tyzhu commited on
Commit
2d78fe1
1 Parent(s): b6fd915

Training in progress, epoch 20, checkpoint

Browse files
checkpoint-3937/adapter_config.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "auto_mapping": null,
3
- "base_model_name_or_path": "Qwen/Qwen1.5-4B",
4
  "bias": "none",
5
  "fan_in_fan_out": false,
6
  "inference_mode": true,
 
1
  {
2
  "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
4
  "bias": "none",
5
  "fan_in_fan_out": false,
6
  "inference_mode": true,
checkpoint-3937/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39e36a12ff10f44cd1952313be615ea4bd063eb34189d26cb7492299464afa5f
3
+ size 143269386
checkpoint-3937/added_tokens.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "<|endoftext|>": 151643,
3
- "<|im_end|>": 151645,
4
- "<|im_start|>": 151644
5
  }
 
1
  {
2
+ "</s>": 2,
3
+ "<s>": 1,
4
+ "<unk>": 0
5
  }
checkpoint-3937/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e22ade3f04b512e7b9d4f92800c12d3cb52c026616061eb1b703cf47d11c395a
3
- size 224537202
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:855d476007fd095d29fed7676bbdb2fc54f859c7f8c2f0063d7d41e1d643e55e
3
+ size 286585234
checkpoint-3937/rng_state_0.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:215dbeff526b0a6fa6ce1ae3c4e6fc275ea6439582935fe49630dd2a165b5319
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d76e41340562ceadc1d8dd12c2dd6079490d9ae73414458c8ef8967290454997
3
  size 15024
checkpoint-3937/rng_state_1.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f6d3aef1e1a3a55daf430a4d94e5c814314395b1a9e70237961f02a541e2650e
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:206c6d345ab07bfeb1e9ee8216eee52922fcd6b95d25de850abc6e23927fedd4
3
  size 15024
checkpoint-3937/rng_state_2.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7f0817181ee46964366400eb5c3aa7fa009c13b90a600966295d09df203b2d34
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9273a5b3fbbeb61232b690544c9936d2f58fc97029b6025fb5d3ae9037d45548
3
  size 15024
checkpoint-3937/rng_state_3.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:622c9149d5b7fc9a45ae05c02e48bd5a696f09c9505f6bdf00663ebc597d0a25
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b71253ad64212a21a10ff174bc5c02dd841d69a0c9674adb085f219f44827da3
3
  size 15024
checkpoint-3937/special_tokens_map.json CHANGED
@@ -1,14 +1,6 @@
1
  {
2
- "additional_special_tokens": [
3
- "<|im_start|>",
4
- "<|im_end|>"
5
- ],
6
- "eos_token": {
7
- "content": "<|endoftext|>",
8
- "lstrip": false,
9
- "normalized": false,
10
- "rstrip": false,
11
- "single_word": false
12
- },
13
- "pad_token": "<|endoftext|>"
14
  }
 
1
  {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "pad_token": "</s>",
5
+ "unk_token": "<unk>"
 
 
 
 
 
 
 
 
6
  }
checkpoint-3937/tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
 
checkpoint-3937/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-3937/tokenizer_config.json CHANGED
@@ -1,24 +1,23 @@
1
  {
2
- "add_prefix_space": false,
3
  "added_tokens_decoder": {
4
- "151643": {
5
- "content": "<|endoftext|>",
6
  "lstrip": false,
7
  "normalized": false,
8
  "rstrip": false,
9
  "single_word": false,
10
  "special": true
11
  },
12
- "151644": {
13
- "content": "<|im_start|>",
14
  "lstrip": false,
15
  "normalized": false,
16
  "rstrip": false,
17
  "single_word": false,
18
  "special": true
19
  },
20
- "151645": {
21
- "content": "<|im_end|>",
22
  "lstrip": false,
23
  "normalized": false,
24
  "rstrip": false,
@@ -26,18 +25,16 @@
26
  "special": true
27
  }
28
  },
29
- "additional_special_tokens": [
30
- "<|im_start|>",
31
- "<|im_end|>"
32
- ],
33
- "bos_token": null,
34
- "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
35
  "clean_up_tokenization_spaces": false,
36
- "eos_token": "<|endoftext|>",
37
- "errors": "replace",
38
- "model_max_length": 32768,
39
- "pad_token": "<|endoftext|>",
40
- "split_special_tokens": false,
41
- "tokenizer_class": "Qwen2Tokenizer",
42
- "unk_token": null
 
 
43
  }
 
1
  {
 
2
  "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<unk>",
5
  "lstrip": false,
6
  "normalized": false,
7
  "rstrip": false,
8
  "single_word": false,
9
  "special": true
10
  },
11
+ "1": {
12
+ "content": "<s>",
13
  "lstrip": false,
14
  "normalized": false,
15
  "rstrip": false,
16
  "single_word": false,
17
  "special": true
18
  },
19
+ "2": {
20
+ "content": "</s>",
21
  "lstrip": false,
22
  "normalized": false,
23
  "rstrip": false,
 
25
  "special": true
26
  }
27
  },
28
+ "additional_special_tokens": [],
29
+ "bos_token": "<s>",
 
 
 
 
30
  "clean_up_tokenization_spaces": false,
31
+ "eos_token": "</s>",
32
+ "legacy": false,
33
+ "model_max_length": 1000000000000000019884624838656,
34
+ "pad_token": "</s>",
35
+ "padding_side": "left",
36
+ "sp_model_kwargs": {},
37
+ "tokenizer_class": "LlamaTokenizer",
38
+ "unk_token": "<unk>",
39
+ "use_default_system_prompt": true
40
  }
checkpoint-3937/trainer_state.json CHANGED
@@ -9,613 +9,560 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.5333333333333333,
13
- "grad_norm": 0.9283596873283386,
14
  "learning_rate": 3e-05,
15
- "loss": 1.963,
16
  "step": 100
17
  },
18
  {
19
- "epoch": 0.9973333333333333,
20
- "eval_accuracy": 0.5695384615384615,
21
- "eval_loss": 1.6438876390457153,
22
- "eval_runtime": 5.0534,
23
- "eval_samples_per_second": 98.943,
24
- "eval_steps_per_second": 12.467,
25
  "step": 187
26
  },
27
  {
28
- "epoch": 0.9973333333333333,
29
- "eval_exact_match": 16.2,
30
- "eval_f1": 25.888571428571442,
31
  "step": 187
32
  },
33
  {
34
- "epoch": 1.0666666666666667,
35
- "grad_norm": 1.0407154560089111,
36
  "learning_rate": 3e-05,
37
- "loss": 1.6668,
38
  "step": 200
39
  },
40
  {
41
  "epoch": 1.6,
42
- "grad_norm": 1.1281720399856567,
43
  "learning_rate": 3e-05,
44
- "loss": 1.6099,
45
  "step": 300
46
  },
47
  {
48
  "epoch": 2.0,
49
- "eval_accuracy": 0.5733333333333334,
50
- "eval_loss": 1.618270754814148,
51
- "eval_runtime": 5.1983,
52
- "eval_samples_per_second": 96.186,
53
- "eval_steps_per_second": 12.119,
54
  "step": 375
55
  },
56
  {
57
  "epoch": 2.0,
58
- "eval_exact_match": 16.6,
59
- "eval_f1": 26.287619047619057,
60
  "step": 375
61
  },
62
  {
63
- "epoch": 2.1333333333333333,
64
- "grad_norm": 1.2508225440979004,
65
  "learning_rate": 3e-05,
66
- "loss": 1.5866,
67
  "step": 400
68
  },
69
  {
70
- "epoch": 2.6666666666666665,
71
- "grad_norm": 1.3197834491729736,
72
  "learning_rate": 3e-05,
73
- "loss": 1.524,
74
  "step": 500
75
  },
76
  {
77
- "epoch": 2.997333333333333,
78
- "eval_accuracy": 0.5743589743589743,
79
- "eval_loss": 1.6164497137069702,
80
- "eval_runtime": 5.0495,
81
- "eval_samples_per_second": 99.019,
82
- "eval_steps_per_second": 12.476,
83
  "step": 562
84
  },
85
  {
86
- "epoch": 2.997333333333333,
87
- "eval_exact_match": 17.8,
88
- "eval_f1": 27.35428571428571,
89
  "step": 562
90
  },
91
  {
92
  "epoch": 3.2,
93
- "grad_norm": 1.6320394277572632,
94
  "learning_rate": 3e-05,
95
- "loss": 1.4752,
96
  "step": 600
97
  },
98
  {
99
- "epoch": 3.7333333333333334,
100
- "grad_norm": 1.7101725339889526,
101
  "learning_rate": 3e-05,
102
- "loss": 1.3938,
103
  "step": 700
104
  },
105
  {
106
  "epoch": 4.0,
107
- "eval_accuracy": 0.5728717948717948,
108
- "eval_loss": 1.6376241445541382,
109
- "eval_runtime": 4.9662,
110
- "eval_samples_per_second": 100.68,
111
- "eval_steps_per_second": 12.686,
112
  "step": 750
113
  },
114
  {
115
  "epoch": 4.0,
116
- "eval_exact_match": 18.0,
117
- "eval_f1": 27.498571428571427,
118
  "step": 750
119
  },
120
  {
121
- "epoch": 4.266666666666667,
122
- "grad_norm": 2.1727664470672607,
123
  "learning_rate": 3e-05,
124
- "loss": 1.342,
125
  "step": 800
126
  },
127
  {
128
  "epoch": 4.8,
129
- "grad_norm": 2.341384172439575,
130
  "learning_rate": 3e-05,
131
- "loss": 1.2685,
132
  "step": 900
133
  },
134
  {
135
- "epoch": 4.997333333333334,
136
- "eval_accuracy": 0.5699487179487179,
137
- "eval_loss": 1.684618353843689,
138
- "eval_runtime": 5.5457,
139
- "eval_samples_per_second": 90.16,
140
- "eval_steps_per_second": 11.36,
141
  "step": 937
142
  },
143
  {
144
- "epoch": 4.997333333333334,
145
- "eval_exact_match": 19.2,
146
- "eval_f1": 28.190952380952375,
147
  "step": 937
148
  },
149
  {
150
- "epoch": 5.333333333333333,
151
- "grad_norm": 2.5624172687530518,
152
  "learning_rate": 3e-05,
153
- "loss": 1.1847,
154
  "step": 1000
155
  },
156
  {
157
- "epoch": 5.866666666666667,
158
- "grad_norm": 2.8159019947052,
159
  "learning_rate": 3e-05,
160
- "loss": 1.1591,
161
  "step": 1100
162
  },
163
  {
164
  "epoch": 6.0,
165
- "eval_accuracy": 0.5672820512820513,
166
- "eval_loss": 1.7546838521957397,
167
- "eval_runtime": 5.0186,
168
- "eval_samples_per_second": 99.629,
169
- "eval_steps_per_second": 12.553,
170
  "step": 1125
171
  },
172
  {
173
  "epoch": 6.0,
174
- "eval_exact_match": 16.6,
175
- "eval_f1": 25.054761904761914,
176
  "step": 1125
177
  },
178
  {
179
  "epoch": 6.4,
180
- "grad_norm": 3.056166410446167,
181
  "learning_rate": 3e-05,
182
- "loss": 1.0486,
183
  "step": 1200
184
  },
185
  {
186
- "epoch": 6.933333333333334,
187
- "grad_norm": 2.873281240463257,
188
  "learning_rate": 3e-05,
189
- "loss": 1.0444,
190
  "step": 1300
191
  },
192
  {
193
- "epoch": 6.997333333333334,
194
- "eval_accuracy": 0.5643076923076923,
195
- "eval_loss": 1.8394943475723267,
196
- "eval_runtime": 5.5931,
197
- "eval_samples_per_second": 89.396,
198
- "eval_steps_per_second": 11.264,
199
  "step": 1312
200
  },
201
  {
202
- "epoch": 6.997333333333334,
203
- "eval_exact_match": 15.8,
204
- "eval_f1": 23.925714285714296,
205
  "step": 1312
206
  },
207
  {
208
- "epoch": 7.466666666666667,
209
- "grad_norm": 3.069322109222412,
210
  "learning_rate": 3e-05,
211
- "loss": 0.9374,
212
  "step": 1400
213
  },
214
  {
215
  "epoch": 8.0,
216
- "grad_norm": 3.2732129096984863,
217
  "learning_rate": 3e-05,
218
- "loss": 0.9535,
219
  "step": 1500
220
  },
221
  {
222
  "epoch": 8.0,
223
- "eval_accuracy": 0.5612820512820513,
224
- "eval_loss": 1.9008216857910156,
225
- "eval_runtime": 5.1357,
226
- "eval_samples_per_second": 97.358,
227
- "eval_steps_per_second": 12.267,
228
  "step": 1500
229
  },
230
  {
231
  "epoch": 8.0,
232
- "eval_exact_match": 15.2,
233
- "eval_f1": 23.12063492063493,
234
  "step": 1500
235
  },
236
  {
237
- "epoch": 8.533333333333333,
238
- "grad_norm": 3.382610321044922,
239
  "learning_rate": 3e-05,
240
- "loss": 0.8235,
241
  "step": 1600
242
  },
243
  {
244
- "epoch": 8.997333333333334,
245
- "eval_accuracy": 0.5592307692307692,
246
- "eval_loss": 2.026776075363159,
247
- "eval_runtime": 5.5056,
248
- "eval_samples_per_second": 90.817,
249
- "eval_steps_per_second": 11.443,
250
  "step": 1687
251
  },
252
  {
253
- "epoch": 8.997333333333334,
254
- "eval_exact_match": 13.8,
255
- "eval_f1": 21.42285714285715,
256
  "step": 1687
257
  },
258
  {
259
- "epoch": 9.066666666666666,
260
- "grad_norm": 4.363819122314453,
261
  "learning_rate": 3e-05,
262
- "loss": 0.8547,
263
  "step": 1700
264
  },
265
  {
266
  "epoch": 9.6,
267
- "grad_norm": 3.685805320739746,
268
  "learning_rate": 3e-05,
269
- "loss": 0.7635,
270
  "step": 1800
271
  },
272
  {
273
  "epoch": 10.0,
274
- "eval_accuracy": 0.5567692307692308,
275
- "eval_loss": 2.0937435626983643,
276
- "eval_runtime": 5.235,
277
- "eval_samples_per_second": 95.512,
278
- "eval_steps_per_second": 12.034,
279
  "step": 1875
280
  },
281
  {
282
  "epoch": 10.0,
283
- "eval_exact_match": 15.2,
284
- "eval_f1": 22.031904761904762,
285
  "step": 1875
286
  },
287
  {
288
- "epoch": 10.133333333333333,
289
- "grad_norm": 3.5318357944488525,
290
  "learning_rate": 3e-05,
291
- "loss": 0.7554,
292
  "step": 1900
293
  },
294
  {
295
- "epoch": 10.666666666666666,
296
- "grad_norm": 3.6454901695251465,
297
  "learning_rate": 3e-05,
298
- "loss": 0.6978,
299
  "step": 2000
300
  },
301
  {
302
- "epoch": 10.997333333333334,
303
- "eval_accuracy": 0.556974358974359,
304
- "eval_loss": 2.1557741165161133,
305
- "eval_runtime": 5.1556,
306
- "eval_samples_per_second": 96.981,
307
- "eval_steps_per_second": 12.22,
308
  "step": 2062
309
  },
310
  {
311
- "epoch": 10.997333333333334,
312
- "eval_exact_match": 15.0,
313
- "eval_f1": 23.010000000000012,
314
  "step": 2062
315
  },
316
  {
317
  "epoch": 11.2,
318
- "grad_norm": 3.658707618713379,
319
  "learning_rate": 3e-05,
320
- "loss": 0.6969,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 11.733333333333333,
325
- "grad_norm": 3.7350716590881348,
326
  "learning_rate": 3e-05,
327
- "loss": 0.6615,
328
  "step": 2200
329
  },
330
  {
331
  "epoch": 12.0,
332
- "eval_accuracy": 0.5551794871794872,
333
- "eval_loss": 2.239973545074463,
334
- "eval_runtime": 4.9026,
335
- "eval_samples_per_second": 101.986,
336
- "eval_steps_per_second": 12.85,
337
  "step": 2250
338
  },
339
  {
340
  "epoch": 12.0,
341
- "eval_exact_match": 15.6,
342
- "eval_f1": 23.333492063492066,
343
  "step": 2250
344
  },
345
  {
346
- "epoch": 12.266666666666667,
347
- "grad_norm": 4.4434709548950195,
348
  "learning_rate": 3e-05,
349
- "loss": 0.632,
350
  "step": 2300
351
  },
352
  {
353
  "epoch": 12.8,
354
- "grad_norm": 3.4198718070983887,
355
  "learning_rate": 3e-05,
356
- "loss": 0.6262,
357
  "step": 2400
358
  },
359
  {
360
- "epoch": 12.997333333333334,
361
- "eval_accuracy": 0.5556410256410257,
362
- "eval_loss": 2.268662452697754,
363
- "eval_runtime": 5.327,
364
- "eval_samples_per_second": 93.862,
365
- "eval_steps_per_second": 11.827,
366
  "step": 2437
367
  },
368
  {
369
- "epoch": 12.997333333333334,
370
- "eval_exact_match": 14.4,
371
- "eval_f1": 22.583492063492077,
372
  "step": 2437
373
  },
374
  {
375
- "epoch": 13.333333333333334,
376
- "grad_norm": 3.158496618270874,
377
  "learning_rate": 3e-05,
378
- "loss": 0.5973,
379
  "step": 2500
380
  },
381
  {
382
- "epoch": 13.866666666666667,
383
- "grad_norm": 3.5720713138580322,
384
  "learning_rate": 3e-05,
385
- "loss": 0.5958,
386
  "step": 2600
387
  },
388
  {
389
  "epoch": 14.0,
390
- "eval_accuracy": 0.5536923076923077,
391
- "eval_loss": 2.358220100402832,
392
- "eval_runtime": 5.236,
393
- "eval_samples_per_second": 95.493,
394
- "eval_steps_per_second": 12.032,
395
  "step": 2625
396
  },
397
  {
398
  "epoch": 14.0,
399
- "eval_exact_match": 14.6,
400
- "eval_f1": 21.79904761904763,
401
  "step": 2625
402
  },
403
  {
404
  "epoch": 14.4,
405
- "grad_norm": 5.126227855682373,
406
  "learning_rate": 3e-05,
407
- "loss": 0.5692,
408
  "step": 2700
409
  },
410
  {
411
- "epoch": 14.933333333333334,
412
- "grad_norm": 3.3391811847686768,
413
  "learning_rate": 3e-05,
414
- "loss": 0.5778,
415
  "step": 2800
416
  },
417
  {
418
- "epoch": 14.997333333333334,
419
- "eval_accuracy": 0.5533846153846154,
420
- "eval_loss": 2.395972728729248,
421
- "eval_runtime": 5.1135,
422
- "eval_samples_per_second": 97.78,
423
- "eval_steps_per_second": 12.32,
424
  "step": 2812
425
  },
426
  {
427
- "epoch": 14.997333333333334,
428
- "eval_exact_match": 14.6,
429
- "eval_f1": 22.537301587301595,
430
  "step": 2812
431
  },
432
  {
433
- "epoch": 15.466666666666667,
434
- "grad_norm": 2.8773677349090576,
435
  "learning_rate": 3e-05,
436
- "loss": 0.5461,
437
  "step": 2900
438
  },
439
  {
440
  "epoch": 16.0,
441
- "grad_norm": 3.580195903778076,
442
  "learning_rate": 3e-05,
443
- "loss": 0.5661,
444
  "step": 3000
445
  },
446
  {
447
  "epoch": 16.0,
448
- "eval_accuracy": 0.5533846153846154,
449
- "eval_loss": 2.4321775436401367,
450
- "eval_runtime": 5.0238,
451
- "eval_samples_per_second": 99.527,
452
- "eval_steps_per_second": 12.54,
453
  "step": 3000
454
  },
455
  {
456
  "epoch": 16.0,
457
- "eval_exact_match": 14.0,
458
- "eval_f1": 22.300808080808093,
459
  "step": 3000
460
  },
461
  {
462
- "epoch": 16.533333333333335,
463
- "grad_norm": 3.6829302310943604,
464
  "learning_rate": 3e-05,
465
- "loss": 0.5277,
466
  "step": 3100
467
  },
468
  {
469
- "epoch": 16.997333333333334,
470
- "eval_accuracy": 0.5515384615384615,
471
- "eval_loss": 2.482792615890503,
472
- "eval_runtime": 5.3006,
473
- "eval_samples_per_second": 94.328,
474
- "eval_steps_per_second": 11.885,
475
  "step": 3187
476
  },
477
  {
478
- "epoch": 16.997333333333334,
479
- "eval_exact_match": 13.4,
480
- "eval_f1": 21.439682539682543,
481
  "step": 3187
482
  },
483
  {
484
- "epoch": 17.066666666666666,
485
- "grad_norm": 3.1766672134399414,
486
  "learning_rate": 3e-05,
487
- "loss": 0.5475,
488
  "step": 3200
489
  },
490
  {
491
  "epoch": 17.6,
492
- "grad_norm": 3.070977210998535,
493
  "learning_rate": 3e-05,
494
- "loss": 0.5211,
495
  "step": 3300
496
  },
497
  {
498
  "epoch": 18.0,
499
- "eval_accuracy": 0.5516410256410257,
500
- "eval_loss": 2.510619640350342,
501
- "eval_runtime": 5.0519,
502
- "eval_samples_per_second": 98.973,
503
- "eval_steps_per_second": 12.471,
504
  "step": 3375
505
  },
506
  {
507
  "epoch": 18.0,
508
- "eval_exact_match": 12.6,
509
- "eval_f1": 20.136349206349216,
510
  "step": 3375
511
  },
512
  {
513
- "epoch": 18.133333333333333,
514
- "grad_norm": 4.728194713592529,
515
  "learning_rate": 3e-05,
516
- "loss": 0.5344,
517
  "step": 3400
518
  },
519
  {
520
- "epoch": 18.666666666666668,
521
- "grad_norm": 3.1506965160369873,
522
  "learning_rate": 3e-05,
523
- "loss": 0.5189,
524
  "step": 3500
525
  },
526
  {
527
- "epoch": 18.997333333333334,
528
- "eval_accuracy": 0.5514871794871795,
529
- "eval_loss": 2.5705809593200684,
530
- "eval_runtime": 4.9081,
531
- "eval_samples_per_second": 101.872,
532
- "eval_steps_per_second": 12.836,
533
  "step": 3562
534
  },
535
  {
536
- "epoch": 18.997333333333334,
537
- "eval_exact_match": 13.8,
538
- "eval_f1": 21.51634920634921,
539
  "step": 3562
540
  },
541
  {
542
  "epoch": 19.2,
543
- "grad_norm": 2.7202513217926025,
544
  "learning_rate": 3e-05,
545
- "loss": 0.5198,
546
  "step": 3600
547
  },
548
  {
549
- "epoch": 19.733333333333334,
550
- "grad_norm": 2.581169366836548,
551
  "learning_rate": 3e-05,
552
- "loss": 0.5166,
553
  "step": 3700
554
  },
555
  {
556
  "epoch": 20.0,
557
- "eval_accuracy": 0.5526153846153846,
558
- "eval_loss": 2.542161226272583,
559
- "eval_runtime": 5.0196,
560
- "eval_samples_per_second": 99.609,
561
- "eval_steps_per_second": 12.551,
562
  "step": 3750
563
  },
564
  {
565
  "epoch": 20.0,
566
- "eval_exact_match": 14.4,
567
- "eval_f1": 21.563492063492074,
568
  "step": 3750
569
  },
570
  {
571
- "epoch": 20.266666666666666,
572
- "grad_norm": 2.6741623878479004,
573
  "learning_rate": 3e-05,
574
- "loss": 0.5112,
575
  "step": 3800
576
  },
577
  {
578
  "epoch": 20.8,
579
- "grad_norm": 3.241607666015625,
580
  "learning_rate": 3e-05,
581
- "loss": 0.5132,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 20.997333333333334,
586
- "eval_accuracy": 0.550923076923077,
587
- "eval_loss": 2.594764232635498,
588
- "eval_runtime": 5.2553,
589
- "eval_samples_per_second": 95.142,
590
- "eval_steps_per_second": 11.988,
591
  "step": 3937
592
  },
593
  {
594
- "epoch": 20.997333333333334,
595
- "eval_exact_match": 13.4,
596
- "eval_f1": 21.25015873015874,
597
  "step": 3937
598
  }
599
  ],
600
  "logging_steps": 100,
601
  "max_steps": 9350,
602
- "num_input_tokens_seen": 0,
603
  "num_train_epochs": 50,
604
  "save_steps": 500,
605
- "stateful_callbacks": {
606
- "TrainerControl": {
607
- "args": {
608
- "should_epoch_stop": false,
609
- "should_evaluate": false,
610
- "should_log": false,
611
- "should_save": true,
612
- "should_training_stop": false
613
- },
614
- "attributes": {}
615
- }
616
- },
617
- "total_flos": 7.212341785106842e+16,
618
- "train_batch_size": 2,
619
  "trial_name": null,
620
  "trial_params": null
621
  }
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.53,
 
13
  "learning_rate": 3e-05,
14
+ "loss": 2.0369,
15
  "step": 100
16
  },
17
  {
18
+ "epoch": 1.0,
19
+ "eval_accuracy": 0.6128205128205129,
20
+ "eval_loss": 1.2953312397003174,
21
+ "eval_runtime": 4.1114,
22
+ "eval_samples_per_second": 121.614,
23
+ "eval_steps_per_second": 15.323,
24
  "step": 187
25
  },
26
  {
27
+ "epoch": 1.0,
28
+ "eval_exact_match": 27.2,
29
+ "eval_f1": 37.44218938149973,
30
  "step": 187
31
  },
32
  {
33
+ "epoch": 1.07,
 
34
  "learning_rate": 3e-05,
35
+ "loss": 1.3264,
36
  "step": 200
37
  },
38
  {
39
  "epoch": 1.6,
 
40
  "learning_rate": 3e-05,
41
+ "loss": 1.2821,
42
  "step": 300
43
  },
44
  {
45
  "epoch": 2.0,
46
+ "eval_accuracy": 0.6146153846153846,
47
+ "eval_loss": 1.2741013765335083,
48
+ "eval_runtime": 4.8302,
49
+ "eval_samples_per_second": 103.516,
50
+ "eval_steps_per_second": 13.043,
51
  "step": 375
52
  },
53
  {
54
  "epoch": 2.0,
55
+ "eval_exact_match": 25.6,
56
+ "eval_f1": 38.026666666666664,
57
  "step": 375
58
  },
59
  {
60
+ "epoch": 2.13,
 
61
  "learning_rate": 3e-05,
62
+ "loss": 1.2348,
63
  "step": 400
64
  },
65
  {
66
+ "epoch": 2.67,
 
67
  "learning_rate": 3e-05,
68
+ "loss": 1.1987,
69
  "step": 500
70
  },
71
  {
72
+ "epoch": 3.0,
73
+ "eval_accuracy": 0.6161538461538462,
74
+ "eval_loss": 1.2714948654174805,
75
+ "eval_runtime": 4.868,
76
+ "eval_samples_per_second": 102.712,
77
+ "eval_steps_per_second": 12.942,
78
  "step": 562
79
  },
80
  {
81
+ "epoch": 3.0,
82
+ "eval_exact_match": 28.0,
83
+ "eval_f1": 40.6962091503268,
84
  "step": 562
85
  },
86
  {
87
  "epoch": 3.2,
 
88
  "learning_rate": 3e-05,
89
+ "loss": 1.1537,
90
  "step": 600
91
  },
92
  {
93
+ "epoch": 3.73,
 
94
  "learning_rate": 3e-05,
95
+ "loss": 1.066,
96
  "step": 700
97
  },
98
  {
99
  "epoch": 4.0,
100
+ "eval_accuracy": 0.6151282051282051,
101
+ "eval_loss": 1.3011157512664795,
102
+ "eval_runtime": 4.407,
103
+ "eval_samples_per_second": 113.455,
104
+ "eval_steps_per_second": 14.295,
105
  "step": 750
106
  },
107
  {
108
  "epoch": 4.0,
109
+ "eval_exact_match": 26.2,
110
+ "eval_f1": 38.19190476190477,
111
  "step": 750
112
  },
113
  {
114
+ "epoch": 4.27,
 
115
  "learning_rate": 3e-05,
116
+ "loss": 1.001,
117
  "step": 800
118
  },
119
  {
120
  "epoch": 4.8,
 
121
  "learning_rate": 3e-05,
122
+ "loss": 0.9381,
123
  "step": 900
124
  },
125
  {
126
+ "epoch": 5.0,
127
+ "eval_accuracy": 0.6125641025641025,
128
+ "eval_loss": 1.3728467226028442,
129
+ "eval_runtime": 5.153,
130
+ "eval_samples_per_second": 97.03,
131
+ "eval_steps_per_second": 12.226,
132
  "step": 937
133
  },
134
  {
135
+ "epoch": 5.0,
136
+ "eval_exact_match": 24.6,
137
+ "eval_f1": 37.08666666666668,
138
  "step": 937
139
  },
140
  {
141
+ "epoch": 5.33,
 
142
  "learning_rate": 3e-05,
143
+ "loss": 0.8667,
144
  "step": 1000
145
  },
146
  {
147
+ "epoch": 5.87,
 
148
  "learning_rate": 3e-05,
149
+ "loss": 0.8238,
150
  "step": 1100
151
  },
152
  {
153
  "epoch": 6.0,
154
+ "eval_accuracy": 0.6090769230769231,
155
+ "eval_loss": 1.4598820209503174,
156
+ "eval_runtime": 4.0994,
157
+ "eval_samples_per_second": 121.97,
158
+ "eval_steps_per_second": 15.368,
159
  "step": 1125
160
  },
161
  {
162
  "epoch": 6.0,
163
+ "eval_exact_match": 23.0,
164
+ "eval_f1": 35.324444444444445,
165
  "step": 1125
166
  },
167
  {
168
  "epoch": 6.4,
 
169
  "learning_rate": 3e-05,
170
+ "loss": 0.7461,
171
  "step": 1200
172
  },
173
  {
174
+ "epoch": 6.93,
 
175
  "learning_rate": 3e-05,
176
+ "loss": 0.7289,
177
  "step": 1300
178
  },
179
  {
180
+ "epoch": 7.0,
181
+ "eval_accuracy": 0.6064102564102564,
182
+ "eval_loss": 1.5455493927001953,
183
+ "eval_runtime": 4.1683,
184
+ "eval_samples_per_second": 119.952,
185
+ "eval_steps_per_second": 15.114,
186
  "step": 1312
187
  },
188
  {
189
+ "epoch": 7.0,
190
+ "eval_exact_match": 21.8,
191
+ "eval_f1": 34.57222222222222,
192
  "step": 1312
193
  },
194
  {
195
+ "epoch": 7.47,
 
196
  "learning_rate": 3e-05,
197
+ "loss": 0.6464,
198
  "step": 1400
199
  },
200
  {
201
  "epoch": 8.0,
 
202
  "learning_rate": 3e-05,
203
+ "loss": 0.6559,
204
  "step": 1500
205
  },
206
  {
207
  "epoch": 8.0,
208
+ "eval_accuracy": 0.6025641025641025,
209
+ "eval_loss": 1.6359158754348755,
210
+ "eval_runtime": 4.3161,
211
+ "eval_samples_per_second": 115.844,
212
+ "eval_steps_per_second": 14.596,
213
  "step": 1500
214
  },
215
  {
216
  "epoch": 8.0,
217
+ "eval_exact_match": 20.6,
218
+ "eval_f1": 33.70507936507938,
219
  "step": 1500
220
  },
221
  {
222
+ "epoch": 8.53,
 
223
  "learning_rate": 3e-05,
224
+ "loss": 0.5733,
225
  "step": 1600
226
  },
227
  {
228
+ "epoch": 9.0,
229
+ "eval_accuracy": 0.6006153846153847,
230
+ "eval_loss": 1.7149169445037842,
231
+ "eval_runtime": 4.1084,
232
+ "eval_samples_per_second": 121.702,
233
+ "eval_steps_per_second": 15.334,
234
  "step": 1687
235
  },
236
  {
237
+ "epoch": 9.0,
238
+ "eval_exact_match": 22.0,
239
+ "eval_f1": 33.980476190476196,
240
  "step": 1687
241
  },
242
  {
243
+ "epoch": 9.07,
 
244
  "learning_rate": 3e-05,
245
+ "loss": 0.5859,
246
  "step": 1700
247
  },
248
  {
249
  "epoch": 9.6,
 
250
  "learning_rate": 3e-05,
251
+ "loss": 0.5336,
252
  "step": 1800
253
  },
254
  {
255
  "epoch": 10.0,
256
+ "eval_accuracy": 0.5989230769230769,
257
+ "eval_loss": 1.8006359338760376,
258
+ "eval_runtime": 4.4039,
259
+ "eval_samples_per_second": 113.535,
260
+ "eval_steps_per_second": 14.305,
261
  "step": 1875
262
  },
263
  {
264
  "epoch": 10.0,
265
+ "eval_exact_match": 20.4,
266
+ "eval_f1": 32.68793650793652,
267
  "step": 1875
268
  },
269
  {
270
+ "epoch": 10.13,
 
271
  "learning_rate": 3e-05,
272
+ "loss": 0.5379,
273
  "step": 1900
274
  },
275
  {
276
+ "epoch": 10.67,
 
277
  "learning_rate": 3e-05,
278
+ "loss": 0.5116,
279
  "step": 2000
280
  },
281
  {
282
+ "epoch": 11.0,
283
+ "eval_accuracy": 0.5982051282051282,
284
+ "eval_loss": 1.8851306438446045,
285
+ "eval_runtime": 4.8059,
286
+ "eval_samples_per_second": 104.038,
287
+ "eval_steps_per_second": 13.109,
288
  "step": 2062
289
  },
290
  {
291
+ "epoch": 11.0,
292
+ "eval_exact_match": 19.8,
293
+ "eval_f1": 31.627460317460333,
294
  "step": 2062
295
  },
296
  {
297
  "epoch": 11.2,
 
298
  "learning_rate": 3e-05,
299
+ "loss": 0.5025,
300
  "step": 2100
301
  },
302
  {
303
+ "epoch": 11.73,
 
304
  "learning_rate": 3e-05,
305
+ "loss": 0.4934,
306
  "step": 2200
307
  },
308
  {
309
  "epoch": 12.0,
310
+ "eval_accuracy": 0.5982051282051282,
311
+ "eval_loss": 1.9262274503707886,
312
+ "eval_runtime": 4.2024,
313
+ "eval_samples_per_second": 118.98,
314
+ "eval_steps_per_second": 14.992,
315
  "step": 2250
316
  },
317
  {
318
  "epoch": 12.0,
319
+ "eval_exact_match": 19.2,
320
+ "eval_f1": 30.531746031746042,
321
  "step": 2250
322
  },
323
  {
324
+ "epoch": 12.27,
 
325
  "learning_rate": 3e-05,
326
+ "loss": 0.4827,
327
  "step": 2300
328
  },
329
  {
330
  "epoch": 12.8,
 
331
  "learning_rate": 3e-05,
332
+ "loss": 0.4823,
333
  "step": 2400
334
  },
335
  {
336
+ "epoch": 13.0,
337
+ "eval_accuracy": 0.5974358974358974,
338
+ "eval_loss": 1.9412546157836914,
339
+ "eval_runtime": 4.4921,
340
+ "eval_samples_per_second": 111.306,
341
+ "eval_steps_per_second": 14.025,
342
  "step": 2437
343
  },
344
  {
345
+ "epoch": 13.0,
346
+ "eval_exact_match": 18.8,
347
+ "eval_f1": 30.277619047619062,
348
  "step": 2437
349
  },
350
  {
351
+ "epoch": 13.33,
 
352
  "learning_rate": 3e-05,
353
+ "loss": 0.467,
354
  "step": 2500
355
  },
356
  {
357
+ "epoch": 13.87,
 
358
  "learning_rate": 3e-05,
359
+ "loss": 0.47,
360
  "step": 2600
361
  },
362
  {
363
  "epoch": 14.0,
364
+ "eval_accuracy": 0.5967179487179487,
365
+ "eval_loss": 2.01212739944458,
366
+ "eval_runtime": 4.0908,
367
+ "eval_samples_per_second": 122.226,
368
+ "eval_steps_per_second": 15.4,
369
  "step": 2625
370
  },
371
  {
372
  "epoch": 14.0,
373
+ "eval_exact_match": 18.8,
374
+ "eval_f1": 30.478571428571446,
375
  "step": 2625
376
  },
377
  {
378
  "epoch": 14.4,
 
379
  "learning_rate": 3e-05,
380
+ "loss": 0.454,
381
  "step": 2700
382
  },
383
  {
384
+ "epoch": 14.93,
 
385
  "learning_rate": 3e-05,
386
+ "loss": 0.4661,
387
  "step": 2800
388
  },
389
  {
390
+ "epoch": 15.0,
391
+ "eval_accuracy": 0.5968205128205128,
392
+ "eval_loss": 2.0249853134155273,
393
+ "eval_runtime": 4.0875,
394
+ "eval_samples_per_second": 122.323,
395
+ "eval_steps_per_second": 15.413,
396
  "step": 2812
397
  },
398
  {
399
+ "epoch": 15.0,
400
+ "eval_exact_match": 20.6,
401
+ "eval_f1": 32.526031746031755,
402
  "step": 2812
403
  },
404
  {
405
+ "epoch": 15.47,
 
406
  "learning_rate": 3e-05,
407
+ "loss": 0.4423,
408
  "step": 2900
409
  },
410
  {
411
  "epoch": 16.0,
 
412
  "learning_rate": 3e-05,
413
+ "loss": 0.462,
414
  "step": 3000
415
  },
416
  {
417
  "epoch": 16.0,
418
+ "eval_accuracy": 0.598974358974359,
419
+ "eval_loss": 1.9804810285568237,
420
+ "eval_runtime": 4.0893,
421
+ "eval_samples_per_second": 122.269,
422
+ "eval_steps_per_second": 15.406,
423
  "step": 3000
424
  },
425
  {
426
  "epoch": 16.0,
427
+ "eval_exact_match": 19.8,
428
+ "eval_f1": 31.533809523809534,
429
  "step": 3000
430
  },
431
  {
432
+ "epoch": 16.53,
 
433
  "learning_rate": 3e-05,
434
+ "loss": 0.4357,
435
  "step": 3100
436
  },
437
  {
438
+ "epoch": 17.0,
439
+ "eval_accuracy": 0.5975897435897436,
440
+ "eval_loss": 2.0656166076660156,
441
+ "eval_runtime": 4.1949,
442
+ "eval_samples_per_second": 119.193,
443
+ "eval_steps_per_second": 15.018,
444
  "step": 3187
445
  },
446
  {
447
+ "epoch": 17.0,
448
+ "eval_exact_match": 18.8,
449
+ "eval_f1": 31.295238095238105,
450
  "step": 3187
451
  },
452
  {
453
+ "epoch": 17.07,
 
454
  "learning_rate": 3e-05,
455
+ "loss": 0.4511,
456
  "step": 3200
457
  },
458
  {
459
  "epoch": 17.6,
 
460
  "learning_rate": 3e-05,
461
+ "loss": 0.4348,
462
  "step": 3300
463
  },
464
  {
465
  "epoch": 18.0,
466
+ "eval_accuracy": 0.5978974358974359,
467
+ "eval_loss": 2.030846118927002,
468
+ "eval_runtime": 4.1974,
469
+ "eval_samples_per_second": 119.122,
470
+ "eval_steps_per_second": 15.009,
471
  "step": 3375
472
  },
473
  {
474
  "epoch": 18.0,
475
+ "eval_exact_match": 19.6,
476
+ "eval_f1": 31.137619047619065,
477
  "step": 3375
478
  },
479
  {
480
+ "epoch": 18.13,
 
481
  "learning_rate": 3e-05,
482
+ "loss": 0.4436,
483
  "step": 3400
484
  },
485
  {
486
+ "epoch": 18.67,
 
487
  "learning_rate": 3e-05,
488
+ "loss": 0.4331,
489
  "step": 3500
490
  },
491
  {
492
+ "epoch": 19.0,
493
+ "eval_accuracy": 0.598974358974359,
494
+ "eval_loss": 2.062859058380127,
495
+ "eval_runtime": 4.7948,
496
+ "eval_samples_per_second": 104.279,
497
+ "eval_steps_per_second": 13.139,
498
  "step": 3562
499
  },
500
  {
501
+ "epoch": 19.0,
502
+ "eval_exact_match": 20.6,
503
+ "eval_f1": 32.42380952380953,
504
  "step": 3562
505
  },
506
  {
507
  "epoch": 19.2,
 
508
  "learning_rate": 3e-05,
509
+ "loss": 0.4338,
510
  "step": 3600
511
  },
512
  {
513
+ "epoch": 19.73,
 
514
  "learning_rate": 3e-05,
515
+ "loss": 0.4341,
516
  "step": 3700
517
  },
518
  {
519
  "epoch": 20.0,
520
+ "eval_accuracy": 0.5983076923076923,
521
+ "eval_loss": 2.0815415382385254,
522
+ "eval_runtime": 4.4041,
523
+ "eval_samples_per_second": 113.531,
524
+ "eval_steps_per_second": 14.305,
525
  "step": 3750
526
  },
527
  {
528
  "epoch": 20.0,
529
+ "eval_exact_match": 21.6,
530
+ "eval_f1": 33.345238095238116,
531
  "step": 3750
532
  },
533
  {
534
+ "epoch": 20.27,
 
535
  "learning_rate": 3e-05,
536
+ "loss": 0.4316,
537
  "step": 3800
538
  },
539
  {
540
  "epoch": 20.8,
 
541
  "learning_rate": 3e-05,
542
+ "loss": 0.434,
543
  "step": 3900
544
  },
545
  {
546
+ "epoch": 21.0,
547
+ "eval_accuracy": 0.5968205128205128,
548
+ "eval_loss": 2.1252710819244385,
549
+ "eval_runtime": 4.3531,
550
+ "eval_samples_per_second": 114.862,
551
+ "eval_steps_per_second": 14.473,
552
  "step": 3937
553
  },
554
  {
555
+ "epoch": 21.0,
556
+ "eval_exact_match": 21.4,
557
+ "eval_f1": 32.32285714285715,
558
  "step": 3937
559
  }
560
  ],
561
  "logging_steps": 100,
562
  "max_steps": 9350,
 
563
  "num_train_epochs": 50,
564
  "save_steps": 500,
565
+ "total_flos": 1.469200178777948e+17,
 
 
 
 
 
 
 
 
 
 
 
 
 
566
  "trial_name": null,
567
  "trial_params": null
568
  }
checkpoint-3937/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1a175de730fa6917b3a611bb9e53766640b7a5e4a426f825f2a602cf98661e47
3
- size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52c53f755ad4d3669bc1e92a0dd04ac5e6ef02b39406da28fac147ebfc373982
3
+ size 4728