xzuyn commited on
Commit
9ad2376
1 Parent(s): 0ea25b5

Upload folder using huggingface_hub

Browse files
checkpoint-205/README.md ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
checkpoint-205/adapter_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "layers_pattern": null,
8
+ "layers_to_transform": null,
9
+ "lora_alpha": 32,
10
+ "lora_dropout": 0.05,
11
+ "modules_to_save": null,
12
+ "peft_type": "LORA",
13
+ "r": 32,
14
+ "revision": null,
15
+ "target_modules": [
16
+ "gate_proj",
17
+ "down_proj",
18
+ "up_proj",
19
+ "q_proj",
20
+ "k_proj",
21
+ "v_proj",
22
+ "o_proj"
23
+ ],
24
+ "task_type": "CAUSAL_LM"
25
+ }
checkpoint-205/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:59e9f20c937859c8db6b5f12e2d812e6a831702164c317718a628f673c83976d
3
+ size 319977674
checkpoint-205/adapter_model/README.md ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
checkpoint-205/adapter_model/adapter_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "layers_pattern": null,
8
+ "layers_to_transform": null,
9
+ "lora_alpha": 32,
10
+ "lora_dropout": 0.05,
11
+ "modules_to_save": null,
12
+ "peft_type": "LORA",
13
+ "r": 32,
14
+ "revision": null,
15
+ "target_modules": [
16
+ "gate_proj",
17
+ "down_proj",
18
+ "up_proj",
19
+ "q_proj",
20
+ "k_proj",
21
+ "v_proj",
22
+ "o_proj"
23
+ ],
24
+ "task_type": "CAUSAL_LM"
25
+ }
checkpoint-205/adapter_model/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:59e9f20c937859c8db6b5f12e2d812e6a831702164c317718a628f673c83976d
3
+ size 319977674
checkpoint-205/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a31662ead7a3d99c3c6f37ea672cea1ae8ff7aa860f1b1139483f071feadf132
3
+ size 160746900
checkpoint-205/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79869b9b034b4439925c29bd5bb0c59128318a87df8dbb9db894e225ca1f35ac
3
+ size 14244
checkpoint-205/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2550c3e8252ec38fc99437c1d4407da0e3f59b0feae0e3159594bac01e24a060
3
+ size 1064
checkpoint-205/trainer_state.json ADDED
@@ -0,0 +1,1246 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 4.776699029126213,
5
+ "global_step": 205,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.631578947368421e-06,
13
+ "loss": 1.9044,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 5.263157894736842e-06,
19
+ "loss": 2.018,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 7.894736842105263e-06,
25
+ "loss": 2.0221,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.09,
30
+ "learning_rate": 1.0526315789473684e-05,
31
+ "loss": 1.8884,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1.3157894736842106e-05,
37
+ "loss": 1.9326,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 1.5789473684210526e-05,
43
+ "loss": 2.1193,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.16,
48
+ "learning_rate": 1.8421052631578947e-05,
49
+ "loss": 2.0827,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 2.105263157894737e-05,
55
+ "loss": 1.8846,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 2.368421052631579e-05,
61
+ "loss": 2.0943,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.23,
66
+ "learning_rate": 2.6315789473684212e-05,
67
+ "loss": 2.0881,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.26,
72
+ "learning_rate": 2.8947368421052634e-05,
73
+ "loss": 1.9275,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.28,
78
+ "learning_rate": 3.157894736842105e-05,
79
+ "loss": 1.6874,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.3,
84
+ "learning_rate": 3.421052631578947e-05,
85
+ "loss": 1.9874,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.33,
90
+ "learning_rate": 3.6842105263157895e-05,
91
+ "loss": 1.8901,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.35,
96
+ "learning_rate": 3.9473684210526316e-05,
97
+ "loss": 1.7683,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.37,
102
+ "learning_rate": 4.210526315789474e-05,
103
+ "loss": 2.0108,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.4,
108
+ "learning_rate": 4.473684210526316e-05,
109
+ "loss": 1.6579,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.42,
114
+ "learning_rate": 4.736842105263158e-05,
115
+ "loss": 1.9758,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.44,
120
+ "learning_rate": 5e-05,
121
+ "loss": 1.9118,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.47,
126
+ "learning_rate": 5.2631578947368424e-05,
127
+ "loss": 1.8964,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.49,
132
+ "learning_rate": 5.526315789473685e-05,
133
+ "loss": 1.9288,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.51,
138
+ "learning_rate": 5.789473684210527e-05,
139
+ "loss": 1.8706,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.54,
144
+ "learning_rate": 6.052631578947369e-05,
145
+ "loss": 1.8604,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.56,
150
+ "learning_rate": 6.31578947368421e-05,
151
+ "loss": 1.9001,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.58,
156
+ "learning_rate": 6.578947368421054e-05,
157
+ "loss": 1.8254,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.61,
162
+ "learning_rate": 6.842105263157895e-05,
163
+ "loss": 1.8111,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.63,
168
+ "learning_rate": 7.105263157894737e-05,
169
+ "loss": 1.7496,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.65,
174
+ "learning_rate": 7.368421052631579e-05,
175
+ "loss": 1.7604,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.68,
180
+ "learning_rate": 7.631578947368422e-05,
181
+ "loss": 1.6285,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.7,
186
+ "learning_rate": 7.894736842105263e-05,
187
+ "loss": 1.7226,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.72,
192
+ "learning_rate": 8.157894736842105e-05,
193
+ "loss": 1.9689,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.75,
198
+ "learning_rate": 8.421052631578948e-05,
199
+ "loss": 1.7737,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.77,
204
+ "learning_rate": 8.68421052631579e-05,
205
+ "loss": 1.8093,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.79,
210
+ "learning_rate": 8.947368421052632e-05,
211
+ "loss": 1.6292,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.82,
216
+ "learning_rate": 9.210526315789474e-05,
217
+ "loss": 1.8351,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.84,
222
+ "learning_rate": 9.473684210526316e-05,
223
+ "loss": 1.918,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.86,
228
+ "learning_rate": 9.736842105263158e-05,
229
+ "loss": 1.8161,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.89,
234
+ "learning_rate": 0.0001,
235
+ "loss": 1.8184,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.91,
240
+ "learning_rate": 9.983108108108109e-05,
241
+ "loss": 1.5839,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.93,
246
+ "learning_rate": 9.966216216216217e-05,
247
+ "loss": 1.8736,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.96,
252
+ "learning_rate": 9.949324324324325e-05,
253
+ "loss": 1.77,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.98,
258
+ "learning_rate": 9.932432432432433e-05,
259
+ "loss": 1.7912,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 1.0,
264
+ "learning_rate": 9.915540540540541e-05,
265
+ "loss": 1.6989,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 1.03,
270
+ "learning_rate": 9.89864864864865e-05,
271
+ "loss": 1.7327,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 1.05,
276
+ "learning_rate": 9.881756756756757e-05,
277
+ "loss": 1.7799,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 1.07,
282
+ "learning_rate": 9.864864864864865e-05,
283
+ "loss": 1.8298,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 1.1,
288
+ "learning_rate": 9.847972972972973e-05,
289
+ "loss": 1.8616,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 1.12,
294
+ "learning_rate": 9.831081081081081e-05,
295
+ "loss": 1.7984,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 1.14,
300
+ "learning_rate": 9.81418918918919e-05,
301
+ "loss": 1.7626,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 1.17,
306
+ "learning_rate": 9.797297297297297e-05,
307
+ "loss": 1.8424,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 1.19,
312
+ "learning_rate": 9.780405405405407e-05,
313
+ "loss": 1.6436,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 1.21,
318
+ "learning_rate": 9.763513513513513e-05,
319
+ "loss": 1.6303,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 1.23,
324
+ "learning_rate": 9.746621621621623e-05,
325
+ "loss": 1.7199,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 1.26,
330
+ "learning_rate": 9.729729729729731e-05,
331
+ "loss": 1.6789,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 1.28,
336
+ "learning_rate": 9.712837837837838e-05,
337
+ "loss": 1.8519,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "learning_rate": 9.695945945945947e-05,
343
+ "loss": 1.7961,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 1.33,
348
+ "learning_rate": 9.679054054054054e-05,
349
+ "loss": 1.6725,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 1.35,
354
+ "learning_rate": 9.662162162162163e-05,
355
+ "loss": 1.7249,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 1.37,
360
+ "learning_rate": 9.64527027027027e-05,
361
+ "loss": 1.5996,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 1.4,
366
+ "learning_rate": 9.628378378378379e-05,
367
+ "loss": 1.7033,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 1.42,
372
+ "learning_rate": 9.611486486486487e-05,
373
+ "loss": 1.7387,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 1.44,
378
+ "learning_rate": 9.594594594594595e-05,
379
+ "loss": 1.7804,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 1.47,
384
+ "learning_rate": 9.577702702702703e-05,
385
+ "loss": 1.7484,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 1.49,
390
+ "learning_rate": 9.56081081081081e-05,
391
+ "loss": 1.7526,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 1.51,
396
+ "learning_rate": 9.54391891891892e-05,
397
+ "loss": 1.7247,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 1.54,
402
+ "learning_rate": 9.527027027027028e-05,
403
+ "loss": 1.6276,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 1.56,
408
+ "learning_rate": 9.510135135135136e-05,
409
+ "loss": 1.8279,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 1.58,
414
+ "learning_rate": 9.493243243243244e-05,
415
+ "loss": 1.7067,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 1.61,
420
+ "learning_rate": 9.476351351351352e-05,
421
+ "loss": 1.7865,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 1.63,
426
+ "learning_rate": 9.45945945945946e-05,
427
+ "loss": 1.6693,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 1.65,
432
+ "learning_rate": 9.442567567567568e-05,
433
+ "loss": 1.6939,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 1.68,
438
+ "learning_rate": 9.425675675675676e-05,
439
+ "loss": 1.8667,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 1.7,
444
+ "learning_rate": 9.408783783783784e-05,
445
+ "loss": 1.746,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 1.72,
450
+ "learning_rate": 9.391891891891892e-05,
451
+ "loss": 1.6951,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 1.75,
456
+ "learning_rate": 9.375e-05,
457
+ "loss": 1.6749,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 1.77,
462
+ "learning_rate": 9.35810810810811e-05,
463
+ "loss": 1.8773,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 1.79,
468
+ "learning_rate": 9.341216216216216e-05,
469
+ "loss": 1.8064,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 1.82,
474
+ "learning_rate": 9.324324324324324e-05,
475
+ "loss": 1.6596,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 1.84,
480
+ "learning_rate": 9.307432432432432e-05,
481
+ "loss": 1.7321,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 1.86,
486
+ "learning_rate": 9.29054054054054e-05,
487
+ "loss": 1.7029,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 1.89,
492
+ "learning_rate": 9.27364864864865e-05,
493
+ "loss": 1.7392,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 1.91,
498
+ "learning_rate": 9.256756756756757e-05,
499
+ "loss": 1.7904,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 1.93,
504
+ "learning_rate": 9.239864864864866e-05,
505
+ "loss": 1.6794,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 1.96,
510
+ "learning_rate": 9.222972972972973e-05,
511
+ "loss": 1.8827,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 1.98,
516
+ "learning_rate": 9.206081081081082e-05,
517
+ "loss": 1.7051,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 2.0,
522
+ "learning_rate": 9.18918918918919e-05,
523
+ "loss": 1.7069,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 2.03,
528
+ "learning_rate": 9.172297297297297e-05,
529
+ "loss": 1.8051,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 2.05,
534
+ "learning_rate": 9.155405405405406e-05,
535
+ "loss": 1.6665,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 2.07,
540
+ "learning_rate": 9.138513513513513e-05,
541
+ "loss": 1.6218,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 2.1,
546
+ "learning_rate": 9.121621621621623e-05,
547
+ "loss": 1.5848,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 2.12,
552
+ "learning_rate": 9.10472972972973e-05,
553
+ "loss": 1.7544,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 2.14,
558
+ "learning_rate": 9.087837837837839e-05,
559
+ "loss": 1.6454,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 2.17,
564
+ "learning_rate": 9.070945945945947e-05,
565
+ "loss": 1.8591,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 2.19,
570
+ "learning_rate": 9.054054054054055e-05,
571
+ "loss": 1.7025,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 2.21,
576
+ "learning_rate": 9.037162162162163e-05,
577
+ "loss": 1.7824,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 2.24,
582
+ "learning_rate": 9.02027027027027e-05,
583
+ "loss": 1.6049,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 2.26,
588
+ "learning_rate": 9.003378378378379e-05,
589
+ "loss": 1.6961,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 2.28,
594
+ "learning_rate": 8.986486486486487e-05,
595
+ "loss": 1.8705,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 2.31,
600
+ "learning_rate": 8.969594594594595e-05,
601
+ "loss": 1.4498,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 2.33,
606
+ "learning_rate": 8.952702702702703e-05,
607
+ "loss": 1.6967,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 2.35,
612
+ "learning_rate": 8.935810810810811e-05,
613
+ "loss": 1.6345,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 2.38,
618
+ "learning_rate": 8.918918918918919e-05,
619
+ "loss": 1.8037,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 2.4,
624
+ "learning_rate": 8.902027027027027e-05,
625
+ "loss": 1.7896,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 2.42,
630
+ "learning_rate": 8.885135135135135e-05,
631
+ "loss": 1.751,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 2.45,
636
+ "learning_rate": 8.868243243243243e-05,
637
+ "loss": 1.482,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 2.47,
642
+ "learning_rate": 8.851351351351352e-05,
643
+ "loss": 1.6049,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 2.49,
648
+ "learning_rate": 8.83445945945946e-05,
649
+ "loss": 1.6029,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 2.52,
654
+ "learning_rate": 8.817567567567569e-05,
655
+ "loss": 1.875,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 2.54,
660
+ "learning_rate": 8.800675675675676e-05,
661
+ "loss": 1.5798,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 2.56,
666
+ "learning_rate": 8.783783783783784e-05,
667
+ "loss": 1.4883,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 2.59,
672
+ "learning_rate": 8.766891891891892e-05,
673
+ "loss": 1.7249,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 2.61,
678
+ "learning_rate": 8.75e-05,
679
+ "loss": 1.5086,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 2.63,
684
+ "learning_rate": 8.73310810810811e-05,
685
+ "loss": 1.39,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 2.66,
690
+ "learning_rate": 8.716216216216216e-05,
691
+ "loss": 1.7481,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 2.68,
696
+ "learning_rate": 8.699324324324325e-05,
697
+ "loss": 1.6514,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 2.7,
702
+ "learning_rate": 8.682432432432432e-05,
703
+ "loss": 1.8731,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 2.73,
708
+ "learning_rate": 8.665540540540542e-05,
709
+ "loss": 1.6415,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 2.75,
714
+ "learning_rate": 8.64864864864865e-05,
715
+ "loss": 1.7659,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 2.77,
720
+ "learning_rate": 8.631756756756756e-05,
721
+ "loss": 1.5691,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 2.8,
726
+ "learning_rate": 8.614864864864866e-05,
727
+ "loss": 1.6315,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 2.82,
732
+ "learning_rate": 8.597972972972972e-05,
733
+ "loss": 1.367,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 2.84,
738
+ "learning_rate": 8.581081081081082e-05,
739
+ "loss": 1.6088,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 2.87,
744
+ "learning_rate": 8.56418918918919e-05,
745
+ "loss": 1.7598,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 2.89,
750
+ "learning_rate": 8.547297297297298e-05,
751
+ "loss": 1.6255,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 2.91,
756
+ "learning_rate": 8.530405405405406e-05,
757
+ "loss": 1.5739,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 2.94,
762
+ "learning_rate": 8.513513513513514e-05,
763
+ "loss": 1.5542,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 2.96,
768
+ "learning_rate": 8.496621621621622e-05,
769
+ "loss": 1.5758,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 2.98,
774
+ "learning_rate": 8.47972972972973e-05,
775
+ "loss": 1.668,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 3.01,
780
+ "learning_rate": 8.462837837837838e-05,
781
+ "loss": 1.6283,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 3.03,
786
+ "learning_rate": 8.445945945945946e-05,
787
+ "loss": 1.6111,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 3.05,
792
+ "learning_rate": 8.429054054054054e-05,
793
+ "loss": 1.8061,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 3.08,
798
+ "learning_rate": 8.412162162162163e-05,
799
+ "loss": 1.4882,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 3.1,
804
+ "learning_rate": 8.39527027027027e-05,
805
+ "loss": 1.592,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 3.12,
810
+ "learning_rate": 8.378378378378379e-05,
811
+ "loss": 1.638,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 3.15,
816
+ "learning_rate": 8.361486486486487e-05,
817
+ "loss": 1.5036,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 3.17,
822
+ "learning_rate": 8.344594594594595e-05,
823
+ "loss": 1.6721,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 3.19,
828
+ "learning_rate": 8.327702702702703e-05,
829
+ "loss": 1.7742,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 3.22,
834
+ "learning_rate": 8.310810810810811e-05,
835
+ "loss": 1.484,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 3.24,
840
+ "learning_rate": 8.293918918918919e-05,
841
+ "loss": 1.3698,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 3.26,
846
+ "learning_rate": 8.277027027027028e-05,
847
+ "loss": 1.5644,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 3.29,
852
+ "learning_rate": 8.260135135135135e-05,
853
+ "loss": 1.6195,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 3.31,
858
+ "learning_rate": 8.243243243243243e-05,
859
+ "loss": 1.5031,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 3.33,
864
+ "learning_rate": 8.226351351351351e-05,
865
+ "loss": 1.5403,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 3.36,
870
+ "learning_rate": 8.209459459459459e-05,
871
+ "loss": 1.57,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 3.38,
876
+ "learning_rate": 8.192567567567569e-05,
877
+ "loss": 1.6454,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 3.4,
882
+ "learning_rate": 8.175675675675675e-05,
883
+ "loss": 1.5069,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 3.43,
888
+ "learning_rate": 8.158783783783785e-05,
889
+ "loss": 1.3782,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 3.45,
894
+ "learning_rate": 8.141891891891892e-05,
895
+ "loss": 1.4296,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 3.47,
900
+ "learning_rate": 8.125000000000001e-05,
901
+ "loss": 1.4303,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 3.5,
906
+ "learning_rate": 8.108108108108109e-05,
907
+ "loss": 1.6368,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 3.52,
912
+ "learning_rate": 8.091216216216216e-05,
913
+ "loss": 1.5951,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 3.54,
918
+ "learning_rate": 8.074324324324325e-05,
919
+ "loss": 1.5769,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 3.57,
924
+ "learning_rate": 8.057432432432432e-05,
925
+ "loss": 1.482,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 3.59,
930
+ "learning_rate": 8.040540540540541e-05,
931
+ "loss": 1.5415,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 3.61,
936
+ "learning_rate": 8.02364864864865e-05,
937
+ "loss": 1.6084,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 3.63,
942
+ "learning_rate": 8.006756756756757e-05,
943
+ "loss": 1.5483,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 3.66,
948
+ "learning_rate": 7.989864864864865e-05,
949
+ "loss": 1.4329,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 3.68,
954
+ "learning_rate": 7.972972972972974e-05,
955
+ "loss": 1.6068,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 3.7,
960
+ "learning_rate": 7.956081081081082e-05,
961
+ "loss": 1.5519,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 3.73,
966
+ "learning_rate": 7.93918918918919e-05,
967
+ "loss": 1.7387,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 3.75,
972
+ "learning_rate": 7.922297297297298e-05,
973
+ "loss": 1.5583,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 3.77,
978
+ "learning_rate": 7.905405405405406e-05,
979
+ "loss": 1.5003,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 3.8,
984
+ "learning_rate": 7.888513513513514e-05,
985
+ "loss": 1.5025,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 3.82,
990
+ "learning_rate": 7.871621621621622e-05,
991
+ "loss": 1.5322,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 3.84,
996
+ "learning_rate": 7.85472972972973e-05,
997
+ "loss": 1.6915,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 3.87,
1002
+ "learning_rate": 7.837837837837838e-05,
1003
+ "loss": 1.4906,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 3.89,
1008
+ "learning_rate": 7.820945945945946e-05,
1009
+ "loss": 1.4796,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 3.91,
1014
+ "learning_rate": 7.804054054054054e-05,
1015
+ "loss": 1.5712,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 3.94,
1020
+ "learning_rate": 7.787162162162162e-05,
1021
+ "loss": 1.459,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 3.96,
1026
+ "learning_rate": 7.77027027027027e-05,
1027
+ "loss": 1.4917,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 3.98,
1032
+ "learning_rate": 7.753378378378378e-05,
1033
+ "loss": 1.5149,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 4.01,
1038
+ "learning_rate": 7.736486486486488e-05,
1039
+ "loss": 1.4344,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 4.03,
1044
+ "learning_rate": 7.719594594594595e-05,
1045
+ "loss": 1.4152,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 4.05,
1050
+ "learning_rate": 7.702702702702703e-05,
1051
+ "loss": 1.3729,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 4.08,
1056
+ "learning_rate": 7.68581081081081e-05,
1057
+ "loss": 1.4462,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 4.1,
1062
+ "learning_rate": 7.668918918918919e-05,
1063
+ "loss": 1.5449,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 4.12,
1068
+ "learning_rate": 7.652027027027028e-05,
1069
+ "loss": 1.2145,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 4.15,
1074
+ "learning_rate": 7.635135135135135e-05,
1075
+ "loss": 1.4754,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 4.17,
1080
+ "learning_rate": 7.618243243243244e-05,
1081
+ "loss": 1.3256,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 4.19,
1086
+ "learning_rate": 7.601351351351351e-05,
1087
+ "loss": 1.3754,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 4.22,
1092
+ "learning_rate": 7.58445945945946e-05,
1093
+ "loss": 1.4267,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 4.24,
1098
+ "learning_rate": 7.567567567567568e-05,
1099
+ "loss": 1.4389,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 4.26,
1104
+ "learning_rate": 7.550675675675675e-05,
1105
+ "loss": 1.4266,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 4.29,
1110
+ "learning_rate": 7.533783783783785e-05,
1111
+ "loss": 1.4098,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 4.31,
1116
+ "learning_rate": 7.516891891891891e-05,
1117
+ "loss": 1.0715,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 4.33,
1122
+ "learning_rate": 7.500000000000001e-05,
1123
+ "loss": 1.4547,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 4.36,
1128
+ "learning_rate": 7.483108108108109e-05,
1129
+ "loss": 1.5758,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 4.38,
1134
+ "learning_rate": 7.466216216216217e-05,
1135
+ "loss": 1.3765,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 4.4,
1140
+ "learning_rate": 7.449324324324325e-05,
1141
+ "loss": 1.4207,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 4.43,
1146
+ "learning_rate": 7.432432432432433e-05,
1147
+ "loss": 1.4264,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 4.45,
1152
+ "learning_rate": 7.415540540540541e-05,
1153
+ "loss": 1.2975,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 4.47,
1158
+ "learning_rate": 7.398648648648649e-05,
1159
+ "loss": 1.5207,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 4.5,
1164
+ "learning_rate": 7.381756756756757e-05,
1165
+ "loss": 1.5805,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 4.52,
1170
+ "learning_rate": 7.364864864864865e-05,
1171
+ "loss": 1.3546,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 4.54,
1176
+ "learning_rate": 7.347972972972973e-05,
1177
+ "loss": 1.6047,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 4.57,
1182
+ "learning_rate": 7.331081081081081e-05,
1183
+ "loss": 1.4313,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 4.59,
1188
+ "learning_rate": 7.31418918918919e-05,
1189
+ "loss": 1.371,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 4.61,
1194
+ "learning_rate": 7.297297297297297e-05,
1195
+ "loss": 1.4083,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 4.64,
1200
+ "learning_rate": 7.280405405405406e-05,
1201
+ "loss": 1.6423,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 4.66,
1206
+ "learning_rate": 7.263513513513514e-05,
1207
+ "loss": 1.4229,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 4.68,
1212
+ "learning_rate": 7.246621621621622e-05,
1213
+ "loss": 1.4711,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 4.71,
1218
+ "learning_rate": 7.229729729729731e-05,
1219
+ "loss": 1.3806,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 4.73,
1224
+ "learning_rate": 7.212837837837838e-05,
1225
+ "loss": 1.4641,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 4.75,
1230
+ "learning_rate": 7.195945945945947e-05,
1231
+ "loss": 1.6534,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 4.78,
1236
+ "learning_rate": 7.179054054054054e-05,
1237
+ "loss": 1.3054,
1238
+ "step": 205
1239
+ }
1240
+ ],
1241
+ "max_steps": 630,
1242
+ "num_train_epochs": 15,
1243
+ "total_flos": 7.419797463583949e+16,
1244
+ "trial_name": null,
1245
+ "trial_params": null
1246
+ }
checkpoint-205/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba2fe107dab361c51487de4fc8c9ce45e8134a0096f693cda683706e603b37d9
3
+ size 4344