Hartunka commited on
Commit
46f2ef8
·
verified ·
1 Parent(s): a9dae29

End of training

Browse files
Files changed (5) hide show
  1. README.md +16 -4
  2. all_results.json +16 -0
  3. eval_results.json +10 -0
  4. train_results.json +9 -0
  5. trainer_state.json +921 -0
README.md CHANGED
@@ -2,11 +2,23 @@
2
  library_name: transformers
3
  tags:
4
  - generated_from_trainer
 
 
5
  metrics:
6
  - accuracy
7
  model-index:
8
  - name: distilbert_rand_20_v2
9
- results: []
 
 
 
 
 
 
 
 
 
 
10
  ---
11
 
12
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -14,10 +26,10 @@ should probably proofread and complete it, then remove this comment. -->
14
 
15
  # distilbert_rand_20_v2
16
 
17
- This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 12.8140
20
- - Accuracy: 0.1511
21
 
22
  ## Model description
23
 
 
2
  library_name: transformers
3
  tags:
4
  - generated_from_trainer
5
+ datasets:
6
+ - Hartunka/processed_wikitext-103-raw-v1-rand-20_v2
7
  metrics:
8
  - accuracy
9
  model-index:
10
  - name: distilbert_rand_20_v2
11
+ results:
12
+ - task:
13
+ name: Masked Language Modeling
14
+ type: fill-mask
15
+ dataset:
16
+ name: Hartunka/processed_wikitext-103-raw-v1-rand-20_v2
17
+ type: Hartunka/processed_wikitext-103-raw-v1-rand-20_v2
18
+ metrics:
19
+ - name: Accuracy
20
+ type: accuracy
21
+ value: 0.15333107306323104
22
  ---
23
 
24
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
26
 
27
  # distilbert_rand_20_v2
28
 
29
+ This model is a fine-tuned version of [](https://huggingface.co/) on the Hartunka/processed_wikitext-103-raw-v1-rand-20_v2 dataset.
30
  It achieves the following results on the evaluation set:
31
+ - Loss: 9.1981
32
+ - Accuracy: 0.1533
33
 
34
  ## Model description
35
 
all_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 25.0,
3
+ "eval_accuracy": 0.15333107306323104,
4
+ "eval_loss": 9.198078155517578,
5
+ "eval_runtime": 0.8079,
6
+ "eval_samples": 479,
7
+ "eval_samples_per_second": 592.86,
8
+ "eval_steps_per_second": 6.189,
9
+ "perplexity": 9878.126581622239,
10
+ "total_flos": 7.68356804521728e+17,
11
+ "train_loss": 7.6570130803823275,
12
+ "train_runtime": 16738.0825,
13
+ "train_samples": 228639,
14
+ "train_samples_per_second": 341.495,
15
+ "train_steps_per_second": 3.558
16
+ }
eval_results.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 25.0,
3
+ "eval_accuracy": 0.15333107306323104,
4
+ "eval_loss": 9.198078155517578,
5
+ "eval_runtime": 0.8079,
6
+ "eval_samples": 479,
7
+ "eval_samples_per_second": 592.86,
8
+ "eval_steps_per_second": 6.189,
9
+ "perplexity": 9878.126581622239
10
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 25.0,
3
+ "total_flos": 7.68356804521728e+17,
4
+ "train_loss": 7.6570130803823275,
5
+ "train_runtime": 16738.0825,
6
+ "train_samples": 228639,
7
+ "train_samples_per_second": 341.495,
8
+ "train_steps_per_second": 3.558
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,921 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 10000,
3
+ "best_metric": 9.169220924377441,
4
+ "best_model_checkpoint": "distilbert_rand_20_v2/checkpoint-10000",
5
+ "epoch": 25.0,
6
+ "eval_steps": 10000,
7
+ "global_step": 59550,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.20990764063811923,
14
+ "grad_norm": 2.839494466781616,
15
+ "learning_rate": 5e-06,
16
+ "loss": 12.295,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.41981528127623846,
21
+ "grad_norm": 2.4242842197418213,
22
+ "learning_rate": 1e-05,
23
+ "loss": 10.5891,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.6297229219143576,
28
+ "grad_norm": 1.7410306930541992,
29
+ "learning_rate": 1.5e-05,
30
+ "loss": 10.0631,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 0.8396305625524769,
35
+ "grad_norm": 1.7641091346740723,
36
+ "learning_rate": 2e-05,
37
+ "loss": 9.9119,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 1.0495382031905962,
42
+ "grad_norm": 2.0337624549865723,
43
+ "learning_rate": 2.5e-05,
44
+ "loss": 9.7769,
45
+ "step": 2500
46
+ },
47
+ {
48
+ "epoch": 1.2594458438287153,
49
+ "grad_norm": 2.2860662937164307,
50
+ "learning_rate": 3e-05,
51
+ "loss": 9.6625,
52
+ "step": 3000
53
+ },
54
+ {
55
+ "epoch": 1.4693534844668346,
56
+ "grad_norm": 2.1490907669067383,
57
+ "learning_rate": 3.5e-05,
58
+ "loss": 9.5659,
59
+ "step": 3500
60
+ },
61
+ {
62
+ "epoch": 1.6792611251049538,
63
+ "grad_norm": 1.8452446460723877,
64
+ "learning_rate": 4e-05,
65
+ "loss": 9.4905,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 1.8891687657430731,
70
+ "grad_norm": 2.248866558074951,
71
+ "learning_rate": 4.5e-05,
72
+ "loss": 9.4311,
73
+ "step": 4500
74
+ },
75
+ {
76
+ "epoch": 2.0990764063811924,
77
+ "grad_norm": 2.205521583557129,
78
+ "learning_rate": 5e-05,
79
+ "loss": 9.3799,
80
+ "step": 5000
81
+ },
82
+ {
83
+ "epoch": 2.3089840470193117,
84
+ "grad_norm": 2.088688850402832,
85
+ "learning_rate": 5.500000000000001e-05,
86
+ "loss": 9.3392,
87
+ "step": 5500
88
+ },
89
+ {
90
+ "epoch": 2.5188916876574305,
91
+ "grad_norm": 2.5509796142578125,
92
+ "learning_rate": 6e-05,
93
+ "loss": 9.304,
94
+ "step": 6000
95
+ },
96
+ {
97
+ "epoch": 2.72879932829555,
98
+ "grad_norm": 2.811840772628784,
99
+ "learning_rate": 6.500000000000001e-05,
100
+ "loss": 9.2708,
101
+ "step": 6500
102
+ },
103
+ {
104
+ "epoch": 2.938706968933669,
105
+ "grad_norm": 2.690173864364624,
106
+ "learning_rate": 7e-05,
107
+ "loss": 9.2417,
108
+ "step": 7000
109
+ },
110
+ {
111
+ "epoch": 3.1486146095717884,
112
+ "grad_norm": 2.387943983078003,
113
+ "learning_rate": 7.500000000000001e-05,
114
+ "loss": 9.2027,
115
+ "step": 7500
116
+ },
117
+ {
118
+ "epoch": 3.3585222502099077,
119
+ "grad_norm": 2.3690969944000244,
120
+ "learning_rate": 8e-05,
121
+ "loss": 9.1765,
122
+ "step": 8000
123
+ },
124
+ {
125
+ "epoch": 3.568429890848027,
126
+ "grad_norm": 3.117220163345337,
127
+ "learning_rate": 8.5e-05,
128
+ "loss": 9.162,
129
+ "step": 8500
130
+ },
131
+ {
132
+ "epoch": 3.7783375314861463,
133
+ "grad_norm": 3.0220932960510254,
134
+ "learning_rate": 9e-05,
135
+ "loss": 9.1394,
136
+ "step": 9000
137
+ },
138
+ {
139
+ "epoch": 3.988245172124265,
140
+ "grad_norm": 3.0078608989715576,
141
+ "learning_rate": 9.5e-05,
142
+ "loss": 9.1265,
143
+ "step": 9500
144
+ },
145
+ {
146
+ "epoch": 4.198152812762385,
147
+ "grad_norm": 3.7584919929504395,
148
+ "learning_rate": 0.0001,
149
+ "loss": 9.0191,
150
+ "step": 10000
151
+ },
152
+ {
153
+ "epoch": 4.198152812762385,
154
+ "eval_accuracy": 0.15056067225020173,
155
+ "eval_loss": 9.169220924377441,
156
+ "eval_runtime": 0.8946,
157
+ "eval_samples_per_second": 535.46,
158
+ "eval_steps_per_second": 5.589,
159
+ "step": 10000
160
+ },
161
+ {
162
+ "epoch": 4.408060453400504,
163
+ "grad_norm": 2.652250289916992,
164
+ "learning_rate": 9.899091826437942e-05,
165
+ "loss": 9.0207,
166
+ "step": 10500
167
+ },
168
+ {
169
+ "epoch": 4.617968094038623,
170
+ "grad_norm": 2.9136319160461426,
171
+ "learning_rate": 9.798183652875883e-05,
172
+ "loss": 9.02,
173
+ "step": 11000
174
+ },
175
+ {
176
+ "epoch": 4.827875734676742,
177
+ "grad_norm": 3.528905153274536,
178
+ "learning_rate": 9.697275479313825e-05,
179
+ "loss": 9.0215,
180
+ "step": 11500
181
+ },
182
+ {
183
+ "epoch": 5.037783375314861,
184
+ "grad_norm": 3.356295347213745,
185
+ "learning_rate": 9.596367305751767e-05,
186
+ "loss": 8.9843,
187
+ "step": 12000
188
+ },
189
+ {
190
+ "epoch": 5.247691015952981,
191
+ "grad_norm": 3.516859531402588,
192
+ "learning_rate": 9.495459132189707e-05,
193
+ "loss": 8.864,
194
+ "step": 12500
195
+ },
196
+ {
197
+ "epoch": 5.4575986565911,
198
+ "grad_norm": 3.8037853240966797,
199
+ "learning_rate": 9.394550958627649e-05,
200
+ "loss": 8.8851,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 5.667506297229219,
205
+ "grad_norm": 2.843942880630493,
206
+ "learning_rate": 9.293642785065591e-05,
207
+ "loss": 8.8948,
208
+ "step": 13500
209
+ },
210
+ {
211
+ "epoch": 5.877413937867338,
212
+ "grad_norm": 3.761806011199951,
213
+ "learning_rate": 9.192734611503532e-05,
214
+ "loss": 8.8942,
215
+ "step": 14000
216
+ },
217
+ {
218
+ "epoch": 6.087321578505458,
219
+ "grad_norm": 3.2125625610351562,
220
+ "learning_rate": 9.091826437941474e-05,
221
+ "loss": 8.8272,
222
+ "step": 14500
223
+ },
224
+ {
225
+ "epoch": 6.297229219143577,
226
+ "grad_norm": 2.9537875652313232,
227
+ "learning_rate": 8.990918264379415e-05,
228
+ "loss": 8.7602,
229
+ "step": 15000
230
+ },
231
+ {
232
+ "epoch": 6.507136859781696,
233
+ "grad_norm": 2.773268461227417,
234
+ "learning_rate": 8.890010090817356e-05,
235
+ "loss": 8.776,
236
+ "step": 15500
237
+ },
238
+ {
239
+ "epoch": 6.717044500419815,
240
+ "grad_norm": 3.131239891052246,
241
+ "learning_rate": 8.789101917255298e-05,
242
+ "loss": 8.7919,
243
+ "step": 16000
244
+ },
245
+ {
246
+ "epoch": 6.926952141057934,
247
+ "grad_norm": 4.001720428466797,
248
+ "learning_rate": 8.68819374369324e-05,
249
+ "loss": 8.7982,
250
+ "step": 16500
251
+ },
252
+ {
253
+ "epoch": 7.136859781696054,
254
+ "grad_norm": 3.204725503921509,
255
+ "learning_rate": 8.58728557013118e-05,
256
+ "loss": 8.6755,
257
+ "step": 17000
258
+ },
259
+ {
260
+ "epoch": 7.346767422334173,
261
+ "grad_norm": 3.509864568710327,
262
+ "learning_rate": 8.486377396569122e-05,
263
+ "loss": 8.632,
264
+ "step": 17500
265
+ },
266
+ {
267
+ "epoch": 7.5566750629722925,
268
+ "grad_norm": 3.831369638442993,
269
+ "learning_rate": 8.385469223007064e-05,
270
+ "loss": 8.6673,
271
+ "step": 18000
272
+ },
273
+ {
274
+ "epoch": 7.766582703610411,
275
+ "grad_norm": 3.5109853744506836,
276
+ "learning_rate": 8.284561049445006e-05,
277
+ "loss": 8.6815,
278
+ "step": 18500
279
+ },
280
+ {
281
+ "epoch": 7.97649034424853,
282
+ "grad_norm": 3.758779764175415,
283
+ "learning_rate": 8.183652875882947e-05,
284
+ "loss": 8.6855,
285
+ "step": 19000
286
+ },
287
+ {
288
+ "epoch": 8.18639798488665,
289
+ "grad_norm": 3.7631304264068604,
290
+ "learning_rate": 8.082744702320889e-05,
291
+ "loss": 8.4605,
292
+ "step": 19500
293
+ },
294
+ {
295
+ "epoch": 8.39630562552477,
296
+ "grad_norm": 3.984078884124756,
297
+ "learning_rate": 7.98183652875883e-05,
298
+ "loss": 8.4686,
299
+ "step": 20000
300
+ },
301
+ {
302
+ "epoch": 8.39630562552477,
303
+ "eval_accuracy": 0.1512668705631471,
304
+ "eval_loss": 9.461175918579102,
305
+ "eval_runtime": 0.7985,
306
+ "eval_samples_per_second": 599.851,
307
+ "eval_steps_per_second": 6.261,
308
+ "step": 20000
309
+ },
310
+ {
311
+ "epoch": 8.606213266162888,
312
+ "grad_norm": 4.056648254394531,
313
+ "learning_rate": 7.880928355196771e-05,
314
+ "loss": 8.4752,
315
+ "step": 20500
316
+ },
317
+ {
318
+ "epoch": 8.816120906801007,
319
+ "grad_norm": 4.289050579071045,
320
+ "learning_rate": 7.780020181634713e-05,
321
+ "loss": 8.5025,
322
+ "step": 21000
323
+ },
324
+ {
325
+ "epoch": 9.026028547439127,
326
+ "grad_norm": 4.93964147567749,
327
+ "learning_rate": 7.679112008072655e-05,
328
+ "loss": 8.462,
329
+ "step": 21500
330
+ },
331
+ {
332
+ "epoch": 9.235936188077247,
333
+ "grad_norm": 5.5340776443481445,
334
+ "learning_rate": 7.578203834510595e-05,
335
+ "loss": 8.1657,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 9.445843828715365,
340
+ "grad_norm": 5.351807594299316,
341
+ "learning_rate": 7.477295660948537e-05,
342
+ "loss": 8.1843,
343
+ "step": 22500
344
+ },
345
+ {
346
+ "epoch": 9.655751469353484,
347
+ "grad_norm": 5.357039451599121,
348
+ "learning_rate": 7.376387487386479e-05,
349
+ "loss": 8.2259,
350
+ "step": 23000
351
+ },
352
+ {
353
+ "epoch": 9.865659109991604,
354
+ "grad_norm": 5.8689751625061035,
355
+ "learning_rate": 7.27547931382442e-05,
356
+ "loss": 8.2237,
357
+ "step": 23500
358
+ },
359
+ {
360
+ "epoch": 10.075566750629722,
361
+ "grad_norm": 5.938015460968018,
362
+ "learning_rate": 7.174571140262362e-05,
363
+ "loss": 8.0956,
364
+ "step": 24000
365
+ },
366
+ {
367
+ "epoch": 10.285474391267842,
368
+ "grad_norm": 6.712480545043945,
369
+ "learning_rate": 7.073662966700304e-05,
370
+ "loss": 7.874,
371
+ "step": 24500
372
+ },
373
+ {
374
+ "epoch": 10.495382031905962,
375
+ "grad_norm": 5.908409595489502,
376
+ "learning_rate": 6.972754793138244e-05,
377
+ "loss": 7.8939,
378
+ "step": 25000
379
+ },
380
+ {
381
+ "epoch": 10.705289672544081,
382
+ "grad_norm": 5.897430419921875,
383
+ "learning_rate": 6.871846619576186e-05,
384
+ "loss": 7.9246,
385
+ "step": 25500
386
+ },
387
+ {
388
+ "epoch": 10.9151973131822,
389
+ "grad_norm": 6.488427639007568,
390
+ "learning_rate": 6.770938446014128e-05,
391
+ "loss": 7.953,
392
+ "step": 26000
393
+ },
394
+ {
395
+ "epoch": 11.125104953820319,
396
+ "grad_norm": 7.22756290435791,
397
+ "learning_rate": 6.670030272452068e-05,
398
+ "loss": 7.7254,
399
+ "step": 26500
400
+ },
401
+ {
402
+ "epoch": 11.335012594458439,
403
+ "grad_norm": 7.378805637359619,
404
+ "learning_rate": 6.56912209889001e-05,
405
+ "loss": 7.6003,
406
+ "step": 27000
407
+ },
408
+ {
409
+ "epoch": 11.544920235096557,
410
+ "grad_norm": 6.353366374969482,
411
+ "learning_rate": 6.468213925327952e-05,
412
+ "loss": 7.6439,
413
+ "step": 27500
414
+ },
415
+ {
416
+ "epoch": 11.754827875734676,
417
+ "grad_norm": 8.001075744628906,
418
+ "learning_rate": 6.367305751765893e-05,
419
+ "loss": 7.655,
420
+ "step": 28000
421
+ },
422
+ {
423
+ "epoch": 11.964735516372796,
424
+ "grad_norm": 6.13035774230957,
425
+ "learning_rate": 6.266397578203835e-05,
426
+ "loss": 7.6915,
427
+ "step": 28500
428
+ },
429
+ {
430
+ "epoch": 12.174643157010916,
431
+ "grad_norm": 7.017744064331055,
432
+ "learning_rate": 6.165489404641777e-05,
433
+ "loss": 7.386,
434
+ "step": 29000
435
+ },
436
+ {
437
+ "epoch": 12.384550797649034,
438
+ "grad_norm": 8.221517562866211,
439
+ "learning_rate": 6.064581231079718e-05,
440
+ "loss": 7.3464,
441
+ "step": 29500
442
+ },
443
+ {
444
+ "epoch": 12.594458438287154,
445
+ "grad_norm": 7.083924293518066,
446
+ "learning_rate": 5.963673057517659e-05,
447
+ "loss": 7.4003,
448
+ "step": 30000
449
+ },
450
+ {
451
+ "epoch": 12.594458438287154,
452
+ "eval_accuracy": 0.1536782606262417,
453
+ "eval_loss": 10.904170036315918,
454
+ "eval_runtime": 0.802,
455
+ "eval_samples_per_second": 597.253,
456
+ "eval_steps_per_second": 6.234,
457
+ "step": 30000
458
+ },
459
+ {
460
+ "epoch": 12.804366078925273,
461
+ "grad_norm": 8.425161361694336,
462
+ "learning_rate": 5.8627648839556e-05,
463
+ "loss": 7.4267,
464
+ "step": 30500
465
+ },
466
+ {
467
+ "epoch": 13.014273719563391,
468
+ "grad_norm": 8.940080642700195,
469
+ "learning_rate": 5.7618567103935416e-05,
470
+ "loss": 7.4089,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 13.224181360201511,
475
+ "grad_norm": 8.634459495544434,
476
+ "learning_rate": 5.660948536831484e-05,
477
+ "loss": 7.0836,
478
+ "step": 31500
479
+ },
480
+ {
481
+ "epoch": 13.43408900083963,
482
+ "grad_norm": 7.144838333129883,
483
+ "learning_rate": 5.5600403632694254e-05,
484
+ "loss": 7.1341,
485
+ "step": 32000
486
+ },
487
+ {
488
+ "epoch": 13.64399664147775,
489
+ "grad_norm": 8.895588874816895,
490
+ "learning_rate": 5.4591321897073666e-05,
491
+ "loss": 7.1816,
492
+ "step": 32500
493
+ },
494
+ {
495
+ "epoch": 13.853904282115868,
496
+ "grad_norm": 7.708846092224121,
497
+ "learning_rate": 5.358224016145308e-05,
498
+ "loss": 7.1807,
499
+ "step": 33000
500
+ },
501
+ {
502
+ "epoch": 14.063811922753988,
503
+ "grad_norm": 8.748336791992188,
504
+ "learning_rate": 5.257315842583249e-05,
505
+ "loss": 7.0967,
506
+ "step": 33500
507
+ },
508
+ {
509
+ "epoch": 14.273719563392108,
510
+ "grad_norm": 9.473174095153809,
511
+ "learning_rate": 5.15640766902119e-05,
512
+ "loss": 6.8908,
513
+ "step": 34000
514
+ },
515
+ {
516
+ "epoch": 14.483627204030226,
517
+ "grad_norm": 8.400636672973633,
518
+ "learning_rate": 5.055499495459133e-05,
519
+ "loss": 6.9241,
520
+ "step": 34500
521
+ },
522
+ {
523
+ "epoch": 14.693534844668346,
524
+ "grad_norm": 8.310261726379395,
525
+ "learning_rate": 4.954591321897074e-05,
526
+ "loss": 6.9588,
527
+ "step": 35000
528
+ },
529
+ {
530
+ "epoch": 14.903442485306465,
531
+ "grad_norm": 9.908066749572754,
532
+ "learning_rate": 4.853683148335015e-05,
533
+ "loss": 6.967,
534
+ "step": 35500
535
+ },
536
+ {
537
+ "epoch": 15.113350125944585,
538
+ "grad_norm": 9.162622451782227,
539
+ "learning_rate": 4.7527749747729566e-05,
540
+ "loss": 6.8265,
541
+ "step": 36000
542
+ },
543
+ {
544
+ "epoch": 15.323257766582703,
545
+ "grad_norm": 9.188130378723145,
546
+ "learning_rate": 4.6518668012108985e-05,
547
+ "loss": 6.7211,
548
+ "step": 36500
549
+ },
550
+ {
551
+ "epoch": 15.533165407220823,
552
+ "grad_norm": 8.647799491882324,
553
+ "learning_rate": 4.55095862764884e-05,
554
+ "loss": 6.7462,
555
+ "step": 37000
556
+ },
557
+ {
558
+ "epoch": 15.743073047858942,
559
+ "grad_norm": 8.854975700378418,
560
+ "learning_rate": 4.450050454086781e-05,
561
+ "loss": 6.774,
562
+ "step": 37500
563
+ },
564
+ {
565
+ "epoch": 15.95298068849706,
566
+ "grad_norm": 9.705062866210938,
567
+ "learning_rate": 4.349142280524723e-05,
568
+ "loss": 6.7905,
569
+ "step": 38000
570
+ },
571
+ {
572
+ "epoch": 16.162888329135182,
573
+ "grad_norm": 8.551140785217285,
574
+ "learning_rate": 4.248234106962664e-05,
575
+ "loss": 6.6067,
576
+ "step": 38500
577
+ },
578
+ {
579
+ "epoch": 16.3727959697733,
580
+ "grad_norm": 8.363558769226074,
581
+ "learning_rate": 4.147325933400605e-05,
582
+ "loss": 6.5872,
583
+ "step": 39000
584
+ },
585
+ {
586
+ "epoch": 16.582703610411418,
587
+ "grad_norm": 10.148078918457031,
588
+ "learning_rate": 4.046417759838547e-05,
589
+ "loss": 6.5978,
590
+ "step": 39500
591
+ },
592
+ {
593
+ "epoch": 16.79261125104954,
594
+ "grad_norm": 11.427934646606445,
595
+ "learning_rate": 3.9455095862764884e-05,
596
+ "loss": 6.6138,
597
+ "step": 40000
598
+ },
599
+ {
600
+ "epoch": 16.79261125104954,
601
+ "eval_accuracy": 0.15131210048222496,
602
+ "eval_loss": 12.232389450073242,
603
+ "eval_runtime": 0.7967,
604
+ "eval_samples_per_second": 601.267,
605
+ "eval_steps_per_second": 6.276,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 17.002518891687657,
610
+ "grad_norm": 8.80695629119873,
611
+ "learning_rate": 3.84460141271443e-05,
612
+ "loss": 6.6346,
613
+ "step": 40500
614
+ },
615
+ {
616
+ "epoch": 17.212426532325775,
617
+ "grad_norm": 6.59151554107666,
618
+ "learning_rate": 3.7436932391523715e-05,
619
+ "loss": 6.4606,
620
+ "step": 41000
621
+ },
622
+ {
623
+ "epoch": 17.422334172963897,
624
+ "grad_norm": 9.461859703063965,
625
+ "learning_rate": 3.642785065590313e-05,
626
+ "loss": 6.4748,
627
+ "step": 41500
628
+ },
629
+ {
630
+ "epoch": 17.632241813602015,
631
+ "grad_norm": 8.955385208129883,
632
+ "learning_rate": 3.541876892028255e-05,
633
+ "loss": 6.4959,
634
+ "step": 42000
635
+ },
636
+ {
637
+ "epoch": 17.842149454240133,
638
+ "grad_norm": 9.80893611907959,
639
+ "learning_rate": 3.440968718466196e-05,
640
+ "loss": 6.5126,
641
+ "step": 42500
642
+ },
643
+ {
644
+ "epoch": 18.052057094878254,
645
+ "grad_norm": 7.227355003356934,
646
+ "learning_rate": 3.340060544904137e-05,
647
+ "loss": 6.4837,
648
+ "step": 43000
649
+ },
650
+ {
651
+ "epoch": 18.261964735516372,
652
+ "grad_norm": 4.860254764556885,
653
+ "learning_rate": 3.239152371342079e-05,
654
+ "loss": 6.3988,
655
+ "step": 43500
656
+ },
657
+ {
658
+ "epoch": 18.471872376154494,
659
+ "grad_norm": 7.772268295288086,
660
+ "learning_rate": 3.13824419778002e-05,
661
+ "loss": 6.406,
662
+ "step": 44000
663
+ },
664
+ {
665
+ "epoch": 18.68178001679261,
666
+ "grad_norm": 9.385651588439941,
667
+ "learning_rate": 3.0373360242179615e-05,
668
+ "loss": 6.4196,
669
+ "step": 44500
670
+ },
671
+ {
672
+ "epoch": 18.89168765743073,
673
+ "grad_norm": 9.672471046447754,
674
+ "learning_rate": 2.9364278506559034e-05,
675
+ "loss": 6.4373,
676
+ "step": 45000
677
+ },
678
+ {
679
+ "epoch": 19.10159529806885,
680
+ "grad_norm": 8.28320598602295,
681
+ "learning_rate": 2.8355196770938446e-05,
682
+ "loss": 6.3845,
683
+ "step": 45500
684
+ },
685
+ {
686
+ "epoch": 19.31150293870697,
687
+ "grad_norm": 7.17496395111084,
688
+ "learning_rate": 2.734611503531786e-05,
689
+ "loss": 6.3499,
690
+ "step": 46000
691
+ },
692
+ {
693
+ "epoch": 19.521410579345087,
694
+ "grad_norm": 9.062373161315918,
695
+ "learning_rate": 2.6337033299697278e-05,
696
+ "loss": 6.3459,
697
+ "step": 46500
698
+ },
699
+ {
700
+ "epoch": 19.73131821998321,
701
+ "grad_norm": 7.60693883895874,
702
+ "learning_rate": 2.532795156407669e-05,
703
+ "loss": 6.35,
704
+ "step": 47000
705
+ },
706
+ {
707
+ "epoch": 19.941225860621326,
708
+ "grad_norm": 7.330896377563477,
709
+ "learning_rate": 2.4318869828456106e-05,
710
+ "loss": 6.3586,
711
+ "step": 47500
712
+ },
713
+ {
714
+ "epoch": 20.151133501259444,
715
+ "grad_norm": 8.745220184326172,
716
+ "learning_rate": 2.330978809283552e-05,
717
+ "loss": 6.3213,
718
+ "step": 48000
719
+ },
720
+ {
721
+ "epoch": 20.361041141897566,
722
+ "grad_norm": 6.972262859344482,
723
+ "learning_rate": 2.2300706357214934e-05,
724
+ "loss": 6.3084,
725
+ "step": 48500
726
+ },
727
+ {
728
+ "epoch": 20.570948782535684,
729
+ "grad_norm": 6.277470588684082,
730
+ "learning_rate": 2.129162462159435e-05,
731
+ "loss": 6.3001,
732
+ "step": 49000
733
+ },
734
+ {
735
+ "epoch": 20.7808564231738,
736
+ "grad_norm": 5.862802982330322,
737
+ "learning_rate": 2.0282542885973765e-05,
738
+ "loss": 6.3026,
739
+ "step": 49500
740
+ },
741
+ {
742
+ "epoch": 20.990764063811923,
743
+ "grad_norm": 7.365855693817139,
744
+ "learning_rate": 1.927346115035318e-05,
745
+ "loss": 6.312,
746
+ "step": 50000
747
+ },
748
+ {
749
+ "epoch": 20.990764063811923,
750
+ "eval_accuracy": 0.1510920458687302,
751
+ "eval_loss": 12.813960075378418,
752
+ "eval_runtime": 0.8028,
753
+ "eval_samples_per_second": 596.652,
754
+ "eval_steps_per_second": 6.228,
755
+ "step": 50000
756
+ },
757
+ {
758
+ "epoch": 21.20067170445004,
759
+ "grad_norm": 8.677397727966309,
760
+ "learning_rate": 1.8264379414732593e-05,
761
+ "loss": 6.2693,
762
+ "step": 50500
763
+ },
764
+ {
765
+ "epoch": 21.410579345088163,
766
+ "grad_norm": 9.135784149169922,
767
+ "learning_rate": 1.725529767911201e-05,
768
+ "loss": 6.2795,
769
+ "step": 51000
770
+ },
771
+ {
772
+ "epoch": 21.62048698572628,
773
+ "grad_norm": 7.724148750305176,
774
+ "learning_rate": 1.6246215943491424e-05,
775
+ "loss": 6.2656,
776
+ "step": 51500
777
+ },
778
+ {
779
+ "epoch": 21.8303946263644,
780
+ "grad_norm": 5.842483997344971,
781
+ "learning_rate": 1.5237134207870837e-05,
782
+ "loss": 6.2682,
783
+ "step": 52000
784
+ },
785
+ {
786
+ "epoch": 22.04030226700252,
787
+ "grad_norm": 7.466794013977051,
788
+ "learning_rate": 1.4228052472250252e-05,
789
+ "loss": 6.2664,
790
+ "step": 52500
791
+ },
792
+ {
793
+ "epoch": 22.250209907640638,
794
+ "grad_norm": 4.366174697875977,
795
+ "learning_rate": 1.3218970736629668e-05,
796
+ "loss": 6.2354,
797
+ "step": 53000
798
+ },
799
+ {
800
+ "epoch": 22.460117548278756,
801
+ "grad_norm": 7.936322212219238,
802
+ "learning_rate": 1.2209889001009082e-05,
803
+ "loss": 6.2374,
804
+ "step": 53500
805
+ },
806
+ {
807
+ "epoch": 22.670025188916878,
808
+ "grad_norm": 7.256034851074219,
809
+ "learning_rate": 1.1200807265388496e-05,
810
+ "loss": 6.2416,
811
+ "step": 54000
812
+ },
813
+ {
814
+ "epoch": 22.879932829554996,
815
+ "grad_norm": 6.694550514221191,
816
+ "learning_rate": 1.0191725529767912e-05,
817
+ "loss": 6.2335,
818
+ "step": 54500
819
+ },
820
+ {
821
+ "epoch": 23.089840470193113,
822
+ "grad_norm": 5.7823486328125,
823
+ "learning_rate": 9.182643794147326e-06,
824
+ "loss": 6.225,
825
+ "step": 55000
826
+ },
827
+ {
828
+ "epoch": 23.299748110831235,
829
+ "grad_norm": 8.693979263305664,
830
+ "learning_rate": 8.173562058526741e-06,
831
+ "loss": 6.2114,
832
+ "step": 55500
833
+ },
834
+ {
835
+ "epoch": 23.509655751469353,
836
+ "grad_norm": 6.171530723571777,
837
+ "learning_rate": 7.164480322906156e-06,
838
+ "loss": 6.2232,
839
+ "step": 56000
840
+ },
841
+ {
842
+ "epoch": 23.719563392107474,
843
+ "grad_norm": 5.320813179016113,
844
+ "learning_rate": 6.155398587285571e-06,
845
+ "loss": 6.204,
846
+ "step": 56500
847
+ },
848
+ {
849
+ "epoch": 23.929471032745592,
850
+ "grad_norm": 7.7308526039123535,
851
+ "learning_rate": 5.146316851664986e-06,
852
+ "loss": 6.2156,
853
+ "step": 57000
854
+ },
855
+ {
856
+ "epoch": 24.13937867338371,
857
+ "grad_norm": 5.540814399719238,
858
+ "learning_rate": 4.1372351160444e-06,
859
+ "loss": 6.2013,
860
+ "step": 57500
861
+ },
862
+ {
863
+ "epoch": 24.349286314021832,
864
+ "grad_norm": 9.152900695800781,
865
+ "learning_rate": 3.1281533804238145e-06,
866
+ "loss": 6.2049,
867
+ "step": 58000
868
+ },
869
+ {
870
+ "epoch": 24.55919395465995,
871
+ "grad_norm": 6.086564064025879,
872
+ "learning_rate": 2.1190716448032293e-06,
873
+ "loss": 6.2016,
874
+ "step": 58500
875
+ },
876
+ {
877
+ "epoch": 24.769101595298068,
878
+ "grad_norm": 7.7911200523376465,
879
+ "learning_rate": 1.109989909182644e-06,
880
+ "loss": 6.1998,
881
+ "step": 59000
882
+ },
883
+ {
884
+ "epoch": 24.97900923593619,
885
+ "grad_norm": 4.825611591339111,
886
+ "learning_rate": 1.0090817356205853e-07,
887
+ "loss": 6.1863,
888
+ "step": 59500
889
+ },
890
+ {
891
+ "epoch": 25.0,
892
+ "step": 59550,
893
+ "total_flos": 7.68356804521728e+17,
894
+ "train_loss": 7.6570130803823275,
895
+ "train_runtime": 16738.0825,
896
+ "train_samples_per_second": 341.495,
897
+ "train_steps_per_second": 3.558
898
+ }
899
+ ],
900
+ "logging_steps": 500,
901
+ "max_steps": 59550,
902
+ "num_input_tokens_seen": 0,
903
+ "num_train_epochs": 25,
904
+ "save_steps": 10000,
905
+ "stateful_callbacks": {
906
+ "TrainerControl": {
907
+ "args": {
908
+ "should_epoch_stop": false,
909
+ "should_evaluate": false,
910
+ "should_log": false,
911
+ "should_save": true,
912
+ "should_training_stop": true
913
+ },
914
+ "attributes": {}
915
+ }
916
+ },
917
+ "total_flos": 7.68356804521728e+17,
918
+ "train_batch_size": 96,
919
+ "trial_name": null,
920
+ "trial_params": null
921
+ }