gokulsrinivasagan commited on
Commit
9074521
·
verified ·
1 Parent(s): d991b4c

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
@@ -4,11 +4,23 @@ license: apache-2.0
4
  base_model: distilbert-base-uncased
5
  tags:
6
  - generated_from_trainer
 
 
7
  metrics:
8
  - accuracy
9
  model-index:
10
  - name: tinybert_base_train_kd
11
- results: []
 
 
 
 
 
 
 
 
 
 
12
  ---
13
 
14
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,10 +28,10 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  # tinybert_base_train_kd
18
 
19
- This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 63.0908
22
- - Accuracy: 0.5510
23
 
24
  ## Model description
25
 
 
4
  base_model: distilbert-base-uncased
5
  tags:
6
  - generated_from_trainer
7
+ datasets:
8
+ - gokulsrinivasagan/processed_wikitext-103-raw-v1-ld
9
  metrics:
10
  - accuracy
11
  model-index:
12
  - name: tinybert_base_train_kd
13
+ results:
14
+ - task:
15
+ name: Masked Language Modeling
16
+ type: fill-mask
17
+ dataset:
18
+ name: gokulsrinivasagan/processed_wikitext-103-raw-v1-ld
19
+ type: gokulsrinivasagan/processed_wikitext-103-raw-v1-ld
20
+ metrics:
21
+ - name: Accuracy
22
+ type: accuracy
23
+ value: 0.5510538509351868
24
  ---
25
 
26
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
28
 
29
  # tinybert_base_train_kd
30
 
31
+ This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the gokulsrinivasagan/processed_wikitext-103-raw-v1-ld dataset.
32
  It achieves the following results on the evaluation set:
33
+ - Loss: 62.6034
34
+ - Accuracy: 0.5511
35
 
36
  ## Model description
37
 
all_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 25.0,
3
+ "eval_accuracy": 0.5510538509351868,
4
+ "eval_loss": 62.603424072265625,
5
+ "eval_runtime": 3.155,
6
+ "eval_samples": 479,
7
+ "eval_samples_per_second": 151.823,
8
+ "eval_steps_per_second": 1.585,
9
+ "perplexity": 1.5428426035822972e+27,
10
+ "total_flos": 7.577162542158336e+17,
11
+ "train_loss": 301.8068682845954,
12
+ "train_runtime": 49685.0337,
13
+ "train_samples": 228639,
14
+ "train_samples_per_second": 115.044,
15
+ "train_steps_per_second": 1.199
16
+ }
eval_results.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 25.0,
3
+ "eval_accuracy": 0.5510538509351868,
4
+ "eval_loss": 62.603424072265625,
5
+ "eval_runtime": 3.155,
6
+ "eval_samples": 479,
7
+ "eval_samples_per_second": 151.823,
8
+ "eval_steps_per_second": 1.585,
9
+ "perplexity": 1.5428426035822972e+27
10
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 25.0,
3
+ "total_flos": 7.577162542158336e+17,
4
+ "train_loss": 301.8068682845954,
5
+ "train_runtime": 49685.0337,
6
+ "train_samples": 228639,
7
+ "train_samples_per_second": 115.044,
8
+ "train_steps_per_second": 1.199
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,921 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 50000,
3
+ "best_metric": 63.090843200683594,
4
+ "best_model_checkpoint": "tinybert_base_train_kd/checkpoint-50000",
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": 1038.087890625,
15
+ "learning_rate": 4.9900000000000005e-06,
16
+ "loss": 6028.7805,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 0.41981528127623846,
21
+ "grad_norm": 1000.9192504882812,
22
+ "learning_rate": 9.990000000000001e-06,
23
+ "loss": 3688.4922,
24
+ "step": 1000
25
+ },
26
+ {
27
+ "epoch": 0.6297229219143576,
28
+ "grad_norm": 1443.9014892578125,
29
+ "learning_rate": 1.499e-05,
30
+ "loss": 2203.5858,
31
+ "step": 1500
32
+ },
33
+ {
34
+ "epoch": 0.8396305625524769,
35
+ "grad_norm": 2327.7587890625,
36
+ "learning_rate": 1.999e-05,
37
+ "loss": 1477.7162,
38
+ "step": 2000
39
+ },
40
+ {
41
+ "epoch": 1.0495382031905962,
42
+ "grad_norm": 1928.0552978515625,
43
+ "learning_rate": 2.4990000000000003e-05,
44
+ "loss": 1145.1917,
45
+ "step": 2500
46
+ },
47
+ {
48
+ "epoch": 1.2594458438287153,
49
+ "grad_norm": 1427.1600341796875,
50
+ "learning_rate": 2.9990000000000003e-05,
51
+ "loss": 941.4575,
52
+ "step": 3000
53
+ },
54
+ {
55
+ "epoch": 1.4693534844668346,
56
+ "grad_norm": 832.8273315429688,
57
+ "learning_rate": 3.499e-05,
58
+ "loss": 812.8064,
59
+ "step": 3500
60
+ },
61
+ {
62
+ "epoch": 1.6792611251049538,
63
+ "grad_norm": 880.94140625,
64
+ "learning_rate": 3.999e-05,
65
+ "loss": 722.3781,
66
+ "step": 4000
67
+ },
68
+ {
69
+ "epoch": 1.8891687657430731,
70
+ "grad_norm": 458.8153381347656,
71
+ "learning_rate": 4.499e-05,
72
+ "loss": 657.1395,
73
+ "step": 4500
74
+ },
75
+ {
76
+ "epoch": 2.0990764063811924,
77
+ "grad_norm": 192.89340209960938,
78
+ "learning_rate": 4.999e-05,
79
+ "loss": 608.8174,
80
+ "step": 5000
81
+ },
82
+ {
83
+ "epoch": 2.3089840470193117,
84
+ "grad_norm": 173.0674285888672,
85
+ "learning_rate": 5.499000000000001e-05,
86
+ "loss": 575.3464,
87
+ "step": 5500
88
+ },
89
+ {
90
+ "epoch": 2.5188916876574305,
91
+ "grad_norm": 119.65222930908203,
92
+ "learning_rate": 5.999e-05,
93
+ "loss": 550.9334,
94
+ "step": 6000
95
+ },
96
+ {
97
+ "epoch": 2.72879932829555,
98
+ "grad_norm": 163.03219604492188,
99
+ "learning_rate": 6.499000000000001e-05,
100
+ "loss": 532.3316,
101
+ "step": 6500
102
+ },
103
+ {
104
+ "epoch": 2.938706968933669,
105
+ "grad_norm": 103.94256591796875,
106
+ "learning_rate": 6.999e-05,
107
+ "loss": 516.1778,
108
+ "step": 7000
109
+ },
110
+ {
111
+ "epoch": 3.1486146095717884,
112
+ "grad_norm": 137.8463134765625,
113
+ "learning_rate": 7.499e-05,
114
+ "loss": 502.3519,
115
+ "step": 7500
116
+ },
117
+ {
118
+ "epoch": 3.3585222502099077,
119
+ "grad_norm": 93.18783569335938,
120
+ "learning_rate": 7.999000000000001e-05,
121
+ "loss": 493.4963,
122
+ "step": 8000
123
+ },
124
+ {
125
+ "epoch": 3.568429890848027,
126
+ "grad_norm": 81.264404296875,
127
+ "learning_rate": 8.499e-05,
128
+ "loss": 486.5324,
129
+ "step": 8500
130
+ },
131
+ {
132
+ "epoch": 3.7783375314861463,
133
+ "grad_norm": 96.80731201171875,
134
+ "learning_rate": 8.999000000000001e-05,
135
+ "loss": 477.9452,
136
+ "step": 9000
137
+ },
138
+ {
139
+ "epoch": 3.988245172124265,
140
+ "grad_norm": 99.205322265625,
141
+ "learning_rate": 9.499e-05,
142
+ "loss": 463.8162,
143
+ "step": 9500
144
+ },
145
+ {
146
+ "epoch": 4.198152812762385,
147
+ "grad_norm": 118.66422271728516,
148
+ "learning_rate": 9.999000000000001e-05,
149
+ "loss": 447.3158,
150
+ "step": 10000
151
+ },
152
+ {
153
+ "epoch": 4.198152812762385,
154
+ "eval_accuracy": 0.16583655638721168,
155
+ "eval_loss": 428.842041015625,
156
+ "eval_runtime": 3.6172,
157
+ "eval_samples_per_second": 132.423,
158
+ "eval_steps_per_second": 1.382,
159
+ "step": 10000
160
+ },
161
+ {
162
+ "epoch": 4.408060453400504,
163
+ "grad_norm": 144.82168579101562,
164
+ "learning_rate": 9.899293642785066e-05,
165
+ "loss": 430.4443,
166
+ "step": 10500
167
+ },
168
+ {
169
+ "epoch": 4.617968094038623,
170
+ "grad_norm": 166.46974182128906,
171
+ "learning_rate": 9.798385469223008e-05,
172
+ "loss": 416.3207,
173
+ "step": 11000
174
+ },
175
+ {
176
+ "epoch": 4.827875734676742,
177
+ "grad_norm": 232.3704071044922,
178
+ "learning_rate": 9.69747729566095e-05,
179
+ "loss": 403.2591,
180
+ "step": 11500
181
+ },
182
+ {
183
+ "epoch": 5.037783375314861,
184
+ "grad_norm": 254.5363311767578,
185
+ "learning_rate": 9.59656912209889e-05,
186
+ "loss": 390.4413,
187
+ "step": 12000
188
+ },
189
+ {
190
+ "epoch": 5.247691015952981,
191
+ "grad_norm": 270.5792236328125,
192
+ "learning_rate": 9.495660948536832e-05,
193
+ "loss": 377.3261,
194
+ "step": 12500
195
+ },
196
+ {
197
+ "epoch": 5.4575986565911,
198
+ "grad_norm": 244.77581787109375,
199
+ "learning_rate": 9.394752774974774e-05,
200
+ "loss": 365.2654,
201
+ "step": 13000
202
+ },
203
+ {
204
+ "epoch": 5.667506297229219,
205
+ "grad_norm": 267.1288757324219,
206
+ "learning_rate": 9.293844601412714e-05,
207
+ "loss": 352.7486,
208
+ "step": 13500
209
+ },
210
+ {
211
+ "epoch": 5.877413937867338,
212
+ "grad_norm": 315.1538391113281,
213
+ "learning_rate": 9.192936427850656e-05,
214
+ "loss": 338.1035,
215
+ "step": 14000
216
+ },
217
+ {
218
+ "epoch": 6.087321578505458,
219
+ "grad_norm": 344.386962890625,
220
+ "learning_rate": 9.092028254288598e-05,
221
+ "loss": 323.3387,
222
+ "step": 14500
223
+ },
224
+ {
225
+ "epoch": 6.297229219143577,
226
+ "grad_norm": 345.59320068359375,
227
+ "learning_rate": 8.991120080726539e-05,
228
+ "loss": 308.7323,
229
+ "step": 15000
230
+ },
231
+ {
232
+ "epoch": 6.507136859781696,
233
+ "grad_norm": 373.200439453125,
234
+ "learning_rate": 8.890211907164481e-05,
235
+ "loss": 293.0822,
236
+ "step": 15500
237
+ },
238
+ {
239
+ "epoch": 6.717044500419815,
240
+ "grad_norm": 387.7672119140625,
241
+ "learning_rate": 8.789303733602423e-05,
242
+ "loss": 276.3089,
243
+ "step": 16000
244
+ },
245
+ {
246
+ "epoch": 6.926952141057934,
247
+ "grad_norm": 380.2518615722656,
248
+ "learning_rate": 8.688395560040363e-05,
249
+ "loss": 257.2038,
250
+ "step": 16500
251
+ },
252
+ {
253
+ "epoch": 7.136859781696054,
254
+ "grad_norm": 306.2325134277344,
255
+ "learning_rate": 8.587487386478305e-05,
256
+ "loss": 221.4098,
257
+ "step": 17000
258
+ },
259
+ {
260
+ "epoch": 7.346767422334173,
261
+ "grad_norm": 244.85455322265625,
262
+ "learning_rate": 8.486579212916247e-05,
263
+ "loss": 200.0532,
264
+ "step": 17500
265
+ },
266
+ {
267
+ "epoch": 7.5566750629722925,
268
+ "grad_norm": 243.97576904296875,
269
+ "learning_rate": 8.385671039354188e-05,
270
+ "loss": 186.9741,
271
+ "step": 18000
272
+ },
273
+ {
274
+ "epoch": 7.766582703610411,
275
+ "grad_norm": 201.4752655029297,
276
+ "learning_rate": 8.28476286579213e-05,
277
+ "loss": 176.1845,
278
+ "step": 18500
279
+ },
280
+ {
281
+ "epoch": 7.97649034424853,
282
+ "grad_norm": 180.40164184570312,
283
+ "learning_rate": 8.183854692230071e-05,
284
+ "loss": 167.2202,
285
+ "step": 19000
286
+ },
287
+ {
288
+ "epoch": 8.18639798488665,
289
+ "grad_norm": 169.40943908691406,
290
+ "learning_rate": 8.082946518668012e-05,
291
+ "loss": 158.6278,
292
+ "step": 19500
293
+ },
294
+ {
295
+ "epoch": 8.39630562552477,
296
+ "grad_norm": 142.1228790283203,
297
+ "learning_rate": 7.982038345105954e-05,
298
+ "loss": 151.7919,
299
+ "step": 20000
300
+ },
301
+ {
302
+ "epoch": 8.39630562552477,
303
+ "eval_accuracy": 0.48157916888801594,
304
+ "eval_loss": 135.78590393066406,
305
+ "eval_runtime": 3.5122,
306
+ "eval_samples_per_second": 136.38,
307
+ "eval_steps_per_second": 1.424,
308
+ "step": 20000
309
+ },
310
+ {
311
+ "epoch": 8.606213266162888,
312
+ "grad_norm": 155.93728637695312,
313
+ "learning_rate": 7.881130171543896e-05,
314
+ "loss": 145.7355,
315
+ "step": 20500
316
+ },
317
+ {
318
+ "epoch": 8.816120906801007,
319
+ "grad_norm": 139.52723693847656,
320
+ "learning_rate": 7.780221997981836e-05,
321
+ "loss": 140.6628,
322
+ "step": 21000
323
+ },
324
+ {
325
+ "epoch": 9.026028547439127,
326
+ "grad_norm": 146.34320068359375,
327
+ "learning_rate": 7.679313824419778e-05,
328
+ "loss": 135.8976,
329
+ "step": 21500
330
+ },
331
+ {
332
+ "epoch": 9.235936188077247,
333
+ "grad_norm": 129.9792938232422,
334
+ "learning_rate": 7.57840565085772e-05,
335
+ "loss": 131.474,
336
+ "step": 22000
337
+ },
338
+ {
339
+ "epoch": 9.445843828715365,
340
+ "grad_norm": 122.61577606201172,
341
+ "learning_rate": 7.477497477295662e-05,
342
+ "loss": 127.8668,
343
+ "step": 22500
344
+ },
345
+ {
346
+ "epoch": 9.655751469353484,
347
+ "grad_norm": 130.75668334960938,
348
+ "learning_rate": 7.376589303733603e-05,
349
+ "loss": 124.6147,
350
+ "step": 23000
351
+ },
352
+ {
353
+ "epoch": 9.865659109991604,
354
+ "grad_norm": 130.1267547607422,
355
+ "learning_rate": 7.275681130171544e-05,
356
+ "loss": 121.6354,
357
+ "step": 23500
358
+ },
359
+ {
360
+ "epoch": 10.075566750629722,
361
+ "grad_norm": 138.63009643554688,
362
+ "learning_rate": 7.174772956609486e-05,
363
+ "loss": 118.5567,
364
+ "step": 24000
365
+ },
366
+ {
367
+ "epoch": 10.285474391267842,
368
+ "grad_norm": 128.6032257080078,
369
+ "learning_rate": 7.073864783047427e-05,
370
+ "loss": 115.905,
371
+ "step": 24500
372
+ },
373
+ {
374
+ "epoch": 10.495382031905962,
375
+ "grad_norm": 117.67427825927734,
376
+ "learning_rate": 6.972956609485369e-05,
377
+ "loss": 113.5025,
378
+ "step": 25000
379
+ },
380
+ {
381
+ "epoch": 10.705289672544081,
382
+ "grad_norm": 132.61322021484375,
383
+ "learning_rate": 6.872048435923311e-05,
384
+ "loss": 111.2961,
385
+ "step": 25500
386
+ },
387
+ {
388
+ "epoch": 10.9151973131822,
389
+ "grad_norm": 124.02651977539062,
390
+ "learning_rate": 6.771140262361251e-05,
391
+ "loss": 109.3261,
392
+ "step": 26000
393
+ },
394
+ {
395
+ "epoch": 11.125104953820319,
396
+ "grad_norm": 116.80554962158203,
397
+ "learning_rate": 6.670232088799193e-05,
398
+ "loss": 106.9513,
399
+ "step": 26500
400
+ },
401
+ {
402
+ "epoch": 11.335012594458439,
403
+ "grad_norm": 120.98764038085938,
404
+ "learning_rate": 6.569323915237135e-05,
405
+ "loss": 105.1095,
406
+ "step": 27000
407
+ },
408
+ {
409
+ "epoch": 11.544920235096557,
410
+ "grad_norm": 109.36351776123047,
411
+ "learning_rate": 6.468415741675076e-05,
412
+ "loss": 103.5172,
413
+ "step": 27500
414
+ },
415
+ {
416
+ "epoch": 11.754827875734676,
417
+ "grad_norm": 115.52029418945312,
418
+ "learning_rate": 6.367507568113018e-05,
419
+ "loss": 101.8046,
420
+ "step": 28000
421
+ },
422
+ {
423
+ "epoch": 11.964735516372796,
424
+ "grad_norm": 110.40377044677734,
425
+ "learning_rate": 6.26659939455096e-05,
426
+ "loss": 100.335,
427
+ "step": 28500
428
+ },
429
+ {
430
+ "epoch": 12.174643157010916,
431
+ "grad_norm": 106.47163391113281,
432
+ "learning_rate": 6.1656912209889e-05,
433
+ "loss": 98.6005,
434
+ "step": 29000
435
+ },
436
+ {
437
+ "epoch": 12.384550797649034,
438
+ "grad_norm": 114.08397674560547,
439
+ "learning_rate": 6.064783047426842e-05,
440
+ "loss": 97.2174,
441
+ "step": 29500
442
+ },
443
+ {
444
+ "epoch": 12.594458438287154,
445
+ "grad_norm": 108.48992156982422,
446
+ "learning_rate": 5.963874873864783e-05,
447
+ "loss": 95.9257,
448
+ "step": 30000
449
+ },
450
+ {
451
+ "epoch": 12.594458438287154,
452
+ "eval_accuracy": 0.5308223969555362,
453
+ "eval_loss": 84.8065414428711,
454
+ "eval_runtime": 3.6726,
455
+ "eval_samples_per_second": 130.426,
456
+ "eval_steps_per_second": 1.361,
457
+ "step": 30000
458
+ },
459
+ {
460
+ "epoch": 12.804366078925273,
461
+ "grad_norm": 116.85979461669922,
462
+ "learning_rate": 5.8629667003027243e-05,
463
+ "loss": 94.7173,
464
+ "step": 30500
465
+ },
466
+ {
467
+ "epoch": 13.014273719563391,
468
+ "grad_norm": 101.37419891357422,
469
+ "learning_rate": 5.762058526740667e-05,
470
+ "loss": 93.6031,
471
+ "step": 31000
472
+ },
473
+ {
474
+ "epoch": 13.224181360201511,
475
+ "grad_norm": 101.53105926513672,
476
+ "learning_rate": 5.661150353178608e-05,
477
+ "loss": 92.2381,
478
+ "step": 31500
479
+ },
480
+ {
481
+ "epoch": 13.43408900083963,
482
+ "grad_norm": 131.7721405029297,
483
+ "learning_rate": 5.5602421796165494e-05,
484
+ "loss": 91.2291,
485
+ "step": 32000
486
+ },
487
+ {
488
+ "epoch": 13.64399664147775,
489
+ "grad_norm": 147.26918029785156,
490
+ "learning_rate": 5.4593340060544906e-05,
491
+ "loss": 90.265,
492
+ "step": 32500
493
+ },
494
+ {
495
+ "epoch": 13.853904282115868,
496
+ "grad_norm": 107.17750549316406,
497
+ "learning_rate": 5.358425832492432e-05,
498
+ "loss": 89.2605,
499
+ "step": 33000
500
+ },
501
+ {
502
+ "epoch": 14.063811922753988,
503
+ "grad_norm": 135.6392364501953,
504
+ "learning_rate": 5.257517658930373e-05,
505
+ "loss": 88.1679,
506
+ "step": 33500
507
+ },
508
+ {
509
+ "epoch": 14.273719563392108,
510
+ "grad_norm": 131.0419158935547,
511
+ "learning_rate": 5.1566094853683156e-05,
512
+ "loss": 87.1749,
513
+ "step": 34000
514
+ },
515
+ {
516
+ "epoch": 14.483627204030226,
517
+ "grad_norm": 110.07307434082031,
518
+ "learning_rate": 5.055701311806257e-05,
519
+ "loss": 86.2937,
520
+ "step": 34500
521
+ },
522
+ {
523
+ "epoch": 14.693534844668346,
524
+ "grad_norm": 107.72781372070312,
525
+ "learning_rate": 4.954793138244198e-05,
526
+ "loss": 85.5873,
527
+ "step": 35000
528
+ },
529
+ {
530
+ "epoch": 14.903442485306465,
531
+ "grad_norm": 91.38671112060547,
532
+ "learning_rate": 4.853884964682139e-05,
533
+ "loss": 84.7775,
534
+ "step": 35500
535
+ },
536
+ {
537
+ "epoch": 15.113350125944585,
538
+ "grad_norm": 107.55229187011719,
539
+ "learning_rate": 4.752976791120081e-05,
540
+ "loss": 83.9663,
541
+ "step": 36000
542
+ },
543
+ {
544
+ "epoch": 15.323257766582703,
545
+ "grad_norm": 93.78266906738281,
546
+ "learning_rate": 4.6520686175580225e-05,
547
+ "loss": 83.2307,
548
+ "step": 36500
549
+ },
550
+ {
551
+ "epoch": 15.533165407220823,
552
+ "grad_norm": 94.07606506347656,
553
+ "learning_rate": 4.551160443995964e-05,
554
+ "loss": 82.4276,
555
+ "step": 37000
556
+ },
557
+ {
558
+ "epoch": 15.743073047858942,
559
+ "grad_norm": 89.64346313476562,
560
+ "learning_rate": 4.4502522704339056e-05,
561
+ "loss": 81.7749,
562
+ "step": 37500
563
+ },
564
+ {
565
+ "epoch": 15.95298068849706,
566
+ "grad_norm": 90.450927734375,
567
+ "learning_rate": 4.349344096871847e-05,
568
+ "loss": 81.2509,
569
+ "step": 38000
570
+ },
571
+ {
572
+ "epoch": 16.162888329135182,
573
+ "grad_norm": 104.07128143310547,
574
+ "learning_rate": 4.248435923309788e-05,
575
+ "loss": 80.4306,
576
+ "step": 38500
577
+ },
578
+ {
579
+ "epoch": 16.3727959697733,
580
+ "grad_norm": 97.70896911621094,
581
+ "learning_rate": 4.14752774974773e-05,
582
+ "loss": 79.7886,
583
+ "step": 39000
584
+ },
585
+ {
586
+ "epoch": 16.582703610411418,
587
+ "grad_norm": 107.2480697631836,
588
+ "learning_rate": 4.046619576185671e-05,
589
+ "loss": 79.2786,
590
+ "step": 39500
591
+ },
592
+ {
593
+ "epoch": 16.79261125104954,
594
+ "grad_norm": 94.23933410644531,
595
+ "learning_rate": 3.9457114026236124e-05,
596
+ "loss": 78.7736,
597
+ "step": 40000
598
+ },
599
+ {
600
+ "epoch": 16.79261125104954,
601
+ "eval_accuracy": 0.5467926432656723,
602
+ "eval_loss": 70.87545013427734,
603
+ "eval_runtime": 2.7701,
604
+ "eval_samples_per_second": 172.917,
605
+ "eval_steps_per_second": 1.805,
606
+ "step": 40000
607
+ },
608
+ {
609
+ "epoch": 17.002518891687657,
610
+ "grad_norm": 95.26321411132812,
611
+ "learning_rate": 3.844803229061554e-05,
612
+ "loss": 78.2565,
613
+ "step": 40500
614
+ },
615
+ {
616
+ "epoch": 17.212426532325775,
617
+ "grad_norm": 97.84280395507812,
618
+ "learning_rate": 3.7438950554994956e-05,
619
+ "loss": 77.6301,
620
+ "step": 41000
621
+ },
622
+ {
623
+ "epoch": 17.422334172963897,
624
+ "grad_norm": 118.63774871826172,
625
+ "learning_rate": 3.642986881937437e-05,
626
+ "loss": 77.0761,
627
+ "step": 41500
628
+ },
629
+ {
630
+ "epoch": 17.632241813602015,
631
+ "grad_norm": 120.7638168334961,
632
+ "learning_rate": 3.542078708375379e-05,
633
+ "loss": 76.71,
634
+ "step": 42000
635
+ },
636
+ {
637
+ "epoch": 17.842149454240133,
638
+ "grad_norm": 96.10567474365234,
639
+ "learning_rate": 3.44117053481332e-05,
640
+ "loss": 76.231,
641
+ "step": 42500
642
+ },
643
+ {
644
+ "epoch": 18.052057094878254,
645
+ "grad_norm": 109.37013244628906,
646
+ "learning_rate": 3.340262361251261e-05,
647
+ "loss": 75.7584,
648
+ "step": 43000
649
+ },
650
+ {
651
+ "epoch": 18.261964735516372,
652
+ "grad_norm": 81.04769134521484,
653
+ "learning_rate": 3.239354187689203e-05,
654
+ "loss": 75.1399,
655
+ "step": 43500
656
+ },
657
+ {
658
+ "epoch": 18.471872376154494,
659
+ "grad_norm": 107.45811462402344,
660
+ "learning_rate": 3.138446014127144e-05,
661
+ "loss": 74.9177,
662
+ "step": 44000
663
+ },
664
+ {
665
+ "epoch": 18.68178001679261,
666
+ "grad_norm": 91.4131088256836,
667
+ "learning_rate": 3.0375378405650862e-05,
668
+ "loss": 74.4926,
669
+ "step": 44500
670
+ },
671
+ {
672
+ "epoch": 18.89168765743073,
673
+ "grad_norm": 103.88982391357422,
674
+ "learning_rate": 2.9366296670030274e-05,
675
+ "loss": 74.1645,
676
+ "step": 45000
677
+ },
678
+ {
679
+ "epoch": 19.10159529806885,
680
+ "grad_norm": 101.31675720214844,
681
+ "learning_rate": 2.8357214934409686e-05,
682
+ "loss": 73.6817,
683
+ "step": 45500
684
+ },
685
+ {
686
+ "epoch": 19.31150293870697,
687
+ "grad_norm": 97.94842529296875,
688
+ "learning_rate": 2.7348133198789106e-05,
689
+ "loss": 73.2026,
690
+ "step": 46000
691
+ },
692
+ {
693
+ "epoch": 19.521410579345087,
694
+ "grad_norm": 85.70989990234375,
695
+ "learning_rate": 2.6339051463168518e-05,
696
+ "loss": 72.9578,
697
+ "step": 46500
698
+ },
699
+ {
700
+ "epoch": 19.73131821998321,
701
+ "grad_norm": 93.49505615234375,
702
+ "learning_rate": 2.532996972754793e-05,
703
+ "loss": 72.7161,
704
+ "step": 47000
705
+ },
706
+ {
707
+ "epoch": 19.941225860621326,
708
+ "grad_norm": 89.21786499023438,
709
+ "learning_rate": 2.4320887991927346e-05,
710
+ "loss": 72.3488,
711
+ "step": 47500
712
+ },
713
+ {
714
+ "epoch": 20.151133501259444,
715
+ "grad_norm": 86.77904510498047,
716
+ "learning_rate": 2.331180625630676e-05,
717
+ "loss": 71.9474,
718
+ "step": 48000
719
+ },
720
+ {
721
+ "epoch": 20.361041141897566,
722
+ "grad_norm": 104.81689453125,
723
+ "learning_rate": 2.2302724520686177e-05,
724
+ "loss": 71.6556,
725
+ "step": 48500
726
+ },
727
+ {
728
+ "epoch": 20.570948782535684,
729
+ "grad_norm": 85.92950439453125,
730
+ "learning_rate": 2.129364278506559e-05,
731
+ "loss": 71.3194,
732
+ "step": 49000
733
+ },
734
+ {
735
+ "epoch": 20.7808564231738,
736
+ "grad_norm": 86.256103515625,
737
+ "learning_rate": 2.0284561049445005e-05,
738
+ "loss": 71.1533,
739
+ "step": 49500
740
+ },
741
+ {
742
+ "epoch": 20.990764063811923,
743
+ "grad_norm": 92.44305419921875,
744
+ "learning_rate": 1.927547931382442e-05,
745
+ "loss": 70.9704,
746
+ "step": 50000
747
+ },
748
+ {
749
+ "epoch": 20.990764063811923,
750
+ "eval_accuracy": 0.5510415790506631,
751
+ "eval_loss": 63.090843200683594,
752
+ "eval_runtime": 2.8803,
753
+ "eval_samples_per_second": 166.303,
754
+ "eval_steps_per_second": 1.736,
755
+ "step": 50000
756
+ },
757
+ {
758
+ "epoch": 21.20067170445004,
759
+ "grad_norm": 80.83171844482422,
760
+ "learning_rate": 1.8266397578203836e-05,
761
+ "loss": 70.5263,
762
+ "step": 50500
763
+ },
764
+ {
765
+ "epoch": 21.410579345088163,
766
+ "grad_norm": 86.68183898925781,
767
+ "learning_rate": 1.725731584258325e-05,
768
+ "loss": 70.3711,
769
+ "step": 51000
770
+ },
771
+ {
772
+ "epoch": 21.62048698572628,
773
+ "grad_norm": 75.10340118408203,
774
+ "learning_rate": 1.6248234106962664e-05,
775
+ "loss": 70.0563,
776
+ "step": 51500
777
+ },
778
+ {
779
+ "epoch": 21.8303946263644,
780
+ "grad_norm": 80.75882720947266,
781
+ "learning_rate": 1.523915237134208e-05,
782
+ "loss": 69.9151,
783
+ "step": 52000
784
+ },
785
+ {
786
+ "epoch": 22.04030226700252,
787
+ "grad_norm": 80.31729888916016,
788
+ "learning_rate": 1.4230070635721492e-05,
789
+ "loss": 69.7056,
790
+ "step": 52500
791
+ },
792
+ {
793
+ "epoch": 22.250209907640638,
794
+ "grad_norm": 79.79283905029297,
795
+ "learning_rate": 1.3220988900100908e-05,
796
+ "loss": 69.3878,
797
+ "step": 53000
798
+ },
799
+ {
800
+ "epoch": 22.460117548278756,
801
+ "grad_norm": 88.36482238769531,
802
+ "learning_rate": 1.2211907164480324e-05,
803
+ "loss": 69.2763,
804
+ "step": 53500
805
+ },
806
+ {
807
+ "epoch": 22.670025188916878,
808
+ "grad_norm": 82.19640350341797,
809
+ "learning_rate": 1.1202825428859738e-05,
810
+ "loss": 69.1046,
811
+ "step": 54000
812
+ },
813
+ {
814
+ "epoch": 22.879932829554996,
815
+ "grad_norm": 77.26073455810547,
816
+ "learning_rate": 1.0193743693239152e-05,
817
+ "loss": 68.8851,
818
+ "step": 54500
819
+ },
820
+ {
821
+ "epoch": 23.089840470193113,
822
+ "grad_norm": 89.86582946777344,
823
+ "learning_rate": 9.184661957618567e-06,
824
+ "loss": 68.7541,
825
+ "step": 55000
826
+ },
827
+ {
828
+ "epoch": 23.299748110831235,
829
+ "grad_norm": 70.37703704833984,
830
+ "learning_rate": 8.175580221997981e-06,
831
+ "loss": 68.5219,
832
+ "step": 55500
833
+ },
834
+ {
835
+ "epoch": 23.509655751469353,
836
+ "grad_norm": 71.59619140625,
837
+ "learning_rate": 7.166498486377397e-06,
838
+ "loss": 68.3648,
839
+ "step": 56000
840
+ },
841
+ {
842
+ "epoch": 23.719563392107474,
843
+ "grad_norm": 69.95408630371094,
844
+ "learning_rate": 6.157416750756812e-06,
845
+ "loss": 68.3101,
846
+ "step": 56500
847
+ },
848
+ {
849
+ "epoch": 23.929471032745592,
850
+ "grad_norm": 71.7118148803711,
851
+ "learning_rate": 5.148335015136227e-06,
852
+ "loss": 68.1746,
853
+ "step": 57000
854
+ },
855
+ {
856
+ "epoch": 24.13937867338371,
857
+ "grad_norm": 67.84393310546875,
858
+ "learning_rate": 4.139253279515641e-06,
859
+ "loss": 67.9893,
860
+ "step": 57500
861
+ },
862
+ {
863
+ "epoch": 24.349286314021832,
864
+ "grad_norm": 70.03734588623047,
865
+ "learning_rate": 3.130171543895056e-06,
866
+ "loss": 67.9278,
867
+ "step": 58000
868
+ },
869
+ {
870
+ "epoch": 24.55919395465995,
871
+ "grad_norm": 67.93866729736328,
872
+ "learning_rate": 2.1210898082744702e-06,
873
+ "loss": 67.8202,
874
+ "step": 58500
875
+ },
876
+ {
877
+ "epoch": 24.769101595298068,
878
+ "grad_norm": 61.62685775756836,
879
+ "learning_rate": 1.112008072653885e-06,
880
+ "loss": 67.7249,
881
+ "step": 59000
882
+ },
883
+ {
884
+ "epoch": 24.97900923593619,
885
+ "grad_norm": 63.67091751098633,
886
+ "learning_rate": 1.0292633703329971e-07,
887
+ "loss": 67.6809,
888
+ "step": 59500
889
+ },
890
+ {
891
+ "epoch": 25.0,
892
+ "step": 59550,
893
+ "total_flos": 7.577162542158336e+17,
894
+ "train_loss": 301.8068682845954,
895
+ "train_runtime": 49685.0337,
896
+ "train_samples_per_second": 115.044,
897
+ "train_steps_per_second": 1.199
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": 7.577162542158336e+17,
918
+ "train_batch_size": 96,
919
+ "trial_name": null,
920
+ "trial_params": null
921
+ }