tyzhu commited on
Commit
b2f25f8
1 Parent(s): 9902d25

Training in progress, epoch 13, checkpoint

Browse files
checkpoint-6877/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-6877/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49df17f9d21f5132ab027ba45caf8d6b504190933f32cb368830781cf0a5a7db
3
+ size 143269386
checkpoint-6877/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-6877/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1a8820a0ea1858420d418c5f5d87ca12ecbcd9dbba50e759a57b683a8be70d55
3
- size 224537202
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de9ab34b21509b011cec165edb94768c124dfe2f48d652a291b91a7dad504565
3
+ size 286585234
checkpoint-6877/rng_state_0.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f9b6721b40312270df8b6bd7b2b36022e55098cf838aad35826be256a3084fed
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b76ce5d2e85149914f210d38d17947f3d51beee48d83ae0cd13f90b96ab02032
3
  size 15024
checkpoint-6877/rng_state_1.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b1a41e64a1e2d0c6e56904c40a32958364ff99503ac13cc25672b0aab48b73ba
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57478f2a8e242519721af343d7b0d2b16b4055eedfa723d37dab4397d096da02
3
  size 15024
checkpoint-6877/rng_state_2.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a539b02b885bbb9bc80ff94139d636fb7ed593e7de6e821aa061fd25c6e66f71
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:911ae0b8bf50c3a0796e33cbc46e9f3a9ddee95a5bd4c7083f88de45d1a5fef2
3
  size 15024
checkpoint-6877/rng_state_3.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5173d25211bca94f9155be0759559e1119622385000e4b63806f064fa84e06b0
3
  size 15024
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bda4ec7d56e586309a8e2f345c3ee7cb218cb5ab34d37a9bf221e119514d7d7d
3
  size 15024
checkpoint-6877/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-6877/tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
 
checkpoint-6877/tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
checkpoint-6877/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-6877/trainer_state.json CHANGED
@@ -9,97 +9,87 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.1890359168241966,
13
- "grad_norm": 0.3393945097923279,
14
  "learning_rate": 0.0003,
15
- "loss": 1.951,
16
  "step": 100
17
  },
18
  {
19
- "epoch": 0.3780718336483932,
20
- "grad_norm": 8017.20361328125,
21
  "learning_rate": 0.0003,
22
- "loss": 2.0285,
23
  "step": 200
24
  },
25
  {
26
- "epoch": 0.5671077504725898,
27
- "grad_norm": 9146.7236328125,
28
  "learning_rate": 0.0003,
29
- "loss": 6.5404,
30
  "step": 300
31
  },
32
  {
33
- "epoch": 0.7561436672967864,
34
- "grad_norm": 574.934326171875,
35
  "learning_rate": 0.0003,
36
- "loss": 7.7357,
37
  "step": 400
38
  },
39
  {
40
- "epoch": 0.945179584120983,
41
- "grad_norm": 25128.55859375,
42
  "learning_rate": 0.0003,
43
- "loss": 6.3771,
44
  "step": 500
45
  },
46
  {
47
  "epoch": 1.0,
48
- "eval_accuracy": 0.24974358974358973,
49
- "eval_loss": 10.19681453704834,
50
- "eval_runtime": 5.1408,
51
- "eval_samples_per_second": 97.26,
52
- "eval_steps_per_second": 12.255,
53
  "step": 529
54
  },
55
  {
56
  "epoch": 1.0,
57
- "eval_exact_match": 0.0,
58
- "eval_f1": 0.0,
59
  "step": 529
60
  },
61
  {
62
- "epoch": 1.1342155009451795,
63
- "grad_norm": 1332.8885498046875,
64
  "learning_rate": 0.0003,
65
- "loss": 11.0858,
66
  "step": 600
67
  },
68
  {
69
- "epoch": 1.3232514177693762,
70
- "grad_norm": 3666.881591796875,
71
  "learning_rate": 0.0003,
72
- "loss": 9.483,
73
  "step": 700
74
  },
75
  {
76
- "epoch": 1.5122873345935728,
77
- "grad_norm": 356847.0,
78
  "learning_rate": 0.0003,
79
- "loss": 8.069,
80
  "step": 800
81
  },
82
  {
83
- "epoch": 1.7013232514177694,
84
- "grad_norm": 1061.877685546875,
85
  "learning_rate": 0.0003,
86
- "loss": 8.7493,
87
  "step": 900
88
  },
89
  {
90
- "epoch": 1.8903591682419658,
91
- "grad_norm": 46670.3046875,
92
  "learning_rate": 0.0003,
93
- "loss": 8.3415,
94
  "step": 1000
95
  },
96
  {
97
  "epoch": 2.0,
98
- "eval_accuracy": 0.20507692307692307,
99
- "eval_loss": 7.504091262817383,
100
- "eval_runtime": 5.2437,
101
- "eval_samples_per_second": 95.352,
102
- "eval_steps_per_second": 12.014,
103
  "step": 1058
104
  },
105
  {
@@ -109,204 +99,183 @@
109
  "step": 1058
110
  },
111
  {
112
- "epoch": 2.0793950850661624,
113
- "grad_norm": 2.0727381706237793,
114
  "learning_rate": 0.0003,
115
- "loss": 8.3511,
116
  "step": 1100
117
  },
118
  {
119
- "epoch": 2.268431001890359,
120
- "grad_norm": 1.8185893297195435,
121
  "learning_rate": 0.0003,
122
- "loss": 7.3065,
123
  "step": 1200
124
  },
125
  {
126
- "epoch": 2.4574669187145557,
127
- "grad_norm": 1620.93701171875,
128
  "learning_rate": 0.0003,
129
- "loss": 7.1302,
130
  "step": 1300
131
  },
132
  {
133
- "epoch": 2.6465028355387523,
134
- "grad_norm": 1.0691380500793457,
135
  "learning_rate": 0.0003,
136
- "loss": 7.4423,
137
  "step": 1400
138
  },
139
  {
140
- "epoch": 2.835538752362949,
141
- "grad_norm": 1.0605100393295288,
142
  "learning_rate": 0.0003,
143
- "loss": 7.6715,
144
  "step": 1500
145
  },
146
  {
147
  "epoch": 3.0,
148
- "eval_accuracy": 0.28507692307692306,
149
- "eval_loss": 7.084368705749512,
150
- "eval_runtime": 5.493,
151
- "eval_samples_per_second": 91.026,
152
- "eval_steps_per_second": 11.469,
153
  "step": 1587
154
  },
155
  {
156
  "epoch": 3.0,
157
  "eval_exact_match": 0.0,
158
- "eval_f1": 0.08,
159
  "step": 1587
160
  },
161
  {
162
- "epoch": 3.0245746691871456,
163
- "grad_norm": 1967676800.0,
164
  "learning_rate": 0.0003,
165
- "loss": 7.8617,
166
  "step": 1600
167
  },
168
  {
169
- "epoch": 3.213610586011342,
170
- "grad_norm": 83918778793984.0,
171
  "learning_rate": 0.0003,
172
- "loss": 8.083,
173
  "step": 1700
174
  },
175
  {
176
- "epoch": 3.402646502835539,
177
- "grad_norm": 12349449.0,
178
  "learning_rate": 0.0003,
179
- "loss": 8.0382,
180
  "step": 1800
181
  },
182
  {
183
- "epoch": 3.5916824196597354,
184
- "grad_norm": 52939337728.0,
185
  "learning_rate": 0.0003,
186
- "loss": 7.8761,
187
  "step": 1900
188
  },
189
  {
190
- "epoch": 3.780718336483932,
191
- "grad_norm": 1381182.25,
192
  "learning_rate": 0.0003,
193
- "loss": 7.4665,
194
  "step": 2000
195
  },
196
  {
197
- "epoch": 3.9697542533081287,
198
- "grad_norm": 105.80550384521484,
199
  "learning_rate": 0.0003,
200
- "loss": 7.8182,
201
  "step": 2100
202
  },
203
  {
204
  "epoch": 4.0,
205
- "eval_accuracy": 0.23933333333333334,
206
- "eval_loss": 7.05845308303833,
207
- "eval_runtime": 5.0891,
208
- "eval_samples_per_second": 98.248,
209
- "eval_steps_per_second": 12.379,
210
  "step": 2116
211
  },
212
  {
213
  "epoch": 4.0,
214
- "eval_exact_match": 0.0,
215
- "eval_f1": 0.0,
216
  "step": 2116
217
  },
218
  {
219
- "epoch": 4.158790170132325,
220
- "grad_norm": 16.110862731933594,
221
  "learning_rate": 0.0003,
222
- "loss": 7.9728,
223
  "step": 2200
224
  },
225
  {
226
- "epoch": 4.3478260869565215,
227
- "grad_norm": 4.362427234649658,
228
  "learning_rate": 0.0003,
229
- "loss": 7.2875,
230
  "step": 2300
231
  },
232
  {
233
- "epoch": 4.536862003780718,
234
- "grad_norm": 18039.095703125,
235
  "learning_rate": 0.0003,
236
- "loss": 7.3972,
237
  "step": 2400
238
  },
239
  {
240
- "epoch": 4.725897920604915,
241
- "grad_norm": 2.3812594051160474e+17,
242
  "learning_rate": 0.0003,
243
- "loss": 7.4631,
244
  "step": 2500
245
  },
246
  {
247
- "epoch": 4.914933837429111,
248
- "grad_norm": 3.9549436569213867,
249
  "learning_rate": 0.0003,
250
- "loss": 7.3322,
251
  "step": 2600
252
  },
253
  {
254
  "epoch": 5.0,
255
- "eval_accuracy": 0.22430769230769232,
256
- "eval_loss": 6.72421932220459,
257
- "eval_runtime": 5.1689,
258
- "eval_samples_per_second": 96.731,
259
- "eval_steps_per_second": 12.188,
260
  "step": 2645
261
  },
262
  {
263
  "epoch": 5.0,
264
  "eval_exact_match": 0.0,
265
- "eval_f1": 0.0,
266
  "step": 2645
267
  },
268
  {
269
- "epoch": 5.103969754253308,
270
- "grad_norm": 120114.2421875,
271
  "learning_rate": 0.0003,
272
- "loss": 7.2428,
273
  "step": 2700
274
  },
275
  {
276
- "epoch": 5.293005671077505,
277
- "grad_norm": 391.66119384765625,
278
  "learning_rate": 0.0003,
279
- "loss": 7.2495,
280
  "step": 2800
281
  },
282
  {
283
- "epoch": 5.482041587901701,
284
- "grad_norm": 599371136.0,
285
  "learning_rate": 0.0003,
286
- "loss": 7.5915,
287
  "step": 2900
288
  },
289
  {
290
- "epoch": 5.671077504725898,
291
- "grad_norm": 5585.97314453125,
292
  "learning_rate": 0.0003,
293
- "loss": 7.3014,
294
  "step": 3000
295
  },
296
  {
297
- "epoch": 5.8601134215500945,
298
- "grad_norm": 268239548448768.0,
299
  "learning_rate": 0.0003,
300
- "loss": 7.259,
301
  "step": 3100
302
  },
303
  {
304
  "epoch": 6.0,
305
- "eval_accuracy": 0.24984615384615386,
306
- "eval_loss": 6.531904220581055,
307
- "eval_runtime": 5.0344,
308
- "eval_samples_per_second": 99.317,
309
- "eval_steps_per_second": 12.514,
310
  "step": 3174
311
  },
312
  {
@@ -316,54 +285,48 @@
316
  "step": 3174
317
  },
318
  {
319
- "epoch": 6.049149338374291,
320
- "grad_norm": 21535.734375,
321
  "learning_rate": 0.0003,
322
- "loss": 7.5829,
323
  "step": 3200
324
  },
325
  {
326
- "epoch": 6.238185255198488,
327
- "grad_norm": 40871530496.0,
328
  "learning_rate": 0.0003,
329
- "loss": 8.6364,
330
  "step": 3300
331
  },
332
  {
333
- "epoch": 6.427221172022684,
334
- "grad_norm": 13.242510795593262,
335
  "learning_rate": 0.0003,
336
- "loss": 7.6599,
337
  "step": 3400
338
  },
339
  {
340
- "epoch": 6.616257088846881,
341
- "grad_norm": 3.0841150283813477,
342
  "learning_rate": 0.0003,
343
- "loss": 7.5796,
344
  "step": 3500
345
  },
346
  {
347
- "epoch": 6.805293005671078,
348
- "grad_norm": 1.6124571561813354,
349
  "learning_rate": 0.0003,
350
- "loss": 7.3499,
351
  "step": 3600
352
  },
353
  {
354
- "epoch": 6.994328922495274,
355
- "grad_norm": 1.8358014822006226,
356
  "learning_rate": 0.0003,
357
- "loss": 7.2206,
358
  "step": 3700
359
  },
360
  {
361
  "epoch": 7.0,
362
- "eval_accuracy": 0.22430769230769232,
363
- "eval_loss": 6.724715709686279,
364
- "eval_runtime": 5.0476,
365
- "eval_samples_per_second": 99.056,
366
- "eval_steps_per_second": 12.481,
367
  "step": 3703
368
  },
369
  {
@@ -373,47 +336,42 @@
373
  "step": 3703
374
  },
375
  {
376
- "epoch": 7.183364839319471,
377
- "grad_norm": 5.786651134490967,
378
  "learning_rate": 0.0003,
379
- "loss": 7.2316,
380
  "step": 3800
381
  },
382
  {
383
- "epoch": 7.3724007561436675,
384
- "grad_norm": 3.7182490825653076,
385
  "learning_rate": 0.0003,
386
- "loss": 7.2591,
387
  "step": 3900
388
  },
389
  {
390
- "epoch": 7.561436672967864,
391
- "grad_norm": 31.241104125976562,
392
  "learning_rate": 0.0003,
393
- "loss": 7.2557,
394
  "step": 4000
395
  },
396
  {
397
- "epoch": 7.750472589792061,
398
- "grad_norm": 896688896.0,
399
  "learning_rate": 0.0003,
400
- "loss": 8.4436,
401
  "step": 4100
402
  },
403
  {
404
- "epoch": 7.939508506616257,
405
- "grad_norm": 1.1150538843415904e+19,
406
  "learning_rate": 0.0003,
407
- "loss": 8.7572,
408
  "step": 4200
409
  },
410
  {
411
  "epoch": 8.0,
412
- "eval_accuracy": 0.26184615384615384,
413
- "eval_loss": 6.479637622833252,
414
- "eval_runtime": 5.1473,
415
- "eval_samples_per_second": 97.138,
416
- "eval_steps_per_second": 12.239,
417
  "step": 4232
418
  },
419
  {
@@ -423,47 +381,42 @@
423
  "step": 4232
424
  },
425
  {
426
- "epoch": 8.128544423440454,
427
- "grad_norm": 7.2513813972473145,
428
  "learning_rate": 0.0003,
429
- "loss": 8.1381,
430
  "step": 4300
431
  },
432
  {
433
- "epoch": 8.31758034026465,
434
- "grad_norm": 14.863628387451172,
435
  "learning_rate": 0.0003,
436
- "loss": 7.3335,
437
  "step": 4400
438
  },
439
  {
440
- "epoch": 8.506616257088847,
441
- "grad_norm": 5.358551979064941,
442
  "learning_rate": 0.0003,
443
- "loss": 8.3921,
444
  "step": 4500
445
  },
446
  {
447
- "epoch": 8.695652173913043,
448
- "grad_norm": 113.71702575683594,
449
  "learning_rate": 0.0003,
450
- "loss": 7.221,
451
  "step": 4600
452
  },
453
  {
454
- "epoch": 8.88468809073724,
455
- "grad_norm": 2.3997488021850586,
456
  "learning_rate": 0.0003,
457
- "loss": 8.4835,
458
  "step": 4700
459
  },
460
  {
461
  "epoch": 9.0,
462
- "eval_accuracy": 0.26184615384615384,
463
- "eval_loss": 6.614655017852783,
464
- "eval_runtime": 5.026,
465
- "eval_samples_per_second": 99.483,
466
- "eval_steps_per_second": 12.535,
467
  "step": 4761
468
  },
469
  {
@@ -473,47 +426,42 @@
473
  "step": 4761
474
  },
475
  {
476
- "epoch": 9.073724007561436,
477
- "grad_norm": 7.822358131408691,
478
  "learning_rate": 0.0003,
479
- "loss": 7.2836,
480
  "step": 4800
481
  },
482
  {
483
- "epoch": 9.262759924385634,
484
- "grad_norm": 89279359352832.0,
485
  "learning_rate": 0.0003,
486
- "loss": 7.2712,
487
  "step": 4900
488
  },
489
  {
490
- "epoch": 9.45179584120983,
491
- "grad_norm": 656.6845703125,
492
  "learning_rate": 0.0003,
493
- "loss": 7.2623,
494
  "step": 5000
495
  },
496
  {
497
- "epoch": 9.640831758034027,
498
- "grad_norm": 13841700.0,
499
  "learning_rate": 0.0003,
500
- "loss": 7.4981,
501
  "step": 5100
502
  },
503
  {
504
- "epoch": 9.829867674858223,
505
- "grad_norm": 768177.9375,
506
  "learning_rate": 0.0003,
507
- "loss": 7.6044,
508
  "step": 5200
509
  },
510
  {
511
  "epoch": 10.0,
512
- "eval_accuracy": 0.2508717948717949,
513
- "eval_loss": 8.167869567871094,
514
- "eval_runtime": 5.1342,
515
- "eval_samples_per_second": 97.387,
516
- "eval_steps_per_second": 12.271,
517
  "step": 5290
518
  },
519
  {
@@ -523,54 +471,48 @@
523
  "step": 5290
524
  },
525
  {
526
- "epoch": 10.01890359168242,
527
- "grad_norm": 389293.03125,
528
  "learning_rate": 0.0003,
529
- "loss": 7.9142,
530
  "step": 5300
531
  },
532
  {
533
- "epoch": 10.207939508506616,
534
- "grad_norm": 3129429.25,
535
  "learning_rate": 0.0003,
536
- "loss": 8.998,
537
  "step": 5400
538
  },
539
  {
540
- "epoch": 10.396975425330814,
541
- "grad_norm": 6832450.5,
542
  "learning_rate": 0.0003,
543
- "loss": 8.9548,
544
  "step": 5500
545
  },
546
  {
547
- "epoch": 10.58601134215501,
548
- "grad_norm": 24994486.0,
549
  "learning_rate": 0.0003,
550
- "loss": 9.3085,
551
  "step": 5600
552
  },
553
  {
554
- "epoch": 10.775047258979207,
555
- "grad_norm": 10634.3076171875,
556
  "learning_rate": 0.0003,
557
- "loss": 9.3052,
558
  "step": 5700
559
  },
560
  {
561
- "epoch": 10.964083175803403,
562
- "grad_norm": 37.100101470947266,
563
  "learning_rate": 0.0003,
564
- "loss": 10.0839,
565
  "step": 5800
566
  },
567
  {
568
  "epoch": 11.0,
569
- "eval_accuracy": 0.2362051282051282,
570
- "eval_loss": 6.820600509643555,
571
- "eval_runtime": 5.0262,
572
- "eval_samples_per_second": 99.478,
573
- "eval_steps_per_second": 12.534,
574
  "step": 5819
575
  },
576
  {
@@ -580,47 +522,42 @@
580
  "step": 5819
581
  },
582
  {
583
- "epoch": 11.1531190926276,
584
- "grad_norm": 319.2048645019531,
585
  "learning_rate": 0.0003,
586
- "loss": 7.4258,
587
  "step": 5900
588
  },
589
  {
590
- "epoch": 11.342155009451796,
591
- "grad_norm": 72.28472900390625,
592
  "learning_rate": 0.0003,
593
- "loss": 7.2609,
594
  "step": 6000
595
  },
596
  {
597
- "epoch": 11.531190926275993,
598
- "grad_norm": 11.851283073425293,
599
  "learning_rate": 0.0003,
600
- "loss": 7.5323,
601
  "step": 6100
602
  },
603
  {
604
- "epoch": 11.720226843100189,
605
- "grad_norm": 6.574771404266357,
606
  "learning_rate": 0.0003,
607
- "loss": 7.3028,
608
  "step": 6200
609
  },
610
  {
611
- "epoch": 11.909262759924385,
612
- "grad_norm": 4.578986644744873,
613
  "learning_rate": 0.0003,
614
- "loss": 7.2957,
615
  "step": 6300
616
  },
617
  {
618
  "epoch": 12.0,
619
- "eval_accuracy": 0.26184615384615384,
620
- "eval_loss": 6.52316427230835,
621
- "eval_runtime": 5.0329,
622
- "eval_samples_per_second": 99.346,
623
- "eval_steps_per_second": 12.518,
624
  "step": 6348
625
  },
626
  {
@@ -630,47 +567,42 @@
630
  "step": 6348
631
  },
632
  {
633
- "epoch": 12.098298676748582,
634
- "grad_norm": 21.886821746826172,
635
  "learning_rate": 0.0003,
636
- "loss": 7.189,
637
  "step": 6400
638
  },
639
  {
640
- "epoch": 12.287334593572778,
641
- "grad_norm": 102.8978500366211,
642
  "learning_rate": 0.0003,
643
- "loss": 7.2076,
644
  "step": 6500
645
  },
646
  {
647
- "epoch": 12.476370510396976,
648
- "grad_norm": 4.47935676574707,
649
  "learning_rate": 0.0003,
650
- "loss": 7.2369,
651
  "step": 6600
652
  },
653
  {
654
- "epoch": 12.665406427221171,
655
- "grad_norm": 259.2627868652344,
656
  "learning_rate": 0.0003,
657
- "loss": 7.3119,
658
  "step": 6700
659
  },
660
  {
661
- "epoch": 12.854442344045369,
662
- "grad_norm": 473.829345703125,
663
  "learning_rate": 0.0003,
664
- "loss": 7.3484,
665
  "step": 6800
666
  },
667
  {
668
  "epoch": 13.0,
669
- "eval_accuracy": 0.26184615384615384,
670
- "eval_loss": 6.655651569366455,
671
- "eval_runtime": 5.7482,
672
- "eval_samples_per_second": 86.984,
673
- "eval_steps_per_second": 10.96,
674
  "step": 6877
675
  },
676
  {
@@ -682,23 +614,9 @@
682
  ],
683
  "logging_steps": 100,
684
  "max_steps": 26450,
685
- "num_input_tokens_seen": 0,
686
  "num_train_epochs": 50,
687
  "save_steps": 500,
688
- "stateful_callbacks": {
689
- "TrainerControl": {
690
- "args": {
691
- "should_epoch_stop": false,
692
- "should_evaluate": false,
693
- "should_log": false,
694
- "should_save": true,
695
- "should_training_stop": false
696
- },
697
- "attributes": {}
698
- }
699
- },
700
- "total_flos": 5.597990780372255e+17,
701
- "train_batch_size": 2,
702
  "trial_name": null,
703
  "trial_params": null
704
  }
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.19,
 
13
  "learning_rate": 0.0003,
14
+ "loss": 1.4758,
15
  "step": 100
16
  },
17
  {
18
+ "epoch": 0.38,
 
19
  "learning_rate": 0.0003,
20
+ "loss": 1.3867,
21
  "step": 200
22
  },
23
  {
24
+ "epoch": 0.57,
 
25
  "learning_rate": 0.0003,
26
+ "loss": 1.4064,
27
  "step": 300
28
  },
29
  {
30
+ "epoch": 0.76,
 
31
  "learning_rate": 0.0003,
32
+ "loss": 1.3789,
33
  "step": 400
34
  },
35
  {
36
+ "epoch": 0.95,
 
37
  "learning_rate": 0.0003,
38
+ "loss": 1.3822,
39
  "step": 500
40
  },
41
  {
42
  "epoch": 1.0,
43
+ "eval_accuracy": 0.6171794871794872,
44
+ "eval_loss": 1.2976911067962646,
45
+ "eval_runtime": 4.224,
46
+ "eval_samples_per_second": 118.372,
47
+ "eval_steps_per_second": 14.915,
48
  "step": 529
49
  },
50
  {
51
  "epoch": 1.0,
52
+ "eval_exact_match": 0.2,
53
+ "eval_f1": 2.031878046337681,
54
  "step": 529
55
  },
56
  {
57
+ "epoch": 1.13,
 
58
  "learning_rate": 0.0003,
59
+ "loss": 1.3048,
60
  "step": 600
61
  },
62
  {
63
+ "epoch": 1.32,
 
64
  "learning_rate": 0.0003,
65
+ "loss": 1.2524,
66
  "step": 700
67
  },
68
  {
69
+ "epoch": 1.51,
 
70
  "learning_rate": 0.0003,
71
+ "loss": 1.3009,
72
  "step": 800
73
  },
74
  {
75
+ "epoch": 1.7,
 
76
  "learning_rate": 0.0003,
77
+ "loss": 1.268,
78
  "step": 900
79
  },
80
  {
81
+ "epoch": 1.89,
 
82
  "learning_rate": 0.0003,
83
+ "loss": 1.2744,
84
  "step": 1000
85
  },
86
  {
87
  "epoch": 2.0,
88
+ "eval_accuracy": 0.6032307692307692,
89
+ "eval_loss": 1.374481201171875,
90
+ "eval_runtime": 4.4141,
91
+ "eval_samples_per_second": 113.274,
92
+ "eval_steps_per_second": 14.272,
93
  "step": 1058
94
  },
95
  {
 
99
  "step": 1058
100
  },
101
  {
102
+ "epoch": 2.08,
 
103
  "learning_rate": 0.0003,
104
+ "loss": 1.2158,
105
  "step": 1100
106
  },
107
  {
108
+ "epoch": 2.27,
 
109
  "learning_rate": 0.0003,
110
+ "loss": 1.1066,
111
  "step": 1200
112
  },
113
  {
114
+ "epoch": 2.46,
 
115
  "learning_rate": 0.0003,
116
+ "loss": 1.1365,
117
  "step": 1300
118
  },
119
  {
120
+ "epoch": 2.65,
 
121
  "learning_rate": 0.0003,
122
+ "loss": 1.1486,
123
  "step": 1400
124
  },
125
  {
126
+ "epoch": 2.84,
 
127
  "learning_rate": 0.0003,
128
+ "loss": 1.1768,
129
  "step": 1500
130
  },
131
  {
132
  "epoch": 3.0,
133
+ "eval_accuracy": 0.6157435897435898,
134
+ "eval_loss": 1.331903100013733,
135
+ "eval_runtime": 4.8466,
136
+ "eval_samples_per_second": 103.164,
137
+ "eval_steps_per_second": 12.999,
138
  "step": 1587
139
  },
140
  {
141
  "epoch": 3.0,
142
  "eval_exact_match": 0.0,
143
+ "eval_f1": 0.0,
144
  "step": 1587
145
  },
146
  {
147
+ "epoch": 3.02,
 
148
  "learning_rate": 0.0003,
149
+ "loss": 1.1769,
150
  "step": 1600
151
  },
152
  {
153
+ "epoch": 3.21,
 
154
  "learning_rate": 0.0003,
155
+ "loss": 0.9541,
156
  "step": 1700
157
  },
158
  {
159
+ "epoch": 3.4,
 
160
  "learning_rate": 0.0003,
161
+ "loss": 1.2939,
162
  "step": 1800
163
  },
164
  {
165
+ "epoch": 3.59,
 
166
  "learning_rate": 0.0003,
167
+ "loss": 0.9852,
168
  "step": 1900
169
  },
170
  {
171
+ "epoch": 3.78,
 
172
  "learning_rate": 0.0003,
173
+ "loss": 0.9538,
174
  "step": 2000
175
  },
176
  {
177
+ "epoch": 3.97,
 
178
  "learning_rate": 0.0003,
179
+ "loss": 0.9247,
180
  "step": 2100
181
  },
182
  {
183
  "epoch": 4.0,
184
+ "eval_accuracy": 0.6101538461538462,
185
+ "eval_loss": 1.4367107152938843,
186
+ "eval_runtime": 4.4146,
187
+ "eval_samples_per_second": 113.259,
188
+ "eval_steps_per_second": 14.271,
189
  "step": 2116
190
  },
191
  {
192
  "epoch": 4.0,
193
+ "eval_exact_match": 6.6,
194
+ "eval_f1": 9.897142857142855,
195
  "step": 2116
196
  },
197
  {
198
+ "epoch": 4.16,
 
199
  "learning_rate": 0.0003,
200
+ "loss": 1.1334,
201
  "step": 2200
202
  },
203
  {
204
+ "epoch": 4.35,
 
205
  "learning_rate": 0.0003,
206
+ "loss": 1.5175,
207
  "step": 2300
208
  },
209
  {
210
+ "epoch": 4.54,
 
211
  "learning_rate": 0.0003,
212
+ "loss": 4.2745,
213
  "step": 2400
214
  },
215
  {
216
+ "epoch": 4.73,
 
217
  "learning_rate": 0.0003,
218
+ "loss": 2.5667,
219
  "step": 2500
220
  },
221
  {
222
+ "epoch": 4.91,
 
223
  "learning_rate": 0.0003,
224
+ "loss": 1.1836,
225
  "step": 2600
226
  },
227
  {
228
  "epoch": 5.0,
229
+ "eval_accuracy": 0.556923076923077,
230
+ "eval_loss": 1.9167817831039429,
231
+ "eval_runtime": 4.8432,
232
+ "eval_samples_per_second": 103.238,
233
+ "eval_steps_per_second": 13.008,
234
  "step": 2645
235
  },
236
  {
237
  "epoch": 5.0,
238
  "eval_exact_match": 0.0,
239
+ "eval_f1": 0.2264729350912329,
240
  "step": 2645
241
  },
242
  {
243
+ "epoch": 5.1,
 
244
  "learning_rate": 0.0003,
245
+ "loss": 1.7226,
246
  "step": 2700
247
  },
248
  {
249
+ "epoch": 5.29,
 
250
  "learning_rate": 0.0003,
251
+ "loss": 2.1466,
252
  "step": 2800
253
  },
254
  {
255
+ "epoch": 5.48,
 
256
  "learning_rate": 0.0003,
257
+ "loss": 1.83,
258
  "step": 2900
259
  },
260
  {
261
+ "epoch": 5.67,
 
262
  "learning_rate": 0.0003,
263
+ "loss": 1.8789,
264
  "step": 3000
265
  },
266
  {
267
+ "epoch": 5.86,
 
268
  "learning_rate": 0.0003,
269
+ "loss": 2.035,
270
  "step": 3100
271
  },
272
  {
273
  "epoch": 6.0,
274
+ "eval_accuracy": 0.5377435897435897,
275
+ "eval_loss": 2.0794410705566406,
276
+ "eval_runtime": 4.1284,
277
+ "eval_samples_per_second": 121.112,
278
+ "eval_steps_per_second": 15.26,
279
  "step": 3174
280
  },
281
  {
 
285
  "step": 3174
286
  },
287
  {
288
+ "epoch": 6.05,
 
289
  "learning_rate": 0.0003,
290
+ "loss": 2.2251,
291
  "step": 3200
292
  },
293
  {
294
+ "epoch": 6.24,
 
295
  "learning_rate": 0.0003,
296
+ "loss": 2.1457,
297
  "step": 3300
298
  },
299
  {
300
+ "epoch": 6.43,
 
301
  "learning_rate": 0.0003,
302
+ "loss": 2.0764,
303
  "step": 3400
304
  },
305
  {
306
+ "epoch": 6.62,
 
307
  "learning_rate": 0.0003,
308
+ "loss": 2.3695,
309
  "step": 3500
310
  },
311
  {
312
+ "epoch": 6.81,
 
313
  "learning_rate": 0.0003,
314
+ "loss": 2.2254,
315
  "step": 3600
316
  },
317
  {
318
+ "epoch": 6.99,
 
319
  "learning_rate": 0.0003,
320
+ "loss": 3.7483,
321
  "step": 3700
322
  },
323
  {
324
  "epoch": 7.0,
325
+ "eval_accuracy": 0.48805128205128206,
326
+ "eval_loss": 2.6722612380981445,
327
+ "eval_runtime": 4.2095,
328
+ "eval_samples_per_second": 118.778,
329
+ "eval_steps_per_second": 14.966,
330
  "step": 3703
331
  },
332
  {
 
336
  "step": 3703
337
  },
338
  {
339
+ "epoch": 7.18,
 
340
  "learning_rate": 0.0003,
341
+ "loss": 6.5533,
342
  "step": 3800
343
  },
344
  {
345
+ "epoch": 7.37,
 
346
  "learning_rate": 0.0003,
347
+ "loss": 6.708,
348
  "step": 3900
349
  },
350
  {
351
+ "epoch": 7.56,
 
352
  "learning_rate": 0.0003,
353
+ "loss": 7.4301,
354
  "step": 4000
355
  },
356
  {
357
+ "epoch": 7.75,
 
358
  "learning_rate": 0.0003,
359
+ "loss": 8.6731,
360
  "step": 4100
361
  },
362
  {
363
+ "epoch": 7.94,
 
364
  "learning_rate": 0.0003,
365
+ "loss": 7.127,
366
  "step": 4200
367
  },
368
  {
369
  "epoch": 8.0,
370
+ "eval_accuracy": 0.1922051282051282,
371
+ "eval_loss": 7.040984153747559,
372
+ "eval_runtime": 4.8262,
373
+ "eval_samples_per_second": 103.6,
374
+ "eval_steps_per_second": 13.054,
375
  "step": 4232
376
  },
377
  {
 
381
  "step": 4232
382
  },
383
  {
384
+ "epoch": 8.13,
 
385
  "learning_rate": 0.0003,
386
+ "loss": 7.8796,
387
  "step": 4300
388
  },
389
  {
390
+ "epoch": 8.32,
 
391
  "learning_rate": 0.0003,
392
+ "loss": 8.0716,
393
  "step": 4400
394
  },
395
  {
396
+ "epoch": 8.51,
 
397
  "learning_rate": 0.0003,
398
+ "loss": 8.2395,
399
  "step": 4500
400
  },
401
  {
402
+ "epoch": 8.7,
 
403
  "learning_rate": 0.0003,
404
+ "loss": 7.9157,
405
  "step": 4600
406
  },
407
  {
408
+ "epoch": 8.88,
 
409
  "learning_rate": 0.0003,
410
+ "loss": 7.5321,
411
  "step": 4700
412
  },
413
  {
414
  "epoch": 9.0,
415
+ "eval_accuracy": 0.19405128205128205,
416
+ "eval_loss": 6.64878511428833,
417
+ "eval_runtime": 4.104,
418
+ "eval_samples_per_second": 121.831,
419
+ "eval_steps_per_second": 15.351,
420
  "step": 4761
421
  },
422
  {
 
426
  "step": 4761
427
  },
428
  {
429
+ "epoch": 9.07,
 
430
  "learning_rate": 0.0003,
431
+ "loss": 7.2432,
432
  "step": 4800
433
  },
434
  {
435
+ "epoch": 9.26,
 
436
  "learning_rate": 0.0003,
437
+ "loss": 7.4268,
438
  "step": 4900
439
  },
440
  {
441
+ "epoch": 9.45,
 
442
  "learning_rate": 0.0003,
443
+ "loss": 7.7517,
444
  "step": 5000
445
  },
446
  {
447
+ "epoch": 9.64,
 
448
  "learning_rate": 0.0003,
449
+ "loss": 7.4674,
450
  "step": 5100
451
  },
452
  {
453
+ "epoch": 9.83,
 
454
  "learning_rate": 0.0003,
455
+ "loss": 7.3806,
456
  "step": 5200
457
  },
458
  {
459
  "epoch": 10.0,
460
+ "eval_accuracy": 0.2196923076923077,
461
+ "eval_loss": 6.842738151550293,
462
+ "eval_runtime": 4.8429,
463
+ "eval_samples_per_second": 103.244,
464
+ "eval_steps_per_second": 13.009,
465
  "step": 5290
466
  },
467
  {
 
471
  "step": 5290
472
  },
473
  {
474
+ "epoch": 10.02,
 
475
  "learning_rate": 0.0003,
476
+ "loss": 7.4491,
477
  "step": 5300
478
  },
479
  {
480
+ "epoch": 10.21,
 
481
  "learning_rate": 0.0003,
482
+ "loss": 7.4851,
483
  "step": 5400
484
  },
485
  {
486
+ "epoch": 10.4,
 
487
  "learning_rate": 0.0003,
488
+ "loss": 7.5747,
489
  "step": 5500
490
  },
491
  {
492
+ "epoch": 10.59,
 
493
  "learning_rate": 0.0003,
494
+ "loss": 7.7881,
495
  "step": 5600
496
  },
497
  {
498
+ "epoch": 10.78,
 
499
  "learning_rate": 0.0003,
500
+ "loss": 7.8095,
501
  "step": 5700
502
  },
503
  {
504
+ "epoch": 10.96,
 
505
  "learning_rate": 0.0003,
506
+ "loss": 7.8159,
507
  "step": 5800
508
  },
509
  {
510
  "epoch": 11.0,
511
+ "eval_accuracy": 0.2196923076923077,
512
+ "eval_loss": 6.883613586425781,
513
+ "eval_runtime": 4.1075,
514
+ "eval_samples_per_second": 121.727,
515
+ "eval_steps_per_second": 15.338,
516
  "step": 5819
517
  },
518
  {
 
522
  "step": 5819
523
  },
524
  {
525
+ "epoch": 11.15,
 
526
  "learning_rate": 0.0003,
527
+ "loss": 7.7222,
528
  "step": 5900
529
  },
530
  {
531
+ "epoch": 11.34,
 
532
  "learning_rate": 0.0003,
533
+ "loss": 7.6203,
534
  "step": 6000
535
  },
536
  {
537
+ "epoch": 11.53,
 
538
  "learning_rate": 0.0003,
539
+ "loss": 7.8391,
540
  "step": 6100
541
  },
542
  {
543
+ "epoch": 11.72,
 
544
  "learning_rate": 0.0003,
545
+ "loss": 7.8888,
546
  "step": 6200
547
  },
548
  {
549
+ "epoch": 11.91,
 
550
  "learning_rate": 0.0003,
551
+ "loss": 7.975,
552
  "step": 6300
553
  },
554
  {
555
  "epoch": 12.0,
556
+ "eval_accuracy": 0.2196923076923077,
557
+ "eval_loss": 6.876262187957764,
558
+ "eval_runtime": 4.4184,
559
+ "eval_samples_per_second": 113.162,
560
+ "eval_steps_per_second": 14.258,
561
  "step": 6348
562
  },
563
  {
 
567
  "step": 6348
568
  },
569
  {
570
+ "epoch": 12.1,
 
571
  "learning_rate": 0.0003,
572
+ "loss": 7.9981,
573
  "step": 6400
574
  },
575
  {
576
+ "epoch": 12.29,
 
577
  "learning_rate": 0.0003,
578
+ "loss": 7.9944,
579
  "step": 6500
580
  },
581
  {
582
+ "epoch": 12.48,
 
583
  "learning_rate": 0.0003,
584
+ "loss": 8.0191,
585
  "step": 6600
586
  },
587
  {
588
+ "epoch": 12.67,
 
589
  "learning_rate": 0.0003,
590
+ "loss": 8.0091,
591
  "step": 6700
592
  },
593
  {
594
+ "epoch": 12.85,
 
595
  "learning_rate": 0.0003,
596
+ "loss": 7.9902,
597
  "step": 6800
598
  },
599
  {
600
  "epoch": 13.0,
601
+ "eval_accuracy": 0.2196923076923077,
602
+ "eval_loss": 6.872635841369629,
603
+ "eval_runtime": 4.943,
604
+ "eval_samples_per_second": 101.152,
605
+ "eval_steps_per_second": 12.745,
606
  "step": 6877
607
  },
608
  {
 
614
  ],
615
  "logging_steps": 100,
616
  "max_steps": 26450,
 
617
  "num_train_epochs": 50,
618
  "save_steps": 500,
619
+ "total_flos": 1.1617845809994793e+18,
 
 
 
 
 
 
 
 
 
 
 
 
 
620
  "trial_name": null,
621
  "trial_params": null
622
  }
checkpoint-6877/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:186f3ae462881ca0bb3a6b5295a97a245e0a78172be9c14948a2d07a12aa4241
3
- size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b59ca157797b57867ac15a1da8ca60d782441a1506bb7cb43703c7748b856900
3
+ size 4728