoch": 1.79,
670
+ "eval_accuracy": 0.46427145708582834,
671
+ "eval_loss": 1.0211690664291382,
672
+ "eval_runtime": 30.4051,
673
+ "eval_samples_per_second": 164.775,
674
+ "eval_steps_per_second": 20.622,
675
+ "step": 44000
676
+ },
677
+ {
678
+ "epoch": 1.83,
679
+ "learning_rate": 1.2494166666666668e-06,
680
+ "loss": 1.0468,
681
+ "step": 45000
682
+ },
683
+ {
684
+ "epoch": 1.83,
685
+ "eval_accuracy": 0.47105788423153694,
686
+ "eval_loss": 1.0185887813568115,
687
+ "eval_runtime": 30.1465,
688
+ "eval_samples_per_second": 166.188,
689
+ "eval_steps_per_second": 20.798,
690
+ "step": 45000
691
+ },
692
+ {
693
+ "epoch": 1.87,
694
+ "learning_rate": 1.2771944444444445e-06,
695
+ "loss": 1.0414,
696
+ "step": 46000
697
+ },
698
+ {
699
+ "epoch": 1.87,
700
+ "eval_accuracy": 0.46966067864271455,
701
+ "eval_loss": 1.0215678215026855,
702
+ "eval_runtime": 30.1615,
703
+ "eval_samples_per_second": 166.106,
704
+ "eval_steps_per_second": 20.788,
705
+ "step": 46000
706
+ },
707
+ {
708
+ "epoch": 1.91,
709
+ "learning_rate": 1.3049444444444446e-06,
710
+ "loss": 1.0374,
711
+ "step": 47000
712
+ },
713
+ {
714
+ "epoch": 1.91,
715
+ "eval_accuracy": 0.47684630738522954,
716
+ "eval_loss": 1.0100239515304565,
717
+ "eval_runtime": 30.0919,
718
+ "eval_samples_per_second": 166.49,
719
+ "eval_steps_per_second": 20.836,
720
+ "step": 47000
721
+ },
722
+ {
723
+ "epoch": 1.96,
724
+ "learning_rate": 1.3327222222222224e-06,
725
+ "loss": 1.0423,
726
+ "step": 48000
727
+ },
728
+ {
729
+ "epoch": 1.96,
730
+ "eval_accuracy": 0.462874251497006,
731
+ "eval_loss": 1.0277544260025024,
732
+ "eval_runtime": 30.0926,
733
+ "eval_samples_per_second": 166.486,
734
+ "eval_steps_per_second": 20.836,
735
+ "step": 48000
736
+ },
737
+ {
738
+ "epoch": 2.0,
739
+ "learning_rate": 1.3604722222222224e-06,
740
+ "loss": 1.0332,
741
+ "step": 49000
742
+ },
743
+ {
744
+ "epoch": 2.0,
745
+ "eval_accuracy": 0.4870259481037924,
746
+ "eval_loss": 1.0046255588531494,
747
+ "eval_runtime": 30.0835,
748
+ "eval_samples_per_second": 166.536,
749
+ "eval_steps_per_second": 20.842,
750
+ "step": 49000
751
+ },
752
+ {
753
+ "epoch": 2.04,
754
+ "learning_rate": 1.3882500000000002e-06,
755
+ "loss": 1.0337,
756
+ "step": 50000
757
+ },
758
+ {
759
+ "epoch": 2.04,
760
+ "eval_accuracy": 0.48542914171656687,
761
+ "eval_loss": 1.00551438331604,
762
+ "eval_runtime": 30.1523,
763
+ "eval_samples_per_second": 166.157,
764
+ "eval_steps_per_second": 20.794,
765
+ "step": 50000
766
+ },
767
+ {
768
+ "epoch": 2.08,
769
+ "learning_rate": 1.416e-06,
770
+ "loss": 1.0335,
771
+ "step": 51000
772
+ },
773
+ {
774
+ "epoch": 2.08,
775
+ "eval_accuracy": 0.4782435129740519,
776
+ "eval_loss": 1.0022392272949219,
777
+ "eval_runtime": 30.0244,
778
+ "eval_samples_per_second": 166.864,
779
+ "eval_steps_per_second": 20.883,
780
+ "step": 51000
781
+ },
782
+ {
783
+ "epoch": 2.12,
784
+ "learning_rate": 1.4437777777777779e-06,
785
+ "loss": 1.0272,
786
+ "step": 52000
787
+ },
788
+ {
789
+ "epoch": 2.12,
790
+ "eval_accuracy": 0.48003992015968067,
791
+ "eval_loss": 1.014728307723999,
792
+ "eval_runtime": 30.0497,
793
+ "eval_samples_per_second": 166.724,
794
+ "eval_steps_per_second": 20.865,
795
+ "step": 52000
796
+ },
797
+ {
798
+ "epoch": 2.16,
799
+ "learning_rate": 1.471527777777778e-06,
800
+ "loss": 1.0315,
801
+ "step": 53000
802
+ },
803
+ {
804
+ "epoch": 2.16,
805
+ "eval_accuracy": 0.4848303393213573,
806
+ "eval_loss": 0.9981946349143982,
807
+ "eval_runtime": 30.2291,
808
+ "eval_samples_per_second": 165.734,
809
+ "eval_steps_per_second": 20.742,
810
+ "step": 53000
811
+ },
812
+ {
813
+ "epoch": 2.2,
814
+ "learning_rate": 1.4993055555555557e-06,
815
+ "loss": 1.0291,
816
+ "step": 54000
817
+ },
818
+ {
819
+ "epoch": 2.2,
820
+ "eval_accuracy": 0.49021956087824353,
821
+ "eval_loss": 0.9988633990287781,
822
+ "eval_runtime": 30.0961,
823
+ "eval_samples_per_second": 166.467,
824
+ "eval_steps_per_second": 20.833,
825
+ "step": 54000
826
+ },
827
+ {
828
+ "epoch": 2.24,
829
+ "learning_rate": 1.5270555555555558e-06,
830
+ "loss": 1.027,
831
+ "step": 55000
832
+ },
833
+ {
834
+ "epoch": 2.24,
835
+ "eval_accuracy": 0.48602794411177647,
836
+ "eval_loss": 0.99713534116745,
837
+ "eval_runtime": 30.1599,
838
+ "eval_samples_per_second": 166.115,
839
+ "eval_steps_per_second": 20.789,
840
+ "step": 55000
841
+ },
842
+ {
843
+ "epoch": 2.28,
844
+ "learning_rate": 1.5548333333333335e-06,
845
+ "loss": 1.0215,
846
+ "step": 56000
847
+ },
848
+ {
849
+ "epoch": 2.28,
850
+ "eval_accuracy": 0.46487025948103794,
851
+ "eval_loss": 1.0212756395339966,
852
+ "eval_runtime": 30.2641,
853
+ "eval_samples_per_second": 165.543,
854
+ "eval_steps_per_second": 20.718,
855
+ "step": 56000
856
+ },
857
+ {
858
+ "epoch": 2.32,
859
+ "learning_rate": 1.5825833333333334e-06,
860
+ "loss": 1.0203,
861
+ "step": 57000
862
+ },
863
+ {
864
+ "epoch": 2.32,
865
+ "eval_accuracy": 0.4874251497005988,
866
+ "eval_loss": 0.9981949925422668,
867
+ "eval_runtime": 30.1111,
868
+ "eval_samples_per_second": 166.384,
869
+ "eval_steps_per_second": 20.823,
870
+ "step": 57000
871
+ },
872
+ {
873
+ "epoch": 2.36,
874
+ "learning_rate": 1.6103611111111112e-06,
875
+ "loss": 1.0238,
876
+ "step": 58000
877
+ },
878
+ {
879
+ "epoch": 2.36,
880
+ "eval_accuracy": 0.48562874251497007,
881
+ "eval_loss": 1.0043253898620605,
882
+ "eval_runtime": 30.4057,
883
+ "eval_samples_per_second": 164.771,
884
+ "eval_steps_per_second": 20.621,
885
+ "step": 58000
886
+ },
887
+ {
888
+ "epoch": 2.4,
889
+ "learning_rate": 1.638111111111111e-06,
890
+ "loss": 1.0211,
891
+ "step": 59000
892
+ },
893
+ {
894
+ "epoch": 2.4,
895
+ "eval_accuracy": 0.49820359281437127,
896
+ "eval_loss": 0.9920447468757629,
897
+ "eval_runtime": 30.4581,
898
+ "eval_samples_per_second": 164.488,
899
+ "eval_steps_per_second": 20.586,
900
+ "step": 59000
901
+ },
902
+ {
903
+ "epoch": 2.44,
904
+ "learning_rate": 1.665888888888889e-06,
905
+ "loss": 1.0236,
906
+ "step": 60000
907
+ },
908
+ {
909
+ "epoch": 2.44,
910
+ "eval_accuracy": 0.49241516966067866,
911
+ "eval_loss": 0.9953387975692749,
912
+ "eval_runtime": 30.3298,
913
+ "eval_samples_per_second": 165.184,
914
+ "eval_steps_per_second": 20.673,
915
+ "step": 60000
916
+ },
917
+ {
918
+ "epoch": 2.49,
919
+ "learning_rate": 1.693638888888889e-06,
920
+ "loss": 1.0207,
921
+ "step": 61000
922
+ },
923
+ {
924
+ "epoch": 2.49,
925
+ "eval_accuracy": 0.49760479041916167,
926
+ "eval_loss": 0.9943307042121887,
927
+ "eval_runtime": 30.3886,
928
+ "eval_samples_per_second": 164.865,
929
+ "eval_steps_per_second": 20.633,
930
+ "step": 61000
931
+ },
932
+ {
933
+ "epoch": 2.53,
934
+ "learning_rate": 1.7214166666666666e-06,
935
+ "loss": 1.0222,
936
+ "step": 62000
937
+ },
938
+ {
939
+ "epoch": 2.53,
940
+ "eval_accuracy": 0.4880239520958084,
941
+ "eval_loss": 1.005075454711914,
942
+ "eval_runtime": 30.41,
943
+ "eval_samples_per_second": 164.748,
944
+ "eval_steps_per_second": 20.618,
945
+ "step": 62000
946
+ },
947
+ {
948
+ "epoch": 2.57,
949
+ "learning_rate": 1.7491666666666667e-06,
950
+ "loss": 1.0221,
951
+ "step": 63000
952
+ },
953
+ {
954
+ "epoch": 2.57,
955
+ "eval_accuracy": 0.5011976047904192,
956
+ "eval_loss": 0.9902356266975403,
957
+ "eval_runtime": 30.3384,
958
+ "eval_samples_per_second": 165.137,
959
+ "eval_steps_per_second": 20.667,
960
+ "step": 63000
961
+ },
962
+ {
963
+ "epoch": 2.61,
964
+ "learning_rate": 1.7769444444444447e-06,
965
+ "loss": 1.0188,
966
+ "step": 64000
967
+ },
968
+ {
969
+ "epoch": 2.61,
970
+ "eval_accuracy": 0.5003992015968064,
971
+ "eval_loss": 0.9907160997390747,
972
+ "eval_runtime": 30.3592,
973
+ "eval_samples_per_second": 165.024,
974
+ "eval_steps_per_second": 20.653,
975
+ "step": 64000
976
+ },
977
+ {
978
+ "epoch": 2.65,
979
+ "learning_rate": 1.8046944444444446e-06,
980
+ "loss": 1.0187,
981
+ "step": 65000
982
+ },
983
+ {
984
+ "epoch": 2.65,
985
+ "eval_accuracy": 0.49820359281437127,
986
+ "eval_loss": 0.9914441704750061,
987
+ "eval_runtime": 30.3614,
988
+ "eval_samples_per_second": 165.012,
989
+ "eval_steps_per_second": 20.651,
990
+ "step": 65000
991
+ },
992
+ {
993
+ "epoch": 2.69,
994
+ "learning_rate": 1.8324722222222223e-06,
995
+ "loss": 1.023,
996
+ "step": 66000
997
+ },
998
+ {
999
+ "epoch": 2.69,
1000
+ "eval_accuracy": 0.5007984031936128,
1001
+ "eval_loss": 0.9877445101737976,
1002
+ "eval_runtime": 30.341,
1003
+ "eval_samples_per_second": 165.123,
1004
+ "eval_steps_per_second": 20.665,
1005
+ "step": 66000
1006
+ },
1007
+ {
1008
+ "epoch": 2.73,
1009
+ "learning_rate": 1.8602222222222222e-06,
1010
+ "loss": 1.0195,
1011
+ "step": 67000
1012
+ },
1013
+ {
1014
+ "epoch": 2.73,
1015
+ "eval_accuracy": 0.49660678642714573,
1016
+ "eval_loss": 0.9916561245918274,
1017
+ "eval_runtime": 30.3344,
1018
+ "eval_samples_per_second": 165.159,
1019
+ "eval_steps_per_second": 20.67,
1020
+ "step": 67000
1021
+ },
1022
+ {
1023
+ "epoch": 2.77,
1024
+ "learning_rate": 1.8880000000000002e-06,
1025
+ "loss": 1.0181,
1026
+ "step": 68000
1027
+ },
1028
+ {
1029
+ "epoch": 2.77,
1030
+ "eval_accuracy": 0.4992015968063872,
1031
+ "eval_loss": 0.985427975654602,
1032
+ "eval_runtime": 30.3261,
1033
+ "eval_samples_per_second": 165.204,
1034
+ "eval_steps_per_second": 20.675,
1035
+ "step": 68000
1036
+ },
1037
+ {
1038
+ "epoch": 2.81,
1039
+ "learning_rate": 1.9157500000000003e-06,
1040
+ "loss": 1.0183,
1041
+ "step": 69000
1042
+ },
1043
+ {
1044
+ "epoch": 2.81,
1045
+ "eval_accuracy": 0.5033932135728543,
1046
+ "eval_loss": 0.9862416982650757,
1047
+ "eval_runtime": 30.3869,
1048
+ "eval_samples_per_second": 164.874,
1049
+ "eval_steps_per_second": 20.634,
1050
+ "step": 69000
1051
+ },
1052
+ {
1053
+ "epoch": 2.85,
1054
+ "learning_rate": 1.943527777777778e-06,
1055
+ "loss": 1.0176,
1056
+ "step": 70000
1057
+ },
1058
+ {
1059
+ "epoch": 2.85,
1060
+ "eval_accuracy": 0.501996007984032,
1061
+ "eval_loss": 0.9887968897819519,
1062
+ "eval_runtime": 30.3292,
1063
+ "eval_samples_per_second": 165.187,
1064
+ "eval_steps_per_second": 20.673,
1065
+ "step": 70000
1066
+ },
1067
+ {
1068
+ "epoch": 2.89,
1069
+ "learning_rate": 1.9712777777777777e-06,
1070
+ "loss": 1.0108,
1071
+ "step": 71000
1072
+ },
1073
+ {
1074
+ "epoch": 2.89,
1075
+ "eval_accuracy": 0.49520958083832334,
1076
+ "eval_loss": 0.992777407169342,
1077
+ "eval_runtime": 30.4,
1078
+ "eval_samples_per_second": 164.803,
1079
+ "eval_steps_per_second": 20.625,
1080
+ "step": 71000
1081
+ },
1082
+ {
1083
+ "epoch": 2.93,
1084
+ "learning_rate": 1.9990555555555557e-06,
1085
+ "loss": 1.0172,
1086
+ "step": 72000
1087
+ },
1088
+ {
1089
+ "epoch": 2.93,
1090
+ "eval_accuracy": 0.49141716566866267,
1091
+ "eval_loss": 0.9982264041900635,
1092
+ "eval_runtime": 30.36,
1093
+ "eval_samples_per_second": 165.02,
1094
+ "eval_steps_per_second": 20.652,
1095
+ "step": 72000
1096
+ },
1097
+ {
1098
+ "epoch": 2.97,
1099
+ "learning_rate": 2.026805555555556e-06,
1100
+ "loss": 1.0151,
1101
+ "step": 73000
1102
+ },
1103
+ {
1104
+ "epoch": 2.97,
1105
+ "eval_accuracy": 0.49720558882235527,
1106
+ "eval_loss": 0.9918563365936279,
1107
+ "eval_runtime": 30.3051,
1108
+ "eval_samples_per_second": 165.318,
1109
+ "eval_steps_per_second": 20.69,
1110
+ "step": 73000
1111
+ },
1112
+ {
1113
+ "epoch": 3.01,
1114
+ "learning_rate": 2.0545833333333335e-06,
1115
+ "loss": 1.0129,
1116
+ "step": 74000
1117
+ },
1118
+ {
1119
+ "epoch": 3.01,
1120
+ "eval_accuracy": 0.49840319361277446,
1121
+ "eval_loss": 0.9832558631896973,
1122
+ "eval_runtime": 30.3597,
1123
+ "eval_samples_per_second": 165.022,
1124
+ "eval_steps_per_second": 20.652,
1125
+ "step": 74000
1126
+ },
1127
+ {
1128
+ "epoch": 3.06,
1129
+ "learning_rate": 2.0823611111111115e-06,
1130
+ "loss": 1.006,
1131
+ "step": 75000
1132
+ },
1133
+ {
1134
+ "epoch": 3.06,
1135
+ "eval_accuracy": 0.5033932135728543,
1136
+ "eval_loss": 0.9880185127258301,
1137
+ "eval_runtime": 30.2854,
1138
+ "eval_samples_per_second": 165.426,
1139
+ "eval_steps_per_second": 20.703,
1140
+ "step": 75000
1141
+ },
1142
+ {
1143
+ "epoch": 3.1,
1144
+ "learning_rate": 2.1101111111111113e-06,
1145
+ "loss": 1.0069,
1146
+ "step": 76000
1147
+ },
1148
+ {
1149
+ "epoch": 3.1,
1150
+ "eval_accuracy": 0.5045908183632735,
1151
+ "eval_loss": 0.9803372025489807,
1152
+ "eval_runtime": 30.3285,
1153
+ "eval_samples_per_second": 165.191,
1154
+ "eval_steps_per_second": 20.674,
1155
+ "step": 76000
1156
+ },
1157
+ {
1158
+ "epoch": 3.14,
1159
+ "learning_rate": 2.137888888888889e-06,
1160
+ "loss": 1.0106,
1161
+ "step": 77000
1162
+ },
1163
+ {
1164
+ "epoch": 3.14,
1165
+ "eval_accuracy": 0.47724550898203594,
1166
+ "eval_loss": 1.0122599601745605,
1167
+ "eval_runtime": 30.3162,
1168
+ "eval_samples_per_second": 165.258,
1169
+ "eval_steps_per_second": 20.682,
1170
+ "step": 77000
1171
+ },
1172
+ {
1173
+ "epoch": 3.18,
1174
+ "learning_rate": 2.1656388888888888e-06,
1175
+ "loss": 1.0019,
1176
+ "step": 78000
1177
+ },
1178
+ {
1179
+ "epoch": 3.18,
1180
+ "eval_accuracy": 0.5045908183632735,
1181
+ "eval_loss": 0.9800828099250793,
1182
+ "eval_runtime": 30.2724,
1183
+ "eval_samples_per_second": 165.497,
1184
+ "eval_steps_per_second": 20.712,
1185
+ "step": 78000
1186
+ },
1187
+ {
1188
+ "epoch": 3.22,
1189
+ "learning_rate": 2.1934166666666667e-06,
1190
+ "loss": 1.0093,
1191
+ "step": 79000
1192
+ },
1193
+ {
1194
+ "epoch": 3.22,
1195
+ "eval_accuracy": 0.493812375249501,
1196
+ "eval_loss": 0.9990217089653015,
1197
+ "eval_runtime": 30.2198,
1198
+ "eval_samples_per_second": 165.785,
1199
+ "eval_steps_per_second": 20.748,
1200
+ "step": 79000
1201
+ },
1202
+ {
1203
+ "epoch": 3.26,
1204
+ "learning_rate": 2.221166666666667e-06,
1205
+ "loss": 1.0039,
1206
+ "step": 80000
1207
+ },
1208
+ {
1209
+ "epoch": 3.26,
1210
+ "eval_accuracy": 0.5067864271457085,
1211
+ "eval_loss": 0.98616623878479,
1212
+ "eval_runtime": 30.3059,
1213
+ "eval_samples_per_second": 165.315,
1214
+ "eval_steps_per_second": 20.689,
1215
+ "step": 80000
1216
+ },
1217
+ {
1218
+ "epoch": 3.3,
1219
+ "learning_rate": 2.2489444444444446e-06,
1220
+ "loss": 1.0062,
1221
+ "step": 81000
1222
+ },
1223
+ {
1224
+ "epoch": 3.3,
1225
+ "eval_accuracy": 0.5013972055888224,
1226
+ "eval_loss": 0.9929732084274292,
1227
+ "eval_runtime": 30.2754,
1228
+ "eval_samples_per_second": 165.481,
1229
+ "eval_steps_per_second": 20.71,
1230
+ "step": 81000
1231
+ },
1232
+ {
1233
+ "epoch": 3.34,
1234
+ "learning_rate": 2.2766944444444444e-06,
1235
+ "loss": 1.0145,
1236
+ "step": 82000
1237
+ },
1238
+ {
1239
+ "epoch": 3.34,
1240
+ "eval_accuracy": 0.5051896207584831,
1241
+ "eval_loss": 0.9891365766525269,
1242
+ "eval_runtime": 30.204,
1243
+ "eval_samples_per_second": 165.872,
1244
+ "eval_steps_per_second": 20.759,
1245
+ "step": 82000
1246
+ },
1247
+ {
1248
+ "epoch": 3.38,
1249
+ "learning_rate": 2.3044722222222224e-06,
1250
+ "loss": 1.0012,
1251
+ "step": 83000
1252
+ },
1253
+ {
1254
+ "epoch": 3.38,
1255
+ "eval_accuracy": 0.5067864271457085,
1256
+ "eval_loss": 0.984894871711731,
1257
+ "eval_runtime": 30.2196,
1258
+ "eval_samples_per_second": 165.787,
1259
+ "eval_steps_per_second": 20.748,
1260
+ "step": 83000
1261
+ },
1262
+ {
1263
+ "epoch": 3.42,
1264
+ "learning_rate": 2.3322222222222223e-06,
1265
+ "loss": 1.0072,
1266
+ "step": 84000
1267
+ },
1268
+ {
1269
+ "epoch": 3.42,
1270
+ "eval_accuracy": 0.5075848303393213,
1271
+ "eval_loss": 0.979631781578064,
1272
+ "eval_runtime": 30.1636,
1273
+ "eval_samples_per_second": 166.094,
1274
+ "eval_steps_per_second": 20.787,
1275
+ "step": 84000
1276
+ },
1277
+ {
1278
+ "epoch": 3.46,
1279
+ "learning_rate": 2.36e-06,
1280
+ "loss": 1.0038,
1281
+ "step": 85000
1282
+ },
1283
+ {
1284
+ "epoch": 3.46,
1285
+ "eval_accuracy": 0.5035928143712575,
1286
+ "eval_loss": 0.9814818501472473,
1287
+ "eval_runtime": 30.2922,
1288
+ "eval_samples_per_second": 165.389,
1289
+ "eval_steps_per_second": 20.698,
1290
+ "step": 85000
1291
+ },
1292
+ {
1293
+ "epoch": 3.5,
1294
+ "learning_rate": 2.38775e-06,
1295
+ "loss": 1.0107,
1296
+ "step": 86000
1297
+ },
1298
+ {
1299
+ "epoch": 3.5,
1300
+ "eval_accuracy": 0.5117764471057884,
1301
+ "eval_loss": 0.9743403792381287,
1302
+ "eval_runtime": 30.2926,
1303
+ "eval_samples_per_second": 165.387,
1304
+ "eval_steps_per_second": 20.698,
1305
+ "step": 86000
1306
+ },
1307
+ {
1308
+ "epoch": 3.54,
1309
+ "learning_rate": 2.415527777777778e-06,
1310
+ "loss": 1.0011,
1311
+ "step": 87000
1312
+ },
1313
+ {
1314
+ "epoch": 3.54,
1315
+ "eval_accuracy": 0.5073852295409181,
1316
+ "eval_loss": 0.9760627150535583,
1317
+ "eval_runtime": 30.3925,
1318
+ "eval_samples_per_second": 164.843,
1319
+ "eval_steps_per_second": 20.63,
1320
+ "step": 87000
1321
+ },
1322
+ {
1323
+ "epoch": 3.59,
1324
+ "learning_rate": 2.443277777777778e-06,
1325
+ "loss": 1.0045,
1326
+ "step": 88000
1327
+ },
1328
+ {
1329
+ "epoch": 3.59,
1330
+ "eval_accuracy": 0.5101796407185629,
1331
+ "eval_loss": 0.9735654592514038,
1332
+ "eval_runtime": 30.2088,
1333
+ "eval_samples_per_second": 165.846,
1334
+ "eval_steps_per_second": 20.756,
1335
+ "step": 88000
1336
+ },
1337
+ {
1338
+ "epoch": 3.63,
1339
+ "learning_rate": 2.4710555555555555e-06,
1340
+ "loss": 0.999,
1341
+ "step": 89000
1342
+ },
1343
+ {
1344
+ "epoch": 3.63,
1345
+ "eval_accuracy": 0.4996007984031936,
1346
+ "eval_loss": 0.9788551926612854,
1347
+ "eval_runtime": 30.2584,
1348
+ "eval_samples_per_second": 165.574,
1349
+ "eval_steps_per_second": 20.721,
1350
+ "step": 89000
1351
+ },
1352
+ {
1353
+ "epoch": 3.67,
1354
+ "learning_rate": 2.498805555555556e-06,
1355
+ "loss": 1.0075,
1356
+ "step": 90000
1357
+ },
1358
+ {
1359
+ "epoch": 3.67,
1360
+ "eval_accuracy": 0.5095808383233533,
1361
+ "eval_loss": 0.972398579120636,
1362
+ "eval_runtime": 30.1697,
1363
+ "eval_samples_per_second": 166.061,
1364
+ "eval_steps_per_second": 20.782,
1365
+ "step": 90000
1366
+ },
1367
+ {
1368
+ "epoch": 3.71,
1369
+ "learning_rate": 2.5265833333333334e-06,
1370
+ "loss": 1.0085,
1371
+ "step": 91000
1372
+ },
1373
+ {
1374
+ "epoch": 3.71,
1375
+ "eval_accuracy": 0.5073852295409181,
1376
+ "eval_loss": 0.9791931509971619,
1377
+ "eval_runtime": 30.1687,
1378
+ "eval_samples_per_second": 166.066,
1379
+ "eval_steps_per_second": 20.783,
1380
+ "step": 91000
1381
+ },
1382
+ {
1383
+ "epoch": 3.75,
1384
+ "learning_rate": 2.5543333333333337e-06,
1385
+ "loss": 1.0022,
1386
+ "step": 92000
1387
+ },
1388
+ {
1389
+ "epoch": 3.75,
1390
+ "eval_accuracy": 0.5105788423153692,
1391
+ "eval_loss": 0.9835652112960815,
1392
+ "eval_runtime": 30.389,
1393
+ "eval_samples_per_second": 164.862,
1394
+ "eval_steps_per_second": 20.632,
1395
+ "step": 92000
1396
+ },
1397
+ {
1398
+ "epoch": 3.79,
1399
+ "learning_rate": 2.5821111111111112e-06,
1400
+ "loss": 1.006,
1401
+ "step": 93000
1402
+ },
1403
+ {
1404
+ "epoch": 3.79,
1405
+ "eval_accuracy": 0.5127744510978044,
1406
+ "eval_loss": 0.9716958403587341,
1407
+ "eval_runtime": 30.2822,
1408
+ "eval_samples_per_second": 165.444,
1409
+ "eval_steps_per_second": 20.705,
1410
+ "step": 93000
1411
+ },
1412
+ {
1413
+ "epoch": 3.83,
1414
+ "learning_rate": 2.609861111111111e-06,
1415
+ "loss": 1.0001,
1416
+ "step": 94000
1417
+ },
1418
+ {
1419
+ "epoch": 3.83,
1420
+ "eval_accuracy": 0.49880239520958086,
1421
+ "eval_loss": 0.982902467250824,
1422
+ "eval_runtime": 30.3141,
1423
+ "eval_samples_per_second": 165.27,
1424
+ "eval_steps_per_second": 20.683,
1425
+ "step": 94000
1426
+ },
1427
+ {
1428
+ "epoch": 3.87,
1429
+ "learning_rate": 2.637638888888889e-06,
1430
+ "loss": 1.0018,
1431
+ "step": 95000
1432
+ },
1433
+ {
1434
+ "epoch": 3.87,
1435
+ "eval_accuracy": 0.5117764471057884,
1436
+ "eval_loss": 0.9711886644363403,
1437
+ "eval_runtime": 30.3419,
1438
+ "eval_samples_per_second": 165.118,
1439
+ "eval_steps_per_second": 20.665,
1440
+ "step": 95000
1441
+ },
1442
+ {
1443
+ "epoch": 3.91,
1444
+ "learning_rate": 2.6653888888888894e-06,
1445
+ "loss": 1.001,
1446
+ "step": 96000
1447
+ },
1448
+ {
1449
+ "epoch": 3.91,
1450
+ "eval_accuracy": 0.5085828343313373,
1451
+ "eval_loss": 0.9692357182502747,
1452
+ "eval_runtime": 30.2793,
1453
+ "eval_samples_per_second": 165.46,
1454
+ "eval_steps_per_second": 20.707,
1455
+ "step": 96000
1456
+ },
1457
+ {
1458
+ "epoch": 3.95,
1459
+ "learning_rate": 2.693166666666667e-06,
1460
+ "loss": 0.9967,
1461
+ "step": 97000
1462
+ },
1463
+ {
1464
+ "epoch": 3.95,
1465
+ "eval_accuracy": 0.5035928143712575,
1466
+ "eval_loss": 0.9811190366744995,
1467
+ "eval_runtime": 30.3522,
1468
+ "eval_samples_per_second": 165.062,
1469
+ "eval_steps_per_second": 20.657,
1470
+ "step": 97000
1471
+ },
1472
+ {
1473
+ "epoch": 3.99,
1474
+ "learning_rate": 2.7209166666666668e-06,
1475
+ "loss": 0.9963,
1476
+ "step": 98000
1477
+ },
1478
+ {
1479
+ "epoch": 3.99,
1480
+ "eval_accuracy": 0.5117764471057884,
1481
+ "eval_loss": 0.9744167923927307,
1482
+ "eval_runtime": 30.2586,
1483
+ "eval_samples_per_second": 165.573,
1484
+ "eval_steps_per_second": 20.721,
1485
+ "step": 98000
1486
+ },
1487
+ {
1488
+ "epoch": 4.03,
1489
+ "learning_rate": 2.7486944444444448e-06,
1490
+ "loss": 0.9923,
1491
+ "step": 99000
1492
+ },
1493
+ {
1494
+ "epoch": 4.03,
1495
+ "eval_accuracy": 0.5155688622754491,
1496
+ "eval_loss": 0.9705401659011841,
1497
+ "eval_runtime": 30.3652,
1498
+ "eval_samples_per_second": 164.992,
1499
+ "eval_steps_per_second": 20.649,
1500
+ "step": 99000
1501
+ },
1502
+ {
1503
+ "epoch": 4.07,
1504
+ "learning_rate": 2.7764444444444446e-06,
1505
+ "loss": 0.9925,
1506
+ "step": 100000
1507
+ },
1508
+ {
1509
+ "epoch": 4.07,
1510
+ "eval_accuracy": 0.5161676646706587,
1511
+ "eval_loss": 0.967511773109436,
1512
+ "eval_runtime": 30.3989,
1513
+ "eval_samples_per_second": 164.809,
1514
+ "eval_steps_per_second": 20.626,
1515
+ "step": 100000
1516
+ },
1517
+ {
1518
+ "epoch": 4.12,
1519
+ "learning_rate": 2.804222222222222e-06,
1520
+ "loss": 0.9965,
1521
+ "step": 101000
1522
+ },
1523
+ {
1524
+ "epoch": 4.12,
1525
+ "eval_accuracy": 0.5179640718562875,
1526
+ "eval_loss": 0.9703009724617004,
1527
+ "eval_runtime": 30.306,
1528
+ "eval_samples_per_second": 165.314,
1529
+ "eval_steps_per_second": 20.689,
1530
+ "step": 101000
1531
+ },
1532
+ {
1533
+ "epoch": 4.16,
1534
+ "learning_rate": 2.8319722222222225e-06,
1535
+ "loss": 0.994,
1536
+ "step": 102000
1537
+ },
1538
+ {
1539
+ "epoch": 4.16,
1540
+ "eval_accuracy": 0.5027944111776447,
1541
+ "eval_loss": 0.9799131155014038,
1542
+ "eval_runtime": 30.4057,
1543
+ "eval_samples_per_second": 164.772,
1544
+ "eval_steps_per_second": 20.621,
1545
+ "step": 102000
1546
+ },
1547
+ {
1548
+ "epoch": 4.2,
1549
+ "learning_rate": 2.8597500000000004e-06,
1550
+ "loss": 0.9989,
1551
+ "step": 103000
1552
+ },
1553
+ {
1554
+ "epoch": 4.2,
1555
+ "eval_accuracy": 0.5175648702594811,
1556
+ "eval_loss": 0.965115487575531,
1557
+ "eval_runtime": 30.5041,
1558
+ "eval_samples_per_second": 164.24,
1559
+ "eval_steps_per_second": 20.555,
1560
+ "step": 103000
1561
+ },
1562
+ {
1563
+ "epoch": 4.24,
1564
+ "learning_rate": 2.8875000000000003e-06,
1565
+ "loss": 0.9909,
1566
+ "step": 104000
1567
+ },
1568
+ {
1569
+ "epoch": 4.24,
1570
+ "eval_accuracy": 0.5005988023952096,
1571
+ "eval_loss": 0.9864978790283203,
1572
+ "eval_runtime": 30.393,
1573
+ "eval_samples_per_second": 164.84,
1574
+ "eval_steps_per_second": 20.63,
1575
+ "step": 104000
1576
+ },
1577
+ {
1578
+ "epoch": 4.28,
1579
+ "learning_rate": 2.915277777777778e-06,
1580
+ "loss": 0.9873,
1581
+ "step": 105000
1582
+ },
1583
+ {
1584
+ "epoch": 4.28,
1585
+ "eval_accuracy": 0.5017964071856288,
1586
+ "eval_loss": 0.9913762211799622,
1587
+ "eval_runtime": 30.3351,
1588
+ "eval_samples_per_second": 165.155,
1589
+ "eval_steps_per_second": 20.669,
1590
+ "step": 105000
1591
+ },
1592
+ {
1593
+ "epoch": 4.32,
1594
+ "learning_rate": 2.9430277777777777e-06,
1595
+ "loss": 0.9826,
1596
+ "step": 106000
1597
+ },
1598
+ {
1599
+ "epoch": 4.32,
1600
+ "eval_accuracy": 0.5201596806387225,
1601
+ "eval_loss": 0.9676294922828674,
1602
+ "eval_runtime": 30.2604,
1603
+ "eval_samples_per_second": 165.563,
1604
+ "eval_steps_per_second": 20.72,
1605
+ "step": 106000
1606
+ }
1607
+ ],
1608
+ "logging_steps": 1000,
1609
+ "max_steps": 10000000,
1610
+ "num_train_epochs": 408,
1611
+ "save_steps": 1000,
1612
+ "total_flos": 1.1079130033107763e+17,
1613
+ "trial_name": null,
1614
+ "trial_params": null
1615
+ }
checkpoint-106000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40a0dfc08574fef5341c0802f4b4b8be93cd36a17a96b8a7b07271ea753deed4
3
+ size 4027
checkpoint-106000/vocab.json ADDED
The diff for this file is too large to render. See raw diff