liswei commited on
Commit
727b22b
1 Parent(s): 6598b31

End of training

Browse files
README.md CHANGED
@@ -2,6 +2,7 @@
2
  base_model: liswei/OpenELM-1_1B-zh-cp
3
  tags:
4
  - llama-factory
 
5
  - generated_from_trainer
6
  model-index:
7
  - name: OpenELM-1_1B-zh-sft
@@ -13,9 +14,9 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  # OpenELM-1_1B-zh-sft
15
 
16
- This model is a fine-tuned version of [liswei/OpenELM-1_1B-zh-cp](https://huggingface.co/liswei/OpenELM-1_1B-zh-cp) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 1.3570
19
 
20
  ## Model description
21
 
 
2
  base_model: liswei/OpenELM-1_1B-zh-cp
3
  tags:
4
  - llama-factory
5
+ - full
6
  - generated_from_trainer
7
  model-index:
8
  - name: OpenELM-1_1B-zh-sft
 
14
 
15
  # OpenELM-1_1B-zh-sft
16
 
17
+ This model is a fine-tuned version of [liswei/OpenELM-1_1B-zh-cp](https://huggingface.co/liswei/OpenELM-1_1B-zh-cp) on the en2tw-alignment-sft and the TaiwanChat datasets.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 1.3355
20
 
21
  ## Model description
22
 
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9994709927702345,
3
+ "eval_loss": 1.3354638814926147,
4
+ "eval_runtime": 1576.9499,
5
+ "eval_samples_per_second": 38.356,
6
+ "eval_steps_per_second": 0.4,
7
+ "total_flos": 3.7897423231228314e+17,
8
+ "train_loss": 1.4671159325642071,
9
+ "train_runtime": 64163.8855,
10
+ "train_samples_per_second": 8.484,
11
+ "train_steps_per_second": 0.022
12
+ }
eval_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9994709927702345,
3
+ "eval_loss": 1.3354638814926147,
4
+ "eval_runtime": 1576.9499,
5
+ "eval_samples_per_second": 38.356,
6
+ "eval_steps_per_second": 0.4
7
+ }
runs/May31_22-40-32_t9onywctr1717164743911-xnk6s/events.out.tfevents.1717232705.t9onywctr1717164743911-xnk6s.11369.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e97b95183e4b3130295361d76c5d9dd2c5f73bf8bff7afe0c8f1b655f62ce87
3
+ size 359
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9994709927702345,
3
+ "total_flos": 3.7897423231228314e+17,
4
+ "train_loss": 1.4671159325642071,
5
+ "train_runtime": 64163.8855,
6
+ "train_samples_per_second": 8.484,
7
+ "train_steps_per_second": 0.022
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1045 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9994709927702345,
5
+ "eval_steps": 500,
6
+ "global_step": 1417,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.007053429730206313,
13
+ "grad_norm": 1.3042659759521484,
14
+ "learning_rate": 7.042253521126762e-06,
15
+ "loss": 2.258,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.014106859460412626,
20
+ "grad_norm": 0.738134503364563,
21
+ "learning_rate": 1.4084507042253523e-05,
22
+ "loss": 2.0805,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.02116028919061894,
27
+ "grad_norm": 0.5946909785270691,
28
+ "learning_rate": 2.112676056338028e-05,
29
+ "loss": 1.9582,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.02821371892082525,
34
+ "grad_norm": 0.5485147833824158,
35
+ "learning_rate": 2.8169014084507046e-05,
36
+ "loss": 1.9138,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.03526714865103157,
41
+ "grad_norm": 0.5394290089607239,
42
+ "learning_rate": 3.5211267605633805e-05,
43
+ "loss": 1.8736,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.04232057838123788,
48
+ "grad_norm": 0.49499747157096863,
49
+ "learning_rate": 4.225352112676056e-05,
50
+ "loss": 1.7826,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.04937400811144419,
55
+ "grad_norm": 0.48521119356155396,
56
+ "learning_rate": 4.929577464788733e-05,
57
+ "loss": 1.7868,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.0564274378416505,
62
+ "grad_norm": 0.5505830645561218,
63
+ "learning_rate": 5.633802816901409e-05,
64
+ "loss": 1.754,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.06348086757185682,
69
+ "grad_norm": 0.512299656867981,
70
+ "learning_rate": 6.338028169014085e-05,
71
+ "loss": 1.7356,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.07053429730206313,
76
+ "grad_norm": 0.47940096259117126,
77
+ "learning_rate": 7.042253521126761e-05,
78
+ "loss": 1.6952,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.07758772703226945,
83
+ "grad_norm": 0.48518213629722595,
84
+ "learning_rate": 7.746478873239437e-05,
85
+ "loss": 1.6853,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.08464115676247576,
90
+ "grad_norm": 0.5208117961883545,
91
+ "learning_rate": 8.450704225352113e-05,
92
+ "loss": 1.7022,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.09169458649268207,
97
+ "grad_norm": 0.5312570333480835,
98
+ "learning_rate": 9.15492957746479e-05,
99
+ "loss": 1.6371,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.09874801622288838,
104
+ "grad_norm": 0.5060498118400574,
105
+ "learning_rate": 9.859154929577466e-05,
106
+ "loss": 1.665,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.1058014459530947,
111
+ "grad_norm": 0.482503205537796,
112
+ "learning_rate": 9.999028628636837e-05,
113
+ "loss": 1.6445,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.112854875683301,
118
+ "grad_norm": 0.5026859045028687,
119
+ "learning_rate": 9.99508307932129e-05,
120
+ "loss": 1.6505,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.11990830541350732,
125
+ "grad_norm": 0.508380651473999,
126
+ "learning_rate": 9.988105034827552e-05,
127
+ "loss": 1.6513,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.12696173514371364,
132
+ "grad_norm": 0.4797461926937103,
133
+ "learning_rate": 9.978098731498833e-05,
134
+ "loss": 1.6382,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.13401516487391996,
139
+ "grad_norm": 0.5087615847587585,
140
+ "learning_rate": 9.965070244122301e-05,
141
+ "loss": 1.6464,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.14106859460412627,
146
+ "grad_norm": 0.4981568455696106,
147
+ "learning_rate": 9.949027482241107e-05,
148
+ "loss": 1.6105,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.14812202433433258,
153
+ "grad_norm": 0.4438849091529846,
154
+ "learning_rate": 9.929980185352526e-05,
155
+ "loss": 1.6194,
156
+ "step": 210
157
+ },
158
+ {
159
+ "epoch": 0.1551754540645389,
160
+ "grad_norm": 0.5086779594421387,
161
+ "learning_rate": 9.907939916995152e-05,
162
+ "loss": 1.6073,
163
+ "step": 220
164
+ },
165
+ {
166
+ "epoch": 0.1622288837947452,
167
+ "grad_norm": 0.5140203833580017,
168
+ "learning_rate": 9.88292005772872e-05,
169
+ "loss": 1.6031,
170
+ "step": 230
171
+ },
172
+ {
173
+ "epoch": 0.16928231352495152,
174
+ "grad_norm": 0.5097965002059937,
175
+ "learning_rate": 9.854935797010821e-05,
176
+ "loss": 1.6216,
177
+ "step": 240
178
+ },
179
+ {
180
+ "epoch": 0.17633574325515783,
181
+ "grad_norm": 0.5518028140068054,
182
+ "learning_rate": 9.824004123975435e-05,
183
+ "loss": 1.5585,
184
+ "step": 250
185
+ },
186
+ {
187
+ "epoch": 0.18338917298536414,
188
+ "grad_norm": 0.48416197299957275,
189
+ "learning_rate": 9.790143817118904e-05,
190
+ "loss": 1.5739,
191
+ "step": 260
192
+ },
193
+ {
194
+ "epoch": 0.19044260271557045,
195
+ "grad_norm": 0.4688737690448761,
196
+ "learning_rate": 9.753375432899581e-05,
197
+ "loss": 1.571,
198
+ "step": 270
199
+ },
200
+ {
201
+ "epoch": 0.19749603244577676,
202
+ "grad_norm": 0.4657231569290161,
203
+ "learning_rate": 9.713721293258078e-05,
204
+ "loss": 1.5614,
205
+ "step": 280
206
+ },
207
+ {
208
+ "epoch": 0.20454946217598308,
209
+ "grad_norm": 0.47847288846969604,
210
+ "learning_rate": 9.671205472065707e-05,
211
+ "loss": 1.5474,
212
+ "step": 290
213
+ },
214
+ {
215
+ "epoch": 0.2116028919061894,
216
+ "grad_norm": 0.5418571829795837,
217
+ "learning_rate": 9.625853780509335e-05,
218
+ "loss": 1.5562,
219
+ "step": 300
220
+ },
221
+ {
222
+ "epoch": 0.2186563216363957,
223
+ "grad_norm": 0.5494666695594788,
224
+ "learning_rate": 9.577693751421506e-05,
225
+ "loss": 1.5585,
226
+ "step": 310
227
+ },
228
+ {
229
+ "epoch": 0.225709751366602,
230
+ "grad_norm": 0.5272155404090881,
231
+ "learning_rate": 9.52675462256537e-05,
232
+ "loss": 1.546,
233
+ "step": 320
234
+ },
235
+ {
236
+ "epoch": 0.23276318109680832,
237
+ "grad_norm": 0.5072800517082214,
238
+ "learning_rate": 9.473067318884552e-05,
239
+ "loss": 1.5395,
240
+ "step": 330
241
+ },
242
+ {
243
+ "epoch": 0.23981661082701464,
244
+ "grad_norm": 0.48417559266090393,
245
+ "learning_rate": 9.416664433728748e-05,
246
+ "loss": 1.5342,
247
+ "step": 340
248
+ },
249
+ {
250
+ "epoch": 0.24687004055722095,
251
+ "grad_norm": 0.49016687273979187,
252
+ "learning_rate": 9.357580209066416e-05,
253
+ "loss": 1.5155,
254
+ "step": 350
255
+ },
256
+ {
257
+ "epoch": 0.2539234702874273,
258
+ "grad_norm": 0.48039939999580383,
259
+ "learning_rate": 9.295850514696628e-05,
260
+ "loss": 1.5194,
261
+ "step": 360
262
+ },
263
+ {
264
+ "epoch": 0.2609769000176336,
265
+ "grad_norm": 0.4710429906845093,
266
+ "learning_rate": 9.23151282647265e-05,
267
+ "loss": 1.4984,
268
+ "step": 370
269
+ },
270
+ {
271
+ "epoch": 0.2680303297478399,
272
+ "grad_norm": 0.46802473068237305,
273
+ "learning_rate": 9.164606203550497e-05,
274
+ "loss": 1.5316,
275
+ "step": 380
276
+ },
277
+ {
278
+ "epoch": 0.2750837594780462,
279
+ "grad_norm": 0.46727123856544495,
280
+ "learning_rate": 9.095171264676283e-05,
281
+ "loss": 1.5478,
282
+ "step": 390
283
+ },
284
+ {
285
+ "epoch": 0.28213718920825254,
286
+ "grad_norm": 0.4999746084213257,
287
+ "learning_rate": 9.023250163526731e-05,
288
+ "loss": 1.5335,
289
+ "step": 400
290
+ },
291
+ {
292
+ "epoch": 0.2891906189384588,
293
+ "grad_norm": 0.4812432825565338,
294
+ "learning_rate": 8.948886563117844e-05,
295
+ "loss": 1.5063,
296
+ "step": 410
297
+ },
298
+ {
299
+ "epoch": 0.29624404866866516,
300
+ "grad_norm": 0.48060664534568787,
301
+ "learning_rate": 8.872125609297271e-05,
302
+ "loss": 1.5039,
303
+ "step": 420
304
+ },
305
+ {
306
+ "epoch": 0.30329747839887145,
307
+ "grad_norm": 0.4972708523273468,
308
+ "learning_rate": 8.793013903336428e-05,
309
+ "loss": 1.5414,
310
+ "step": 430
311
+ },
312
+ {
313
+ "epoch": 0.3103509081290778,
314
+ "grad_norm": 0.4610942006111145,
315
+ "learning_rate": 8.711599473639058e-05,
316
+ "loss": 1.4808,
317
+ "step": 440
318
+ },
319
+ {
320
+ "epoch": 0.31740433785928407,
321
+ "grad_norm": 0.45125675201416016,
322
+ "learning_rate": 8.627931746583378e-05,
323
+ "loss": 1.5097,
324
+ "step": 450
325
+ },
326
+ {
327
+ "epoch": 0.3244577675894904,
328
+ "grad_norm": 0.46617043018341064,
329
+ "learning_rate": 8.542061516515512e-05,
330
+ "loss": 1.4708,
331
+ "step": 460
332
+ },
333
+ {
334
+ "epoch": 0.3315111973196967,
335
+ "grad_norm": 0.5127195119857788,
336
+ "learning_rate": 8.454040914912456e-05,
337
+ "loss": 1.4894,
338
+ "step": 470
339
+ },
340
+ {
341
+ "epoch": 0.33856462704990303,
342
+ "grad_norm": 0.45094090700149536,
343
+ "learning_rate": 8.363923378733255e-05,
344
+ "loss": 1.4554,
345
+ "step": 480
346
+ },
347
+ {
348
+ "epoch": 0.3456180567801093,
349
+ "grad_norm": 0.48269587755203247,
350
+ "learning_rate": 8.271763617977641e-05,
351
+ "loss": 1.4732,
352
+ "step": 490
353
+ },
354
+ {
355
+ "epoch": 0.35267148651031566,
356
+ "grad_norm": 0.4571676552295685,
357
+ "learning_rate": 8.177617582471813e-05,
358
+ "loss": 1.4543,
359
+ "step": 500
360
+ },
361
+ {
362
+ "epoch": 0.35267148651031566,
363
+ "eval_loss": 1.4671958684921265,
364
+ "eval_runtime": 1577.1702,
365
+ "eval_samples_per_second": 38.35,
366
+ "eval_steps_per_second": 0.4,
367
+ "step": 500
368
+ },
369
+ {
370
+ "epoch": 0.35972491624052194,
371
+ "grad_norm": 0.4831838011741638,
372
+ "learning_rate": 8.081542427901503e-05,
373
+ "loss": 1.4805,
374
+ "step": 510
375
+ },
376
+ {
377
+ "epoch": 0.3667783459707283,
378
+ "grad_norm": 0.47973912954330444,
379
+ "learning_rate": 7.983596481113005e-05,
380
+ "loss": 1.4767,
381
+ "step": 520
382
+ },
383
+ {
384
+ "epoch": 0.37383177570093457,
385
+ "grad_norm": 0.4587724208831787,
386
+ "learning_rate": 7.883839204703165e-05,
387
+ "loss": 1.4783,
388
+ "step": 530
389
+ },
390
+ {
391
+ "epoch": 0.3808852054311409,
392
+ "grad_norm": 0.48470768332481384,
393
+ "learning_rate": 7.782331160919877e-05,
394
+ "loss": 1.4691,
395
+ "step": 540
396
+ },
397
+ {
398
+ "epoch": 0.3879386351613472,
399
+ "grad_norm": 0.4692317247390747,
400
+ "learning_rate": 7.679133974894983e-05,
401
+ "loss": 1.4651,
402
+ "step": 550
403
+ },
404
+ {
405
+ "epoch": 0.39499206489155353,
406
+ "grad_norm": 0.5106418132781982,
407
+ "learning_rate": 7.574310297231897e-05,
408
+ "loss": 1.4651,
409
+ "step": 560
410
+ },
411
+ {
412
+ "epoch": 0.4020454946217598,
413
+ "grad_norm": 0.4444584250450134,
414
+ "learning_rate": 7.46792376597067e-05,
415
+ "loss": 1.459,
416
+ "step": 570
417
+ },
418
+ {
419
+ "epoch": 0.40909892435196615,
420
+ "grad_norm": 0.44441869854927063,
421
+ "learning_rate": 7.360038967953598e-05,
422
+ "loss": 1.466,
423
+ "step": 580
424
+ },
425
+ {
426
+ "epoch": 0.41615235408217244,
427
+ "grad_norm": 0.4592408537864685,
428
+ "learning_rate": 7.250721399614786e-05,
429
+ "loss": 1.4527,
430
+ "step": 590
431
+ },
432
+ {
433
+ "epoch": 0.4232057838123788,
434
+ "grad_norm": 0.48659372329711914,
435
+ "learning_rate": 7.140037427217545e-05,
436
+ "loss": 1.4409,
437
+ "step": 600
438
+ },
439
+ {
440
+ "epoch": 0.43025921354258506,
441
+ "grad_norm": 0.49556177854537964,
442
+ "learning_rate": 7.02805424656368e-05,
443
+ "loss": 1.4378,
444
+ "step": 610
445
+ },
446
+ {
447
+ "epoch": 0.4373126432727914,
448
+ "grad_norm": 0.4970695674419403,
449
+ "learning_rate": 6.914839842199187e-05,
450
+ "loss": 1.4209,
451
+ "step": 620
452
+ },
453
+ {
454
+ "epoch": 0.4443660730029977,
455
+ "grad_norm": 0.4574239253997803,
456
+ "learning_rate": 6.800462946141109e-05,
457
+ "loss": 1.4215,
458
+ "step": 630
459
+ },
460
+ {
461
+ "epoch": 0.451419502733204,
462
+ "grad_norm": 0.44600942730903625,
463
+ "learning_rate": 6.684992996150599e-05,
464
+ "loss": 1.4481,
465
+ "step": 640
466
+ },
467
+ {
468
+ "epoch": 0.4584729324634103,
469
+ "grad_norm": 0.4874365031719208,
470
+ "learning_rate": 6.568500093577525e-05,
471
+ "loss": 1.4357,
472
+ "step": 650
473
+ },
474
+ {
475
+ "epoch": 0.46552636219361665,
476
+ "grad_norm": 0.44276514649391174,
477
+ "learning_rate": 6.451054960802224e-05,
478
+ "loss": 1.4229,
479
+ "step": 660
480
+ },
481
+ {
482
+ "epoch": 0.47257979192382293,
483
+ "grad_norm": 0.4567689895629883,
484
+ "learning_rate": 6.332728898300221e-05,
485
+ "loss": 1.3891,
486
+ "step": 670
487
+ },
488
+ {
489
+ "epoch": 0.4796332216540293,
490
+ "grad_norm": 0.4700818359851837,
491
+ "learning_rate": 6.213593741355993e-05,
492
+ "loss": 1.3942,
493
+ "step": 680
494
+ },
495
+ {
496
+ "epoch": 0.48668665138423556,
497
+ "grad_norm": 0.510308027267456,
498
+ "learning_rate": 6.093721816452052e-05,
499
+ "loss": 1.4321,
500
+ "step": 690
501
+ },
502
+ {
503
+ "epoch": 0.4937400811144419,
504
+ "grad_norm": 0.4537096321582794,
505
+ "learning_rate": 5.973185897359828e-05,
506
+ "loss": 1.4058,
507
+ "step": 700
508
+ },
509
+ {
510
+ "epoch": 0.5007935108446482,
511
+ "grad_norm": 0.46621763706207275,
512
+ "learning_rate": 5.852059160958988e-05,
513
+ "loss": 1.4264,
514
+ "step": 710
515
+ },
516
+ {
517
+ "epoch": 0.5078469405748546,
518
+ "grad_norm": 0.4476553201675415,
519
+ "learning_rate": 5.730415142812059e-05,
520
+ "loss": 1.3943,
521
+ "step": 720
522
+ },
523
+ {
524
+ "epoch": 0.5149003703050609,
525
+ "grad_norm": 0.4877742528915405,
526
+ "learning_rate": 5.6083276925212624e-05,
527
+ "loss": 1.4114,
528
+ "step": 730
529
+ },
530
+ {
531
+ "epoch": 0.5219538000352671,
532
+ "grad_norm": 0.4259643256664276,
533
+ "learning_rate": 5.485870928894725e-05,
534
+ "loss": 1.3905,
535
+ "step": 740
536
+ },
537
+ {
538
+ "epoch": 0.5290072297654734,
539
+ "grad_norm": 0.4499744772911072,
540
+ "learning_rate": 5.3631191949492354e-05,
541
+ "loss": 1.3973,
542
+ "step": 750
543
+ },
544
+ {
545
+ "epoch": 0.5360606594956798,
546
+ "grad_norm": 0.45207515358924866,
547
+ "learning_rate": 5.240147012776899e-05,
548
+ "loss": 1.3957,
549
+ "step": 760
550
+ },
551
+ {
552
+ "epoch": 0.5431140892258861,
553
+ "grad_norm": 0.4296036958694458,
554
+ "learning_rate": 5.117029038303065e-05,
555
+ "loss": 1.4023,
556
+ "step": 770
557
+ },
558
+ {
559
+ "epoch": 0.5501675189560924,
560
+ "grad_norm": 0.4590493142604828,
561
+ "learning_rate": 4.993840015963016e-05,
562
+ "loss": 1.414,
563
+ "step": 780
564
+ },
565
+ {
566
+ "epoch": 0.5572209486862987,
567
+ "grad_norm": 0.4364815652370453,
568
+ "learning_rate": 4.870654733324907e-05,
569
+ "loss": 1.3939,
570
+ "step": 790
571
+ },
572
+ {
573
+ "epoch": 0.5642743784165051,
574
+ "grad_norm": 0.4730934798717499,
575
+ "learning_rate": 4.747547975686538e-05,
576
+ "loss": 1.3951,
577
+ "step": 800
578
+ },
579
+ {
580
+ "epoch": 0.5713278081467114,
581
+ "grad_norm": 0.4653618037700653,
582
+ "learning_rate": 4.6245944806734966e-05,
583
+ "loss": 1.3919,
584
+ "step": 810
585
+ },
586
+ {
587
+ "epoch": 0.5783812378769176,
588
+ "grad_norm": 0.44776442646980286,
589
+ "learning_rate": 4.50186889286623e-05,
590
+ "loss": 1.401,
591
+ "step": 820
592
+ },
593
+ {
594
+ "epoch": 0.5854346676071239,
595
+ "grad_norm": 0.47814151644706726,
596
+ "learning_rate": 4.379445718483626e-05,
597
+ "loss": 1.4014,
598
+ "step": 830
599
+ },
600
+ {
601
+ "epoch": 0.5924880973373303,
602
+ "grad_norm": 0.4623441994190216,
603
+ "learning_rate": 4.257399280150576e-05,
604
+ "loss": 1.4101,
605
+ "step": 840
606
+ },
607
+ {
608
+ "epoch": 0.5995415270675366,
609
+ "grad_norm": 0.47123944759368896,
610
+ "learning_rate": 4.135803671777011e-05,
611
+ "loss": 1.3687,
612
+ "step": 850
613
+ },
614
+ {
615
+ "epoch": 0.6065949567977429,
616
+ "grad_norm": 0.4534803628921509,
617
+ "learning_rate": 4.014732713575769e-05,
618
+ "loss": 1.3675,
619
+ "step": 860
620
+ },
621
+ {
622
+ "epoch": 0.6136483865279492,
623
+ "grad_norm": 0.4451278746128082,
624
+ "learning_rate": 3.894259907246652e-05,
625
+ "loss": 1.3705,
626
+ "step": 870
627
+ },
628
+ {
629
+ "epoch": 0.6207018162581556,
630
+ "grad_norm": 0.40993040800094604,
631
+ "learning_rate": 3.774458391353821e-05,
632
+ "loss": 1.3944,
633
+ "step": 880
634
+ },
635
+ {
636
+ "epoch": 0.6277552459883619,
637
+ "grad_norm": 0.43390119075775146,
638
+ "learning_rate": 3.655400896923672e-05,
639
+ "loss": 1.3898,
640
+ "step": 890
641
+ },
642
+ {
643
+ "epoch": 0.6348086757185681,
644
+ "grad_norm": 0.469901442527771,
645
+ "learning_rate": 3.537159703290112e-05,
646
+ "loss": 1.3975,
647
+ "step": 900
648
+ },
649
+ {
650
+ "epoch": 0.6418621054487744,
651
+ "grad_norm": 0.4519672989845276,
652
+ "learning_rate": 3.419806594214063e-05,
653
+ "loss": 1.3705,
654
+ "step": 910
655
+ },
656
+ {
657
+ "epoch": 0.6489155351789808,
658
+ "grad_norm": 0.44010812044143677,
659
+ "learning_rate": 3.303412814303827e-05,
660
+ "loss": 1.3673,
661
+ "step": 920
662
+ },
663
+ {
664
+ "epoch": 0.6559689649091871,
665
+ "grad_norm": 0.45562589168548584,
666
+ "learning_rate": 3.1880490257627646e-05,
667
+ "loss": 1.3754,
668
+ "step": 930
669
+ },
670
+ {
671
+ "epoch": 0.6630223946393934,
672
+ "grad_norm": 0.4654683768749237,
673
+ "learning_rate": 3.0737852654905497e-05,
674
+ "loss": 1.3809,
675
+ "step": 940
676
+ },
677
+ {
678
+ "epoch": 0.6700758243695997,
679
+ "grad_norm": 0.40361422300338745,
680
+ "learning_rate": 2.960690902564044e-05,
681
+ "loss": 1.3432,
682
+ "step": 950
683
+ },
684
+ {
685
+ "epoch": 0.6771292540998061,
686
+ "grad_norm": 0.45382988452911377,
687
+ "learning_rate": 2.8488345961236075e-05,
688
+ "loss": 1.3686,
689
+ "step": 960
690
+ },
691
+ {
692
+ "epoch": 0.6841826838300124,
693
+ "grad_norm": 0.4258984327316284,
694
+ "learning_rate": 2.7382842536904075e-05,
695
+ "loss": 1.3755,
696
+ "step": 970
697
+ },
698
+ {
699
+ "epoch": 0.6912361135602186,
700
+ "grad_norm": 0.5345839858055115,
701
+ "learning_rate": 2.629106989940025e-05,
702
+ "loss": 1.3793,
703
+ "step": 980
704
+ },
705
+ {
706
+ "epoch": 0.6982895432904249,
707
+ "grad_norm": 0.42425984144210815,
708
+ "learning_rate": 2.5213690859574058e-05,
709
+ "loss": 1.3385,
710
+ "step": 990
711
+ },
712
+ {
713
+ "epoch": 0.7053429730206313,
714
+ "grad_norm": 0.43758624792099,
715
+ "learning_rate": 2.4151359489978733e-05,
716
+ "loss": 1.3683,
717
+ "step": 1000
718
+ },
719
+ {
720
+ "epoch": 0.7053429730206313,
721
+ "eval_loss": 1.3570424318313599,
722
+ "eval_runtime": 1575.8807,
723
+ "eval_samples_per_second": 38.382,
724
+ "eval_steps_per_second": 0.4,
725
+ "step": 1000
726
+ },
727
+ {
728
+ "epoch": 0.7123964027508376,
729
+ "grad_norm": 0.4325385093688965,
730
+ "learning_rate": 2.3104720727786467e-05,
731
+ "loss": 1.3411,
732
+ "step": 1010
733
+ },
734
+ {
735
+ "epoch": 0.7194498324810439,
736
+ "grad_norm": 0.41445136070251465,
737
+ "learning_rate": 2.2074409983249545e-05,
738
+ "loss": 1.3495,
739
+ "step": 1020
740
+ },
741
+ {
742
+ "epoch": 0.7265032622112502,
743
+ "grad_norm": 0.4271891415119171,
744
+ "learning_rate": 2.106105275394533e-05,
745
+ "loss": 1.3505,
746
+ "step": 1030
747
+ },
748
+ {
749
+ "epoch": 0.7335566919414566,
750
+ "grad_norm": 0.4772493243217468,
751
+ "learning_rate": 2.0065264245039128e-05,
752
+ "loss": 1.317,
753
+ "step": 1040
754
+ },
755
+ {
756
+ "epoch": 0.7406101216716628,
757
+ "grad_norm": 0.43762919306755066,
758
+ "learning_rate": 1.9087648995795658e-05,
759
+ "loss": 1.349,
760
+ "step": 1050
761
+ },
762
+ {
763
+ "epoch": 0.7476635514018691,
764
+ "grad_norm": 0.4569909870624542,
765
+ "learning_rate": 1.8128800512565513e-05,
766
+ "loss": 1.3473,
767
+ "step": 1060
768
+ },
769
+ {
770
+ "epoch": 0.7547169811320755,
771
+ "grad_norm": 0.4433094561100006,
772
+ "learning_rate": 1.7189300908469912e-05,
773
+ "loss": 1.3386,
774
+ "step": 1070
775
+ },
776
+ {
777
+ "epoch": 0.7617704108622818,
778
+ "grad_norm": 0.48191553354263306,
779
+ "learning_rate": 1.6269720550002077e-05,
780
+ "loss": 1.3765,
781
+ "step": 1080
782
+ },
783
+ {
784
+ "epoch": 0.7688238405924881,
785
+ "grad_norm": 0.40601858496665955,
786
+ "learning_rate": 1.5370617710760004e-05,
787
+ "loss": 1.3354,
788
+ "step": 1090
789
+ },
790
+ {
791
+ "epoch": 0.7758772703226944,
792
+ "grad_norm": 0.43439218401908875,
793
+ "learning_rate": 1.4492538232520791e-05,
794
+ "loss": 1.3404,
795
+ "step": 1100
796
+ },
797
+ {
798
+ "epoch": 0.7829307000529008,
799
+ "grad_norm": 0.47002339363098145,
800
+ "learning_rate": 1.3636015193862212e-05,
801
+ "loss": 1.3562,
802
+ "step": 1110
803
+ },
804
+ {
805
+ "epoch": 0.7899841297831071,
806
+ "grad_norm": 0.44879859685897827,
807
+ "learning_rate": 1.2801568586532837e-05,
808
+ "loss": 1.3652,
809
+ "step": 1120
810
+ },
811
+ {
812
+ "epoch": 0.7970375595133133,
813
+ "grad_norm": 0.44707298278808594,
814
+ "learning_rate": 1.1989704999767149e-05,
815
+ "loss": 1.3332,
816
+ "step": 1130
817
+ },
818
+ {
819
+ "epoch": 0.8040909892435196,
820
+ "grad_norm": 0.45202744007110596,
821
+ "learning_rate": 1.1200917312737135e-05,
822
+ "loss": 1.3488,
823
+ "step": 1140
824
+ },
825
+ {
826
+ "epoch": 0.811144418973726,
827
+ "grad_norm": 0.4314928650856018,
828
+ "learning_rate": 1.043568439532735e-05,
829
+ "loss": 1.3592,
830
+ "step": 1150
831
+ },
832
+ {
833
+ "epoch": 0.8181978487039323,
834
+ "grad_norm": 0.47974613308906555,
835
+ "learning_rate": 9.69447081741487e-06,
836
+ "loss": 1.3273,
837
+ "step": 1160
838
+ },
839
+ {
840
+ "epoch": 0.8252512784341386,
841
+ "grad_norm": 0.49341320991516113,
842
+ "learning_rate": 8.97772656683083e-06,
843
+ "loss": 1.3495,
844
+ "step": 1170
845
+ },
846
+ {
847
+ "epoch": 0.8323047081643449,
848
+ "grad_norm": 0.45349812507629395,
849
+ "learning_rate": 8.2858867761745e-06,
850
+ "loss": 1.3349,
851
+ "step": 1180
852
+ },
853
+ {
854
+ "epoch": 0.8393581378945513,
855
+ "grad_norm": 0.4800957143306732,
856
+ "learning_rate": 7.619371458646007e-06,
857
+ "loss": 1.3282,
858
+ "step": 1190
859
+ },
860
+ {
861
+ "epoch": 0.8464115676247576,
862
+ "grad_norm": 0.41283947229385376,
863
+ "learning_rate": 6.978585253058101e-06,
864
+ "loss": 1.3289,
865
+ "step": 1200
866
+ },
867
+ {
868
+ "epoch": 0.8534649973549638,
869
+ "grad_norm": 0.3885229825973511,
870
+ "learning_rate": 6.363917178181389e-06,
871
+ "loss": 1.3251,
872
+ "step": 1210
873
+ },
874
+ {
875
+ "epoch": 0.8605184270851701,
876
+ "grad_norm": 0.3919270634651184,
877
+ "learning_rate": 5.7757403965726655e-06,
878
+ "loss": 1.3552,
879
+ "step": 1220
880
+ },
881
+ {
882
+ "epoch": 0.8675718568153765,
883
+ "grad_norm": 0.4169398546218872,
884
+ "learning_rate": 5.214411988029355e-06,
885
+ "loss": 1.332,
886
+ "step": 1230
887
+ },
888
+ {
889
+ "epoch": 0.8746252865455828,
890
+ "grad_norm": 0.4624149799346924,
891
+ "learning_rate": 4.680272732807834e-06,
892
+ "loss": 1.3261,
893
+ "step": 1240
894
+ },
895
+ {
896
+ "epoch": 0.8816787162757891,
897
+ "grad_norm": 0.404799222946167,
898
+ "learning_rate": 4.173646904737027e-06,
899
+ "loss": 1.3352,
900
+ "step": 1250
901
+ },
902
+ {
903
+ "epoch": 0.8887321460059954,
904
+ "grad_norm": 0.45688724517822266,
905
+ "learning_rate": 3.694842074353061e-06,
906
+ "loss": 1.3586,
907
+ "step": 1260
908
+ },
909
+ {
910
+ "epoch": 0.8957855757362018,
911
+ "grad_norm": 0.42276087403297424,
912
+ "learning_rate": 3.244148922174445e-06,
913
+ "loss": 1.3653,
914
+ "step": 1270
915
+ },
916
+ {
917
+ "epoch": 0.902839005466408,
918
+ "grad_norm": 0.446267306804657,
919
+ "learning_rate": 2.8218410622310286e-06,
920
+ "loss": 1.3674,
921
+ "step": 1280
922
+ },
923
+ {
924
+ "epoch": 0.9098924351966143,
925
+ "grad_norm": 0.48505640029907227,
926
+ "learning_rate": 2.4281748759540534e-06,
927
+ "loss": 1.3289,
928
+ "step": 1290
929
+ },
930
+ {
931
+ "epoch": 0.9169458649268206,
932
+ "grad_norm": 0.43678343296051025,
933
+ "learning_rate": 2.0633893565279615e-06,
934
+ "loss": 1.3405,
935
+ "step": 1300
936
+ },
937
+ {
938
+ "epoch": 0.923999294657027,
939
+ "grad_norm": 0.4217219650745392,
940
+ "learning_rate": 1.7277059637985925e-06,
941
+ "loss": 1.3356,
942
+ "step": 1310
943
+ },
944
+ {
945
+ "epoch": 0.9310527243872333,
946
+ "grad_norm": 0.40834924578666687,
947
+ "learning_rate": 1.4213284898258073e-06,
948
+ "loss": 1.3514,
949
+ "step": 1320
950
+ },
951
+ {
952
+ "epoch": 0.9381061541174396,
953
+ "grad_norm": 0.43623679876327515,
954
+ "learning_rate": 1.1444429351620933e-06,
955
+ "loss": 1.3239,
956
+ "step": 1330
957
+ },
958
+ {
959
+ "epoch": 0.9451595838476459,
960
+ "grad_norm": 0.3971399664878845,
961
+ "learning_rate": 8.972173959324093e-07,
962
+ "loss": 1.3358,
963
+ "step": 1340
964
+ },
965
+ {
966
+ "epoch": 0.9522130135778523,
967
+ "grad_norm": 0.42954108119010925,
968
+ "learning_rate": 6.798019617836437e-07,
969
+ "loss": 1.336,
970
+ "step": 1350
971
+ },
972
+ {
973
+ "epoch": 0.9592664433080585,
974
+ "grad_norm": 0.43156835436820984,
975
+ "learning_rate": 4.923286247658476e-07,
976
+ "loss": 1.3414,
977
+ "step": 1360
978
+ },
979
+ {
980
+ "epoch": 0.9663198730382648,
981
+ "grad_norm": 0.43554648756980896,
982
+ "learning_rate": 3.349111992003218e-07,
983
+ "loss": 1.3395,
984
+ "step": 1370
985
+ },
986
+ {
987
+ "epoch": 0.9733733027684711,
988
+ "grad_norm": 0.44990482926368713,
989
+ "learning_rate": 2.0764525258341606e-07,
990
+ "loss": 1.3522,
991
+ "step": 1380
992
+ },
993
+ {
994
+ "epoch": 0.9804267324986775,
995
+ "grad_norm": 0.46251362562179565,
996
+ "learning_rate": 1.1060804756790055e-07,
997
+ "loss": 1.3422,
998
+ "step": 1390
999
+ },
1000
+ {
1001
+ "epoch": 0.9874801622288838,
1002
+ "grad_norm": 0.4148382842540741,
1003
+ "learning_rate": 4.385849505708084e-08,
1004
+ "loss": 1.336,
1005
+ "step": 1400
1006
+ },
1007
+ {
1008
+ "epoch": 0.9945335919590901,
1009
+ "grad_norm": 0.39327695965766907,
1010
+ "learning_rate": 7.437118440245528e-09,
1011
+ "loss": 1.367,
1012
+ "step": 1410
1013
+ },
1014
+ {
1015
+ "epoch": 0.9994709927702345,
1016
+ "step": 1417,
1017
+ "total_flos": 3.7897423231228314e+17,
1018
+ "train_loss": 1.4671159325642071,
1019
+ "train_runtime": 64163.8855,
1020
+ "train_samples_per_second": 8.484,
1021
+ "train_steps_per_second": 0.022
1022
+ }
1023
+ ],
1024
+ "logging_steps": 10,
1025
+ "max_steps": 1417,
1026
+ "num_input_tokens_seen": 0,
1027
+ "num_train_epochs": 1,
1028
+ "save_steps": 4000,
1029
+ "stateful_callbacks": {
1030
+ "TrainerControl": {
1031
+ "args": {
1032
+ "should_epoch_stop": false,
1033
+ "should_evaluate": false,
1034
+ "should_log": false,
1035
+ "should_save": false,
1036
+ "should_training_stop": false
1037
+ },
1038
+ "attributes": {}
1039
+ }
1040
+ },
1041
+ "total_flos": 3.7897423231228314e+17,
1042
+ "train_batch_size": 12,
1043
+ "trial_name": null,
1044
+ "trial_params": null
1045
+ }
training_eval_loss.png ADDED
training_loss.png ADDED