HongzheBi commited on
Commit
5eacd2c
·
1 Parent(s): a6c3dfc

Upload 8 files

Browse files
adapter_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "THUDM/chatglm-6b",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "lora_alpha": 32.0,
8
+ "lora_dropout": 0.1,
9
+ "modules_to_save": null,
10
+ "peft_type": "LORA",
11
+ "r": 8,
12
+ "target_modules": [
13
+ "query_key_value"
14
+ ],
15
+ "task_type": "CAUSAL_LM"
16
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09268f26742768d3d1707ad462af2b988a6d595f62f89174a63194640d6da047
3
+ size 14700057
finetuning_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d1b487590fc0e2f1630c709e1d857fe7f42b969b89a6dd35cf99432b03f3799
3
+ size 895
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ca17ca4714d678e5f85d79add8e393b12d1e877d88c1cd87e4b353637d15045
3
+ size 29407173
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1c96494b2e1d9b61c9e04bc026fae37ea5981acaf4cd8b19f8b524f92c836d9
3
+ size 14575
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:193031841c2511a6f76974c5d1670ea8dd6849f75aa64ff306fe78e48fa4f6ff
3
+ size 627
trainer_state.json ADDED
@@ -0,0 +1,1816 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.7198549410698094,
5
+ "global_step": 3000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 4.999908736188902e-05,
13
+ "loss": 1.3169,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.02,
18
+ "learning_rate": 4.999593264902687e-05,
19
+ "loss": 1.0577,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.03,
24
+ "learning_rate": 4.999052487856493e-05,
25
+ "loss": 0.9159,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.04,
30
+ "learning_rate": 4.998286453794323e-05,
31
+ "loss": 0.8264,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.05,
36
+ "learning_rate": 4.9972952317641664e-05,
37
+ "loss": 0.7412,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.05,
42
+ "learning_rate": 4.996078911111772e-05,
43
+ "loss": 0.6837,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.06,
48
+ "learning_rate": 4.994637601472596e-05,
49
+ "loss": 0.7422,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.07,
54
+ "learning_rate": 4.99297143276192e-05,
55
+ "loss": 0.6616,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.08,
60
+ "learning_rate": 4.9910805551631424e-05,
61
+ "loss": 0.7062,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.09,
66
+ "learning_rate": 4.988965139114237e-05,
67
+ "loss": 0.657,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.1,
72
+ "learning_rate": 4.986625375292394e-05,
73
+ "loss": 0.6791,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.11,
78
+ "learning_rate": 4.984061474596833e-05,
79
+ "loss": 0.6274,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.12,
84
+ "learning_rate": 4.9812736681297904e-05,
85
+ "loss": 0.6117,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.13,
90
+ "learning_rate": 4.978262207175689e-05,
91
+ "loss": 0.63,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.14,
96
+ "learning_rate": 4.9750273631784924e-05,
97
+ "loss": 0.6091,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.15,
102
+ "learning_rate": 4.97227884745494e-05,
103
+ "loss": 0.633,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.15,
108
+ "learning_rate": 4.968642662273368e-05,
109
+ "loss": 0.6445,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.16,
114
+ "learning_rate": 4.9647839611253184e-05,
115
+ "loss": 0.6469,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.17,
120
+ "learning_rate": 4.960703091822416e-05,
121
+ "loss": 0.588,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.18,
126
+ "learning_rate": 4.95640042220185e-05,
127
+ "loss": 0.5695,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.19,
132
+ "learning_rate": 4.9518763400932166e-05,
133
+ "loss": 0.6091,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.2,
138
+ "learning_rate": 4.9471312532835645e-05,
139
+ "loss": 0.6153,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.21,
144
+ "learning_rate": 4.942165589480634e-05,
145
+ "loss": 0.6056,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.22,
150
+ "learning_rate": 4.936979796274309e-05,
151
+ "loss": 0.5868,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.23,
156
+ "learning_rate": 4.931574341096269e-05,
157
+ "loss": 0.604,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.24,
162
+ "learning_rate": 4.925949711177859e-05,
163
+ "loss": 0.5964,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.24,
168
+ "learning_rate": 4.9201064135061694e-05,
169
+ "loss": 0.5506,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.25,
174
+ "learning_rate": 4.914044974778341e-05,
175
+ "loss": 0.602,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.26,
180
+ "learning_rate": 4.907765941354085e-05,
181
+ "loss": 0.5943,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.27,
186
+ "learning_rate": 4.9012698792064426e-05,
187
+ "loss": 0.565,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.28,
192
+ "learning_rate": 4.894557373870764e-05,
193
+ "loss": 0.5818,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.29,
198
+ "learning_rate": 4.887629030391932e-05,
199
+ "loss": 0.6095,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.3,
204
+ "learning_rate": 4.880485473269828e-05,
205
+ "loss": 0.5539,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.31,
210
+ "learning_rate": 4.873127346403035e-05,
211
+ "loss": 0.5863,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.32,
216
+ "learning_rate": 4.865555313030805e-05,
217
+ "loss": 0.5869,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.33,
222
+ "learning_rate": 4.857770055673274e-05,
223
+ "loss": 0.5819,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.34,
228
+ "learning_rate": 4.849772276069941e-05,
229
+ "loss": 0.6066,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.34,
234
+ "learning_rate": 4.841562695116415e-05,
235
+ "loss": 0.5623,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.35,
240
+ "learning_rate": 4.833142052799439e-05,
241
+ "loss": 0.5557,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 0.36,
246
+ "learning_rate": 4.824511108130186e-05,
247
+ "loss": 0.5869,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 0.37,
252
+ "learning_rate": 4.815670639075845e-05,
253
+ "loss": 0.5145,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 0.38,
258
+ "learning_rate": 4.8066214424895e-05,
259
+ "loss": 0.5519,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 0.39,
264
+ "learning_rate": 4.7973643340383004e-05,
265
+ "loss": 0.6257,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 0.4,
270
+ "learning_rate": 4.7879001481299404e-05,
271
+ "loss": 0.5976,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 0.41,
276
+ "learning_rate": 4.778229737837451e-05,
277
+ "loss": 0.585,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 0.42,
282
+ "learning_rate": 4.768353974822301e-05,
283
+ "loss": 0.6175,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 0.43,
288
+ "learning_rate": 4.758273749255833e-05,
289
+ "loss": 0.6172,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 0.44,
294
+ "learning_rate": 4.7479899697390235e-05,
295
+ "loss": 0.5003,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 0.44,
300
+ "learning_rate": 4.737503563220586e-05,
301
+ "loss": 0.5583,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 0.45,
306
+ "learning_rate": 4.726815474913417e-05,
307
+ "loss": 0.5686,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 0.46,
312
+ "learning_rate": 4.715926668209398e-05,
313
+ "loss": 0.5621,
314
+ "step": 510
315
+ },
316
+ {
317
+ "epoch": 0.47,
318
+ "learning_rate": 4.704838124592558e-05,
319
+ "loss": 0.558,
320
+ "step": 520
321
+ },
322
+ {
323
+ "epoch": 0.48,
324
+ "learning_rate": 4.693550843550607e-05,
325
+ "loss": 0.5746,
326
+ "step": 530
327
+ },
328
+ {
329
+ "epoch": 0.49,
330
+ "learning_rate": 4.682065842484843e-05,
331
+ "loss": 0.5191,
332
+ "step": 540
333
+ },
334
+ {
335
+ "epoch": 0.5,
336
+ "learning_rate": 4.670384156618448e-05,
337
+ "loss": 0.5243,
338
+ "step": 550
339
+ },
340
+ {
341
+ "epoch": 0.51,
342
+ "learning_rate": 4.658506838903176e-05,
343
+ "loss": 0.5996,
344
+ "step": 560
345
+ },
346
+ {
347
+ "epoch": 0.52,
348
+ "learning_rate": 4.6464349599244425e-05,
349
+ "loss": 0.5799,
350
+ "step": 570
351
+ },
352
+ {
353
+ "epoch": 0.53,
354
+ "learning_rate": 4.634169607804826e-05,
355
+ "loss": 0.5152,
356
+ "step": 580
357
+ },
358
+ {
359
+ "epoch": 0.53,
360
+ "learning_rate": 4.6217118881059864e-05,
361
+ "loss": 0.5641,
362
+ "step": 590
363
+ },
364
+ {
365
+ "epoch": 0.54,
366
+ "learning_rate": 4.609062923729014e-05,
367
+ "loss": 0.5278,
368
+ "step": 600
369
+ },
370
+ {
371
+ "epoch": 0.55,
372
+ "learning_rate": 4.596223854813219e-05,
373
+ "loss": 0.5245,
374
+ "step": 610
375
+ },
376
+ {
377
+ "epoch": 0.56,
378
+ "learning_rate": 4.5831958386333505e-05,
379
+ "loss": 0.503,
380
+ "step": 620
381
+ },
382
+ {
383
+ "epoch": 0.57,
384
+ "learning_rate": 4.569980049495298e-05,
385
+ "loss": 0.5509,
386
+ "step": 630
387
+ },
388
+ {
389
+ "epoch": 0.58,
390
+ "learning_rate": 4.5565776786302336e-05,
391
+ "loss": 0.5467,
392
+ "step": 640
393
+ },
394
+ {
395
+ "epoch": 0.59,
396
+ "learning_rate": 4.5429899340872405e-05,
397
+ "loss": 0.5637,
398
+ "step": 650
399
+ },
400
+ {
401
+ "epoch": 0.6,
402
+ "learning_rate": 4.529218040624425e-05,
403
+ "loss": 0.5836,
404
+ "step": 660
405
+ },
406
+ {
407
+ "epoch": 0.61,
408
+ "learning_rate": 4.5152632395985165e-05,
409
+ "loss": 0.5497,
410
+ "step": 670
411
+ },
412
+ {
413
+ "epoch": 0.62,
414
+ "learning_rate": 4.5011267888529786e-05,
415
+ "loss": 0.5182,
416
+ "step": 680
417
+ },
418
+ {
419
+ "epoch": 0.63,
420
+ "learning_rate": 4.486809962604631e-05,
421
+ "loss": 0.5251,
422
+ "step": 690
423
+ },
424
+ {
425
+ "epoch": 0.63,
426
+ "learning_rate": 4.472314051328794e-05,
427
+ "loss": 0.5453,
428
+ "step": 700
429
+ },
430
+ {
431
+ "epoch": 0.64,
432
+ "learning_rate": 4.457640361642968e-05,
433
+ "loss": 0.5442,
434
+ "step": 710
435
+ },
436
+ {
437
+ "epoch": 0.65,
438
+ "learning_rate": 4.442790216189062e-05,
439
+ "loss": 0.5223,
440
+ "step": 720
441
+ },
442
+ {
443
+ "epoch": 0.66,
444
+ "learning_rate": 4.4277649535141726e-05,
445
+ "loss": 0.5733,
446
+ "step": 730
447
+ },
448
+ {
449
+ "epoch": 0.67,
450
+ "learning_rate": 4.412565927949933e-05,
451
+ "loss": 0.5234,
452
+ "step": 740
453
+ },
454
+ {
455
+ "epoch": 0.68,
456
+ "learning_rate": 4.397194509490433e-05,
457
+ "loss": 0.5091,
458
+ "step": 750
459
+ },
460
+ {
461
+ "epoch": 0.69,
462
+ "learning_rate": 4.381652083668739e-05,
463
+ "loss": 0.506,
464
+ "step": 760
465
+ },
466
+ {
467
+ "epoch": 0.7,
468
+ "learning_rate": 4.3659400514320016e-05,
469
+ "loss": 0.5328,
470
+ "step": 770
471
+ },
472
+ {
473
+ "epoch": 0.71,
474
+ "learning_rate": 4.3500598290151796e-05,
475
+ "loss": 0.5134,
476
+ "step": 780
477
+ },
478
+ {
479
+ "epoch": 0.72,
480
+ "learning_rate": 4.3340128478133834e-05,
481
+ "loss": 0.5423,
482
+ "step": 790
483
+ },
484
+ {
485
+ "epoch": 0.73,
486
+ "learning_rate": 4.317800554252857e-05,
487
+ "loss": 0.5101,
488
+ "step": 800
489
+ },
490
+ {
491
+ "epoch": 0.73,
492
+ "learning_rate": 4.301424409660598e-05,
493
+ "loss": 0.5427,
494
+ "step": 810
495
+ },
496
+ {
497
+ "epoch": 0.74,
498
+ "learning_rate": 4.284885890132639e-05,
499
+ "loss": 0.526,
500
+ "step": 820
501
+ },
502
+ {
503
+ "epoch": 0.75,
504
+ "learning_rate": 4.2681864864009985e-05,
505
+ "loss": 0.5381,
506
+ "step": 830
507
+ },
508
+ {
509
+ "epoch": 0.76,
510
+ "learning_rate": 4.251327703699308e-05,
511
+ "loss": 0.5581,
512
+ "step": 840
513
+ },
514
+ {
515
+ "epoch": 0.77,
516
+ "learning_rate": 4.2343110616271376e-05,
517
+ "loss": 0.5464,
518
+ "step": 850
519
+ },
520
+ {
521
+ "epoch": 0.78,
522
+ "learning_rate": 4.217138094013021e-05,
523
+ "loss": 0.5458,
524
+ "step": 860
525
+ },
526
+ {
527
+ "epoch": 0.79,
528
+ "learning_rate": 4.1998103487762034e-05,
529
+ "loss": 0.5158,
530
+ "step": 870
531
+ },
532
+ {
533
+ "epoch": 0.8,
534
+ "learning_rate": 4.1823293877871164e-05,
535
+ "loss": 0.5544,
536
+ "step": 880
537
+ },
538
+ {
539
+ "epoch": 0.81,
540
+ "learning_rate": 4.164696786726593e-05,
541
+ "loss": 0.5348,
542
+ "step": 890
543
+ },
544
+ {
545
+ "epoch": 0.82,
546
+ "learning_rate": 4.146914134943843e-05,
547
+ "loss": 0.5161,
548
+ "step": 900
549
+ },
550
+ {
551
+ "epoch": 0.83,
552
+ "learning_rate": 4.1289830353131974e-05,
553
+ "loss": 0.5126,
554
+ "step": 910
555
+ },
556
+ {
557
+ "epoch": 0.83,
558
+ "learning_rate": 4.110905104089621e-05,
559
+ "loss": 0.5608,
560
+ "step": 920
561
+ },
562
+ {
563
+ "epoch": 0.84,
564
+ "learning_rate": 4.0926819707630346e-05,
565
+ "loss": 0.5306,
566
+ "step": 930
567
+ },
568
+ {
569
+ "epoch": 0.85,
570
+ "learning_rate": 4.074315277911438e-05,
571
+ "loss": 0.5399,
572
+ "step": 940
573
+ },
574
+ {
575
+ "epoch": 0.86,
576
+ "learning_rate": 4.055806681052852e-05,
577
+ "loss": 0.5085,
578
+ "step": 950
579
+ },
580
+ {
581
+ "epoch": 0.87,
582
+ "learning_rate": 4.0371578484960906e-05,
583
+ "loss": 0.528,
584
+ "step": 960
585
+ },
586
+ {
587
+ "epoch": 0.88,
588
+ "learning_rate": 4.018370461190392e-05,
589
+ "loss": 0.5315,
590
+ "step": 970
591
+ },
592
+ {
593
+ "epoch": 0.89,
594
+ "learning_rate": 3.9994462125739e-05,
595
+ "loss": 0.5357,
596
+ "step": 980
597
+ },
598
+ {
599
+ "epoch": 0.9,
600
+ "learning_rate": 3.980386808421018e-05,
601
+ "loss": 0.5262,
602
+ "step": 990
603
+ },
604
+ {
605
+ "epoch": 0.91,
606
+ "learning_rate": 3.961193966688664e-05,
607
+ "loss": 0.5546,
608
+ "step": 1000
609
+ },
610
+ {
611
+ "epoch": 0.92,
612
+ "learning_rate": 3.941869417361413e-05,
613
+ "loss": 0.5125,
614
+ "step": 1010
615
+ },
616
+ {
617
+ "epoch": 0.92,
618
+ "learning_rate": 3.922414902295562e-05,
619
+ "loss": 0.5429,
620
+ "step": 1020
621
+ },
622
+ {
623
+ "epoch": 0.93,
624
+ "learning_rate": 3.90283217506213e-05,
625
+ "loss": 0.5559,
626
+ "step": 1030
627
+ },
628
+ {
629
+ "epoch": 0.94,
630
+ "learning_rate": 3.883123000788787e-05,
631
+ "loss": 0.5147,
632
+ "step": 1040
633
+ },
634
+ {
635
+ "epoch": 0.95,
636
+ "learning_rate": 3.86328915600076e-05,
637
+ "loss": 0.5537,
638
+ "step": 1050
639
+ },
640
+ {
641
+ "epoch": 0.96,
642
+ "learning_rate": 3.843332428460696e-05,
643
+ "loss": 0.4949,
644
+ "step": 1060
645
+ },
646
+ {
647
+ "epoch": 0.97,
648
+ "learning_rate": 3.823254617007522e-05,
649
+ "loss": 0.514,
650
+ "step": 1070
651
+ },
652
+ {
653
+ "epoch": 0.98,
654
+ "learning_rate": 3.803057531394302e-05,
655
+ "loss": 0.5361,
656
+ "step": 1080
657
+ },
658
+ {
659
+ "epoch": 0.99,
660
+ "learning_rate": 3.782742992125108e-05,
661
+ "loss": 0.5687,
662
+ "step": 1090
663
+ },
664
+ {
665
+ "epoch": 1.0,
666
+ "learning_rate": 3.762312830290932e-05,
667
+ "loss": 0.5287,
668
+ "step": 1100
669
+ },
670
+ {
671
+ "epoch": 1.01,
672
+ "learning_rate": 3.741768887404634e-05,
673
+ "loss": 0.4718,
674
+ "step": 1110
675
+ },
676
+ {
677
+ "epoch": 1.02,
678
+ "learning_rate": 3.7211130152349496e-05,
679
+ "loss": 0.5207,
680
+ "step": 1120
681
+ },
682
+ {
683
+ "epoch": 1.02,
684
+ "learning_rate": 3.7003470756395845e-05,
685
+ "loss": 0.4424,
686
+ "step": 1130
687
+ },
688
+ {
689
+ "epoch": 1.03,
690
+ "learning_rate": 3.679472940397387e-05,
691
+ "loss": 0.5109,
692
+ "step": 1140
693
+ },
694
+ {
695
+ "epoch": 1.04,
696
+ "learning_rate": 3.658492491039631e-05,
697
+ "loss": 0.4586,
698
+ "step": 1150
699
+ },
700
+ {
701
+ "epoch": 1.05,
702
+ "learning_rate": 3.6374076186804276e-05,
703
+ "loss": 0.4981,
704
+ "step": 1160
705
+ },
706
+ {
707
+ "epoch": 1.06,
708
+ "learning_rate": 3.6162202238462546e-05,
709
+ "loss": 0.4991,
710
+ "step": 1170
711
+ },
712
+ {
713
+ "epoch": 1.07,
714
+ "learning_rate": 3.594932216304658e-05,
715
+ "loss": 0.5041,
716
+ "step": 1180
717
+ },
718
+ {
719
+ "epoch": 1.08,
720
+ "learning_rate": 3.573545514892109e-05,
721
+ "loss": 0.4927,
722
+ "step": 1190
723
+ },
724
+ {
725
+ "epoch": 1.09,
726
+ "learning_rate": 3.552062047341045e-05,
727
+ "loss": 0.4954,
728
+ "step": 1200
729
+ },
730
+ {
731
+ "epoch": 1.1,
732
+ "learning_rate": 3.5304837501061075e-05,
733
+ "loss": 0.4924,
734
+ "step": 1210
735
+ },
736
+ {
737
+ "epoch": 1.11,
738
+ "learning_rate": 3.5088125681896e-05,
739
+ "loss": 0.5002,
740
+ "step": 1220
741
+ },
742
+ {
743
+ "epoch": 1.12,
744
+ "learning_rate": 3.487050454966171e-05,
745
+ "loss": 0.5133,
746
+ "step": 1230
747
+ },
748
+ {
749
+ "epoch": 1.12,
750
+ "learning_rate": 3.465199372006738e-05,
751
+ "loss": 0.505,
752
+ "step": 1240
753
+ },
754
+ {
755
+ "epoch": 1.13,
756
+ "learning_rate": 3.443261288901683e-05,
757
+ "loss": 0.5264,
758
+ "step": 1250
759
+ },
760
+ {
761
+ "epoch": 1.14,
762
+ "learning_rate": 3.4212381830833184e-05,
763
+ "loss": 0.5201,
764
+ "step": 1260
765
+ },
766
+ {
767
+ "epoch": 1.15,
768
+ "learning_rate": 3.399132039647644e-05,
769
+ "loss": 0.4936,
770
+ "step": 1270
771
+ },
772
+ {
773
+ "epoch": 1.16,
774
+ "learning_rate": 3.3769448511754186e-05,
775
+ "loss": 0.4969,
776
+ "step": 1280
777
+ },
778
+ {
779
+ "epoch": 1.17,
780
+ "learning_rate": 3.3546786175525566e-05,
781
+ "loss": 0.4781,
782
+ "step": 1290
783
+ },
784
+ {
785
+ "epoch": 1.18,
786
+ "learning_rate": 3.332335345789863e-05,
787
+ "loss": 0.5155,
788
+ "step": 1300
789
+ },
790
+ {
791
+ "epoch": 1.19,
792
+ "learning_rate": 3.309917049842126e-05,
793
+ "loss": 0.5064,
794
+ "step": 1310
795
+ },
796
+ {
797
+ "epoch": 1.2,
798
+ "learning_rate": 3.287425750426585e-05,
799
+ "loss": 0.4699,
800
+ "step": 1320
801
+ },
802
+ {
803
+ "epoch": 1.21,
804
+ "learning_rate": 3.2648634748407945e-05,
805
+ "loss": 0.5182,
806
+ "step": 1330
807
+ },
808
+ {
809
+ "epoch": 1.21,
810
+ "learning_rate": 3.2422322567798815e-05,
811
+ "loss": 0.4881,
812
+ "step": 1340
813
+ },
814
+ {
815
+ "epoch": 1.22,
816
+ "learning_rate": 3.2195341361532424e-05,
817
+ "loss": 0.4931,
818
+ "step": 1350
819
+ },
820
+ {
821
+ "epoch": 1.23,
822
+ "learning_rate": 3.196771158900666e-05,
823
+ "loss": 0.524,
824
+ "step": 1360
825
+ },
826
+ {
827
+ "epoch": 1.24,
828
+ "learning_rate": 3.173945376807921e-05,
829
+ "loss": 0.5238,
830
+ "step": 1370
831
+ },
832
+ {
833
+ "epoch": 1.25,
834
+ "learning_rate": 3.151058847321815e-05,
835
+ "loss": 0.4506,
836
+ "step": 1380
837
+ },
838
+ {
839
+ "epoch": 1.26,
840
+ "learning_rate": 3.1281136333647416e-05,
841
+ "loss": 0.4688,
842
+ "step": 1390
843
+ },
844
+ {
845
+ "epoch": 1.27,
846
+ "learning_rate": 3.1051118031487326e-05,
847
+ "loss": 0.4495,
848
+ "step": 1400
849
+ },
850
+ {
851
+ "epoch": 1.28,
852
+ "learning_rate": 3.0820554299890384e-05,
853
+ "loss": 0.4999,
854
+ "step": 1410
855
+ },
856
+ {
857
+ "epoch": 1.29,
858
+ "learning_rate": 3.058946592117248e-05,
859
+ "loss": 0.513,
860
+ "step": 1420
861
+ },
862
+ {
863
+ "epoch": 1.3,
864
+ "learning_rate": 3.035787372493956e-05,
865
+ "loss": 0.4841,
866
+ "step": 1430
867
+ },
868
+ {
869
+ "epoch": 1.31,
870
+ "learning_rate": 3.012579858621019e-05,
871
+ "loss": 0.4783,
872
+ "step": 1440
873
+ },
874
+ {
875
+ "epoch": 1.31,
876
+ "learning_rate": 2.9893261423533893e-05,
877
+ "loss": 0.4896,
878
+ "step": 1450
879
+ },
880
+ {
881
+ "epoch": 1.32,
882
+ "learning_rate": 2.9660283197105632e-05,
883
+ "loss": 0.4491,
884
+ "step": 1460
885
+ },
886
+ {
887
+ "epoch": 1.33,
888
+ "learning_rate": 2.9426884906876517e-05,
889
+ "loss": 0.5361,
890
+ "step": 1470
891
+ },
892
+ {
893
+ "epoch": 1.34,
894
+ "learning_rate": 2.9193087590660944e-05,
895
+ "loss": 0.4871,
896
+ "step": 1480
897
+ },
898
+ {
899
+ "epoch": 1.35,
900
+ "learning_rate": 2.8958912322240282e-05,
901
+ "loss": 0.4915,
902
+ "step": 1490
903
+ },
904
+ {
905
+ "epoch": 1.36,
906
+ "learning_rate": 2.87243802094634e-05,
907
+ "loss": 0.4772,
908
+ "step": 1500
909
+ },
910
+ {
911
+ "epoch": 1.37,
912
+ "learning_rate": 2.8489512392343993e-05,
913
+ "loss": 0.4973,
914
+ "step": 1510
915
+ },
916
+ {
917
+ "epoch": 1.38,
918
+ "learning_rate": 2.8254330041155163e-05,
919
+ "loss": 0.4987,
920
+ "step": 1520
921
+ },
922
+ {
923
+ "epoch": 1.39,
924
+ "learning_rate": 2.8018854354521152e-05,
925
+ "loss": 0.489,
926
+ "step": 1530
927
+ },
928
+ {
929
+ "epoch": 1.4,
930
+ "learning_rate": 2.778310655750656e-05,
931
+ "loss": 0.4624,
932
+ "step": 1540
933
+ },
934
+ {
935
+ "epoch": 1.41,
936
+ "learning_rate": 2.7547107899703196e-05,
937
+ "loss": 0.5009,
938
+ "step": 1550
939
+ },
940
+ {
941
+ "epoch": 1.41,
942
+ "learning_rate": 2.7310879653314702e-05,
943
+ "loss": 0.5004,
944
+ "step": 1560
945
+ },
946
+ {
947
+ "epoch": 1.42,
948
+ "learning_rate": 2.7074443111239133e-05,
949
+ "loss": 0.4825,
950
+ "step": 1570
951
+ },
952
+ {
953
+ "epoch": 1.43,
954
+ "learning_rate": 2.6837819585149697e-05,
955
+ "loss": 0.4262,
956
+ "step": 1580
957
+ },
958
+ {
959
+ "epoch": 1.44,
960
+ "learning_rate": 2.660103040357376e-05,
961
+ "loss": 0.4871,
962
+ "step": 1590
963
+ },
964
+ {
965
+ "epoch": 1.45,
966
+ "learning_rate": 2.636409690997039e-05,
967
+ "loss": 0.4874,
968
+ "step": 1600
969
+ },
970
+ {
971
+ "epoch": 1.46,
972
+ "learning_rate": 2.6127040460806495e-05,
973
+ "loss": 0.49,
974
+ "step": 1610
975
+ },
976
+ {
977
+ "epoch": 1.47,
978
+ "learning_rate": 2.5889882423631817e-05,
979
+ "loss": 0.5357,
980
+ "step": 1620
981
+ },
982
+ {
983
+ "epoch": 1.48,
984
+ "learning_rate": 2.5652644175152928e-05,
985
+ "loss": 0.499,
986
+ "step": 1630
987
+ },
988
+ {
989
+ "epoch": 1.49,
990
+ "learning_rate": 2.5415347099306436e-05,
991
+ "loss": 0.5093,
992
+ "step": 1640
993
+ },
994
+ {
995
+ "epoch": 1.5,
996
+ "learning_rate": 2.5178012585331433e-05,
997
+ "loss": 0.4934,
998
+ "step": 1650
999
+ },
1000
+ {
1001
+ "epoch": 1.5,
1002
+ "learning_rate": 2.4940662025841592e-05,
1003
+ "loss": 0.5063,
1004
+ "step": 1660
1005
+ },
1006
+ {
1007
+ "epoch": 1.51,
1008
+ "learning_rate": 2.470331681489686e-05,
1009
+ "loss": 0.4538,
1010
+ "step": 1670
1011
+ },
1012
+ {
1013
+ "epoch": 1.52,
1014
+ "learning_rate": 2.4465998346075128e-05,
1015
+ "loss": 0.4856,
1016
+ "step": 1680
1017
+ },
1018
+ {
1019
+ "epoch": 1.53,
1020
+ "learning_rate": 2.422872801054378e-05,
1021
+ "loss": 0.4774,
1022
+ "step": 1690
1023
+ },
1024
+ {
1025
+ "epoch": 1.54,
1026
+ "learning_rate": 2.3991527195131627e-05,
1027
+ "loss": 0.537,
1028
+ "step": 1700
1029
+ },
1030
+ {
1031
+ "epoch": 1.55,
1032
+ "learning_rate": 2.375441728040117e-05,
1033
+ "loss": 0.4859,
1034
+ "step": 1710
1035
+ },
1036
+ {
1037
+ "epoch": 1.56,
1038
+ "learning_rate": 2.3517419638721357e-05,
1039
+ "loss": 0.5388,
1040
+ "step": 1720
1041
+ },
1042
+ {
1043
+ "epoch": 1.57,
1044
+ "learning_rate": 2.3280555632341238e-05,
1045
+ "loss": 0.5142,
1046
+ "step": 1730
1047
+ },
1048
+ {
1049
+ "epoch": 1.58,
1050
+ "learning_rate": 2.3043846611464316e-05,
1051
+ "loss": 0.4888,
1052
+ "step": 1740
1053
+ },
1054
+ {
1055
+ "epoch": 1.59,
1056
+ "learning_rate": 2.2807313912324242e-05,
1057
+ "loss": 0.4642,
1058
+ "step": 1750
1059
+ },
1060
+ {
1061
+ "epoch": 1.6,
1062
+ "learning_rate": 2.2570978855261507e-05,
1063
+ "loss": 0.5062,
1064
+ "step": 1760
1065
+ },
1066
+ {
1067
+ "epoch": 1.6,
1068
+ "learning_rate": 2.2334862742801756e-05,
1069
+ "loss": 0.4726,
1070
+ "step": 1770
1071
+ },
1072
+ {
1073
+ "epoch": 1.61,
1074
+ "learning_rate": 2.2098986857735644e-05,
1075
+ "loss": 0.53,
1076
+ "step": 1780
1077
+ },
1078
+ {
1079
+ "epoch": 1.62,
1080
+ "learning_rate": 2.186337246120045e-05,
1081
+ "loss": 0.5003,
1082
+ "step": 1790
1083
+ },
1084
+ {
1085
+ "epoch": 1.63,
1086
+ "learning_rate": 2.1628040790763667e-05,
1087
+ "loss": 0.4687,
1088
+ "step": 1800
1089
+ },
1090
+ {
1091
+ "epoch": 1.64,
1092
+ "learning_rate": 2.1393013058508708e-05,
1093
+ "loss": 0.4609,
1094
+ "step": 1810
1095
+ },
1096
+ {
1097
+ "epoch": 1.65,
1098
+ "learning_rate": 2.1158310449122948e-05,
1099
+ "loss": 0.5066,
1100
+ "step": 1820
1101
+ },
1102
+ {
1103
+ "epoch": 1.66,
1104
+ "learning_rate": 2.092395411798813e-05,
1105
+ "loss": 0.4846,
1106
+ "step": 1830
1107
+ },
1108
+ {
1109
+ "epoch": 1.67,
1110
+ "learning_rate": 2.0689965189273572e-05,
1111
+ "loss": 0.5068,
1112
+ "step": 1840
1113
+ },
1114
+ {
1115
+ "epoch": 1.68,
1116
+ "learning_rate": 2.045636475403201e-05,
1117
+ "loss": 0.4671,
1118
+ "step": 1850
1119
+ },
1120
+ {
1121
+ "epoch": 1.69,
1122
+ "learning_rate": 2.022317386829857e-05,
1123
+ "loss": 0.4876,
1124
+ "step": 1860
1125
+ },
1126
+ {
1127
+ "epoch": 1.7,
1128
+ "learning_rate": 1.9990413551192843e-05,
1129
+ "loss": 0.501,
1130
+ "step": 1870
1131
+ },
1132
+ {
1133
+ "epoch": 1.7,
1134
+ "learning_rate": 1.9758104783024237e-05,
1135
+ "loss": 0.507,
1136
+ "step": 1880
1137
+ },
1138
+ {
1139
+ "epoch": 1.71,
1140
+ "learning_rate": 1.9526268503400946e-05,
1141
+ "loss": 0.4716,
1142
+ "step": 1890
1143
+ },
1144
+ {
1145
+ "epoch": 1.72,
1146
+ "learning_rate": 1.9294925609342443e-05,
1147
+ "loss": 0.4714,
1148
+ "step": 1900
1149
+ },
1150
+ {
1151
+ "epoch": 1.73,
1152
+ "learning_rate": 1.9064096953395954e-05,
1153
+ "loss": 0.5107,
1154
+ "step": 1910
1155
+ },
1156
+ {
1157
+ "epoch": 1.74,
1158
+ "learning_rate": 1.8833803341756816e-05,
1159
+ "loss": 0.4775,
1160
+ "step": 1920
1161
+ },
1162
+ {
1163
+ "epoch": 1.75,
1164
+ "learning_rate": 1.8604065532393094e-05,
1165
+ "loss": 0.4835,
1166
+ "step": 1930
1167
+ },
1168
+ {
1169
+ "epoch": 1.76,
1170
+ "learning_rate": 1.8374904233174543e-05,
1171
+ "loss": 0.5042,
1172
+ "step": 1940
1173
+ },
1174
+ {
1175
+ "epoch": 1.77,
1176
+ "learning_rate": 1.8146340100006014e-05,
1177
+ "loss": 0.452,
1178
+ "step": 1950
1179
+ },
1180
+ {
1181
+ "epoch": 1.78,
1182
+ "learning_rate": 1.7918393734965645e-05,
1183
+ "loss": 0.5025,
1184
+ "step": 1960
1185
+ },
1186
+ {
1187
+ "epoch": 1.79,
1188
+ "learning_rate": 1.7691085684447802e-05,
1189
+ "loss": 0.4923,
1190
+ "step": 1970
1191
+ },
1192
+ {
1193
+ "epoch": 1.8,
1194
+ "learning_rate": 1.746443643731115e-05,
1195
+ "loss": 0.4653,
1196
+ "step": 1980
1197
+ },
1198
+ {
1199
+ "epoch": 1.8,
1200
+ "learning_rate": 1.723846642303177e-05,
1201
+ "loss": 0.4577,
1202
+ "step": 1990
1203
+ },
1204
+ {
1205
+ "epoch": 1.81,
1206
+ "learning_rate": 1.701319600986179e-05,
1207
+ "loss": 0.5044,
1208
+ "step": 2000
1209
+ },
1210
+ {
1211
+ "epoch": 1.82,
1212
+ "learning_rate": 1.6788645502993396e-05,
1213
+ "loss": 0.5008,
1214
+ "step": 2010
1215
+ },
1216
+ {
1217
+ "epoch": 1.83,
1218
+ "learning_rate": 1.65648351427286e-05,
1219
+ "loss": 0.4531,
1220
+ "step": 2020
1221
+ },
1222
+ {
1223
+ "epoch": 1.84,
1224
+ "learning_rate": 1.634178510265486e-05,
1225
+ "loss": 0.4518,
1226
+ "step": 2030
1227
+ },
1228
+ {
1229
+ "epoch": 1.85,
1230
+ "learning_rate": 1.611951548782665e-05,
1231
+ "loss": 0.4898,
1232
+ "step": 2040
1233
+ },
1234
+ {
1235
+ "epoch": 1.86,
1236
+ "learning_rate": 1.589804633295331e-05,
1237
+ "loss": 0.5001,
1238
+ "step": 2050
1239
+ },
1240
+ {
1241
+ "epoch": 1.87,
1242
+ "learning_rate": 1.5677397600593108e-05,
1243
+ "loss": 0.4747,
1244
+ "step": 2060
1245
+ },
1246
+ {
1247
+ "epoch": 1.88,
1248
+ "learning_rate": 1.5457589179353964e-05,
1249
+ "loss": 0.4872,
1250
+ "step": 2070
1251
+ },
1252
+ {
1253
+ "epoch": 1.89,
1254
+ "learning_rate": 1.5238640882100658e-05,
1255
+ "loss": 0.4893,
1256
+ "step": 2080
1257
+ },
1258
+ {
1259
+ "epoch": 1.89,
1260
+ "learning_rate": 1.5020572444169007e-05,
1261
+ "loss": 0.4987,
1262
+ "step": 2090
1263
+ },
1264
+ {
1265
+ "epoch": 1.9,
1266
+ "learning_rate": 1.4803403521586995e-05,
1267
+ "loss": 0.4679,
1268
+ "step": 2100
1269
+ },
1270
+ {
1271
+ "epoch": 1.91,
1272
+ "learning_rate": 1.4587153689302993e-05,
1273
+ "loss": 0.4659,
1274
+ "step": 2110
1275
+ },
1276
+ {
1277
+ "epoch": 1.92,
1278
+ "learning_rate": 1.4371842439421396e-05,
1279
+ "loss": 0.4961,
1280
+ "step": 2120
1281
+ },
1282
+ {
1283
+ "epoch": 1.93,
1284
+ "learning_rate": 1.4157489179445583e-05,
1285
+ "loss": 0.4935,
1286
+ "step": 2130
1287
+ },
1288
+ {
1289
+ "epoch": 1.94,
1290
+ "learning_rate": 1.3944113230528661e-05,
1291
+ "loss": 0.4561,
1292
+ "step": 2140
1293
+ },
1294
+ {
1295
+ "epoch": 1.95,
1296
+ "learning_rate": 1.3731733825731857e-05,
1297
+ "loss": 0.4755,
1298
+ "step": 2150
1299
+ },
1300
+ {
1301
+ "epoch": 1.96,
1302
+ "learning_rate": 1.3520370108290953e-05,
1303
+ "loss": 0.4951,
1304
+ "step": 2160
1305
+ },
1306
+ {
1307
+ "epoch": 1.97,
1308
+ "learning_rate": 1.3310041129890737e-05,
1309
+ "loss": 0.475,
1310
+ "step": 2170
1311
+ },
1312
+ {
1313
+ "epoch": 1.98,
1314
+ "learning_rate": 1.3100765848947753e-05,
1315
+ "loss": 0.5023,
1316
+ "step": 2180
1317
+ },
1318
+ {
1319
+ "epoch": 1.99,
1320
+ "learning_rate": 1.2892563128901468e-05,
1321
+ "loss": 0.4529,
1322
+ "step": 2190
1323
+ },
1324
+ {
1325
+ "epoch": 1.99,
1326
+ "learning_rate": 1.2685451736513942e-05,
1327
+ "loss": 0.4688,
1328
+ "step": 2200
1329
+ },
1330
+ {
1331
+ "epoch": 2.0,
1332
+ "learning_rate": 1.2479450340178284e-05,
1333
+ "loss": 0.4713,
1334
+ "step": 2210
1335
+ },
1336
+ {
1337
+ "epoch": 2.01,
1338
+ "learning_rate": 1.2274577508235899e-05,
1339
+ "loss": 0.4398,
1340
+ "step": 2220
1341
+ },
1342
+ {
1343
+ "epoch": 2.02,
1344
+ "learning_rate": 1.2070851707302854e-05,
1345
+ "loss": 0.4439,
1346
+ "step": 2230
1347
+ },
1348
+ {
1349
+ "epoch": 2.03,
1350
+ "learning_rate": 1.1868291300605295e-05,
1351
+ "loss": 0.469,
1352
+ "step": 2240
1353
+ },
1354
+ {
1355
+ "epoch": 2.04,
1356
+ "learning_rate": 1.1666914546324254e-05,
1357
+ "loss": 0.4142,
1358
+ "step": 2250
1359
+ },
1360
+ {
1361
+ "epoch": 2.05,
1362
+ "learning_rate": 1.1466739595949954e-05,
1363
+ "loss": 0.4307,
1364
+ "step": 2260
1365
+ },
1366
+ {
1367
+ "epoch": 2.06,
1368
+ "learning_rate": 1.1267784492645628e-05,
1369
+ "loss": 0.472,
1370
+ "step": 2270
1371
+ },
1372
+ {
1373
+ "epoch": 2.07,
1374
+ "learning_rate": 1.1070067169621215e-05,
1375
+ "loss": 0.4416,
1376
+ "step": 2280
1377
+ },
1378
+ {
1379
+ "epoch": 2.08,
1380
+ "learning_rate": 1.0873605448516864e-05,
1381
+ "loss": 0.4922,
1382
+ "step": 2290
1383
+ },
1384
+ {
1385
+ "epoch": 2.09,
1386
+ "learning_rate": 1.0678417037796607e-05,
1387
+ "loss": 0.4398,
1388
+ "step": 2300
1389
+ },
1390
+ {
1391
+ "epoch": 2.09,
1392
+ "learning_rate": 1.0484519531152099e-05,
1393
+ "loss": 0.4936,
1394
+ "step": 2310
1395
+ },
1396
+ {
1397
+ "epoch": 2.1,
1398
+ "learning_rate": 1.0291930405916853e-05,
1399
+ "loss": 0.4647,
1400
+ "step": 2320
1401
+ },
1402
+ {
1403
+ "epoch": 2.11,
1404
+ "learning_rate": 1.0100667021490823e-05,
1405
+ "loss": 0.4681,
1406
+ "step": 2330
1407
+ },
1408
+ {
1409
+ "epoch": 2.12,
1410
+ "learning_rate": 9.910746617775703e-06,
1411
+ "loss": 0.4836,
1412
+ "step": 2340
1413
+ },
1414
+ {
1415
+ "epoch": 2.13,
1416
+ "learning_rate": 9.722186313621004e-06,
1417
+ "loss": 0.4718,
1418
+ "step": 2350
1419
+ },
1420
+ {
1421
+ "epoch": 2.14,
1422
+ "learning_rate": 9.535003105280955e-06,
1423
+ "loss": 0.4677,
1424
+ "step": 2360
1425
+ },
1426
+ {
1427
+ "epoch": 2.15,
1428
+ "learning_rate": 9.349213864882556e-06,
1429
+ "loss": 0.4573,
1430
+ "step": 2370
1431
+ },
1432
+ {
1433
+ "epoch": 2.16,
1434
+ "learning_rate": 9.164835338904753e-06,
1435
+ "loss": 0.4483,
1436
+ "step": 2380
1437
+ },
1438
+ {
1439
+ "epoch": 2.17,
1440
+ "learning_rate": 8.981884146668995e-06,
1441
+ "loss": 0.4143,
1442
+ "step": 2390
1443
+ },
1444
+ {
1445
+ "epoch": 2.18,
1446
+ "learning_rate": 8.800376778841171e-06,
1447
+ "loss": 0.4634,
1448
+ "step": 2400
1449
+ },
1450
+ {
1451
+ "epoch": 2.18,
1452
+ "learning_rate": 8.62032959594521e-06,
1453
+ "loss": 0.4403,
1454
+ "step": 2410
1455
+ },
1456
+ {
1457
+ "epoch": 2.19,
1458
+ "learning_rate": 8.441758826888421e-06,
1459
+ "loss": 0.4536,
1460
+ "step": 2420
1461
+ },
1462
+ {
1463
+ "epoch": 2.2,
1464
+ "learning_rate": 8.264680567498623e-06,
1465
+ "loss": 0.4579,
1466
+ "step": 2430
1467
+ },
1468
+ {
1469
+ "epoch": 2.21,
1470
+ "learning_rate": 8.089110779073362e-06,
1471
+ "loss": 0.4465,
1472
+ "step": 2440
1473
+ },
1474
+ {
1475
+ "epoch": 2.22,
1476
+ "learning_rate": 7.915065286941148e-06,
1477
+ "loss": 0.4767,
1478
+ "step": 2450
1479
+ },
1480
+ {
1481
+ "epoch": 2.23,
1482
+ "learning_rate": 7.742559779035086e-06,
1483
+ "loss": 0.4905,
1484
+ "step": 2460
1485
+ },
1486
+ {
1487
+ "epoch": 2.24,
1488
+ "learning_rate": 7.5716098044787295e-06,
1489
+ "loss": 0.4821,
1490
+ "step": 2470
1491
+ },
1492
+ {
1493
+ "epoch": 2.25,
1494
+ "learning_rate": 7.402230772184609e-06,
1495
+ "loss": 0.4678,
1496
+ "step": 2480
1497
+ },
1498
+ {
1499
+ "epoch": 2.26,
1500
+ "learning_rate": 7.234437949465256e-06,
1501
+ "loss": 0.4078,
1502
+ "step": 2490
1503
+ },
1504
+ {
1505
+ "epoch": 2.27,
1506
+ "learning_rate": 7.068246460657082e-06,
1507
+ "loss": 0.4464,
1508
+ "step": 2500
1509
+ },
1510
+ {
1511
+ "epoch": 2.28,
1512
+ "learning_rate": 6.903671285757118e-06,
1513
+ "loss": 0.5072,
1514
+ "step": 2510
1515
+ },
1516
+ {
1517
+ "epoch": 2.28,
1518
+ "learning_rate": 6.740727259072768e-06,
1519
+ "loss": 0.4362,
1520
+ "step": 2520
1521
+ },
1522
+ {
1523
+ "epoch": 2.29,
1524
+ "learning_rate": 6.579429067884671e-06,
1525
+ "loss": 0.4391,
1526
+ "step": 2530
1527
+ },
1528
+ {
1529
+ "epoch": 2.3,
1530
+ "learning_rate": 6.419791251122828e-06,
1531
+ "loss": 0.4321,
1532
+ "step": 2540
1533
+ },
1534
+ {
1535
+ "epoch": 2.31,
1536
+ "learning_rate": 6.261828198056157e-06,
1537
+ "loss": 0.466,
1538
+ "step": 2550
1539
+ },
1540
+ {
1541
+ "epoch": 2.32,
1542
+ "learning_rate": 6.105554146995429e-06,
1543
+ "loss": 0.471,
1544
+ "step": 2560
1545
+ },
1546
+ {
1547
+ "epoch": 2.33,
1548
+ "learning_rate": 5.950983184009889e-06,
1549
+ "loss": 0.449,
1550
+ "step": 2570
1551
+ },
1552
+ {
1553
+ "epoch": 2.34,
1554
+ "learning_rate": 5.79812924165761e-06,
1555
+ "loss": 0.4913,
1556
+ "step": 2580
1557
+ },
1558
+ {
1559
+ "epoch": 2.35,
1560
+ "learning_rate": 5.647006097729604e-06,
1561
+ "loss": 0.4324,
1562
+ "step": 2590
1563
+ },
1564
+ {
1565
+ "epoch": 2.36,
1566
+ "learning_rate": 5.49762737400798e-06,
1567
+ "loss": 0.4746,
1568
+ "step": 2600
1569
+ },
1570
+ {
1571
+ "epoch": 2.37,
1572
+ "learning_rate": 5.350006535038079e-06,
1573
+ "loss": 0.4356,
1574
+ "step": 2610
1575
+ },
1576
+ {
1577
+ "epoch": 2.38,
1578
+ "learning_rate": 5.204156886914863e-06,
1579
+ "loss": 0.4221,
1580
+ "step": 2620
1581
+ },
1582
+ {
1583
+ "epoch": 2.38,
1584
+ "learning_rate": 5.060091576083487e-06,
1585
+ "loss": 0.4374,
1586
+ "step": 2630
1587
+ },
1588
+ {
1589
+ "epoch": 2.39,
1590
+ "learning_rate": 4.9178235881543855e-06,
1591
+ "loss": 0.426,
1592
+ "step": 2640
1593
+ },
1594
+ {
1595
+ "epoch": 2.4,
1596
+ "learning_rate": 4.7773657467327346e-06,
1597
+ "loss": 0.4299,
1598
+ "step": 2650
1599
+ },
1600
+ {
1601
+ "epoch": 2.41,
1602
+ "learning_rate": 4.638730712262593e-06,
1603
+ "loss": 0.4551,
1604
+ "step": 2660
1605
+ },
1606
+ {
1607
+ "epoch": 2.42,
1608
+ "learning_rate": 4.50193098088573e-06,
1609
+ "loss": 0.4554,
1610
+ "step": 2670
1611
+ },
1612
+ {
1613
+ "epoch": 2.43,
1614
+ "learning_rate": 4.366978883315248e-06,
1615
+ "loss": 0.4505,
1616
+ "step": 2680
1617
+ },
1618
+ {
1619
+ "epoch": 2.44,
1620
+ "learning_rate": 4.233886583724162e-06,
1621
+ "loss": 0.4837,
1622
+ "step": 2690
1623
+ },
1624
+ {
1625
+ "epoch": 2.45,
1626
+ "learning_rate": 4.1026660786489065e-06,
1627
+ "loss": 0.4772,
1628
+ "step": 2700
1629
+ },
1630
+ {
1631
+ "epoch": 2.46,
1632
+ "learning_rate": 3.97332919590806e-06,
1633
+ "loss": 0.4122,
1634
+ "step": 2710
1635
+ },
1636
+ {
1637
+ "epoch": 2.47,
1638
+ "learning_rate": 3.8458875935361686e-06,
1639
+ "loss": 0.4439,
1640
+ "step": 2720
1641
+ },
1642
+ {
1643
+ "epoch": 2.48,
1644
+ "learning_rate": 3.720352758732956e-06,
1645
+ "loss": 0.4806,
1646
+ "step": 2730
1647
+ },
1648
+ {
1649
+ "epoch": 2.48,
1650
+ "learning_rate": 3.5967360068279164e-06,
1651
+ "loss": 0.4608,
1652
+ "step": 2740
1653
+ },
1654
+ {
1655
+ "epoch": 2.49,
1656
+ "learning_rate": 3.47504848026034e-06,
1657
+ "loss": 0.4809,
1658
+ "step": 2750
1659
+ },
1660
+ {
1661
+ "epoch": 2.5,
1662
+ "learning_rate": 3.355301147575016e-06,
1663
+ "loss": 0.4484,
1664
+ "step": 2760
1665
+ },
1666
+ {
1667
+ "epoch": 2.51,
1668
+ "learning_rate": 3.2375048024335163e-06,
1669
+ "loss": 0.4982,
1670
+ "step": 2770
1671
+ },
1672
+ {
1673
+ "epoch": 2.52,
1674
+ "learning_rate": 3.1216700626413324e-06,
1675
+ "loss": 0.4707,
1676
+ "step": 2780
1677
+ },
1678
+ {
1679
+ "epoch": 2.53,
1680
+ "learning_rate": 3.007807369190768e-06,
1681
+ "loss": 0.4655,
1682
+ "step": 2790
1683
+ },
1684
+ {
1685
+ "epoch": 2.54,
1686
+ "learning_rate": 2.89592698531988e-06,
1687
+ "loss": 0.4566,
1688
+ "step": 2800
1689
+ },
1690
+ {
1691
+ "epoch": 2.55,
1692
+ "learning_rate": 2.7860389955873256e-06,
1693
+ "loss": 0.4153,
1694
+ "step": 2810
1695
+ },
1696
+ {
1697
+ "epoch": 2.56,
1698
+ "learning_rate": 2.6781533049633963e-06,
1699
+ "loss": 0.4386,
1700
+ "step": 2820
1701
+ },
1702
+ {
1703
+ "epoch": 2.57,
1704
+ "learning_rate": 2.5722796379372196e-06,
1705
+ "loss": 0.4663,
1706
+ "step": 2830
1707
+ },
1708
+ {
1709
+ "epoch": 2.57,
1710
+ "learning_rate": 2.4684275376402025e-06,
1711
+ "loss": 0.4283,
1712
+ "step": 2840
1713
+ },
1714
+ {
1715
+ "epoch": 2.58,
1716
+ "learning_rate": 2.3666063649858736e-06,
1717
+ "loss": 0.4853,
1718
+ "step": 2850
1719
+ },
1720
+ {
1721
+ "epoch": 2.59,
1722
+ "learning_rate": 2.2668252978260835e-06,
1723
+ "loss": 0.4587,
1724
+ "step": 2860
1725
+ },
1726
+ {
1727
+ "epoch": 2.6,
1728
+ "learning_rate": 2.169093330123767e-06,
1729
+ "loss": 0.5012,
1730
+ "step": 2870
1731
+ },
1732
+ {
1733
+ "epoch": 2.61,
1734
+ "learning_rate": 2.0734192711422406e-06,
1735
+ "loss": 0.4482,
1736
+ "step": 2880
1737
+ },
1738
+ {
1739
+ "epoch": 2.62,
1740
+ "learning_rate": 1.979811744651161e-06,
1741
+ "loss": 0.4864,
1742
+ "step": 2890
1743
+ },
1744
+ {
1745
+ "epoch": 2.63,
1746
+ "learning_rate": 1.8882791881492278e-06,
1747
+ "loss": 0.4551,
1748
+ "step": 2900
1749
+ },
1750
+ {
1751
+ "epoch": 2.64,
1752
+ "learning_rate": 1.7988298521036172e-06,
1753
+ "loss": 0.44,
1754
+ "step": 2910
1755
+ },
1756
+ {
1757
+ "epoch": 2.65,
1758
+ "learning_rate": 1.7114717992063463e-06,
1759
+ "loss": 0.4589,
1760
+ "step": 2920
1761
+ },
1762
+ {
1763
+ "epoch": 2.66,
1764
+ "learning_rate": 1.6262129036474837e-06,
1765
+ "loss": 0.4587,
1766
+ "step": 2930
1767
+ },
1768
+ {
1769
+ "epoch": 2.67,
1770
+ "learning_rate": 1.543060850405445e-06,
1771
+ "loss": 0.4385,
1772
+ "step": 2940
1773
+ },
1774
+ {
1775
+ "epoch": 2.67,
1776
+ "learning_rate": 1.4620231345542497e-06,
1777
+ "loss": 0.4963,
1778
+ "step": 2950
1779
+ },
1780
+ {
1781
+ "epoch": 2.68,
1782
+ "learning_rate": 1.3831070605879598e-06,
1783
+ "loss": 0.4926,
1784
+ "step": 2960
1785
+ },
1786
+ {
1787
+ "epoch": 2.69,
1788
+ "learning_rate": 1.3063197417622602e-06,
1789
+ "loss": 0.452,
1790
+ "step": 2970
1791
+ },
1792
+ {
1793
+ "epoch": 2.7,
1794
+ "learning_rate": 1.2316680994533174e-06,
1795
+ "loss": 0.4193,
1796
+ "step": 2980
1797
+ },
1798
+ {
1799
+ "epoch": 2.71,
1800
+ "learning_rate": 1.1591588625338767e-06,
1801
+ "loss": 0.4179,
1802
+ "step": 2990
1803
+ },
1804
+ {
1805
+ "epoch": 2.72,
1806
+ "learning_rate": 1.0887985667667522e-06,
1807
+ "loss": 0.4563,
1808
+ "step": 3000
1809
+ }
1810
+ ],
1811
+ "max_steps": 3309,
1812
+ "num_train_epochs": 3,
1813
+ "total_flos": 1.728722097093673e+17,
1814
+ "trial_name": null,
1815
+ "trial_params": null
1816
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad3d4177beb35316579d67e7342b9d0f04221e374825305456b1d62441e57f08
3
+ size 4027