GaetanMichelet commited on
Commit
bb3fe5c
1 Parent(s): d484c80

Model save

Browse files
README.md ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
3
+ library_name: peft
4
+ license: llama3.1
5
+ tags:
6
+ - trl
7
+ - sft
8
+ - generated_from_trainer
9
+ model-index:
10
+ - name: Llama-31-8B_task-1_60-samples_config-2_full
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # Llama-31-8B_task-1_60-samples_config-2_full
18
+
19
+ This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on an unknown dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.9806
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0001
41
+ - train_batch_size: 1
42
+ - eval_batch_size: 1
43
+ - seed: 42
44
+ - distributed_type: multi-GPU
45
+ - gradient_accumulation_steps: 16
46
+ - total_train_batch_size: 16
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.1
50
+ - num_epochs: 50
51
+
52
+ ### Training results
53
+
54
+ | Training Loss | Epoch | Step | Validation Loss |
55
+ |:-------------:|:-------:|:----:|:---------------:|
56
+ | 2.5391 | 0.6957 | 2 | 2.3916 |
57
+ | 2.4703 | 1.7391 | 5 | 2.2074 |
58
+ | 2.1328 | 2.7826 | 8 | 1.9537 |
59
+ | 1.9709 | 3.8261 | 11 | 1.6769 |
60
+ | 1.5704 | 4.8696 | 14 | 1.3962 |
61
+ | 1.3765 | 5.9130 | 17 | 1.1358 |
62
+ | 1.0594 | 6.9565 | 20 | 1.0275 |
63
+ | 0.9969 | 8.0 | 23 | 0.9877 |
64
+ | 0.9485 | 8.6957 | 25 | 0.9700 |
65
+ | 0.8932 | 9.7391 | 28 | 0.9503 |
66
+ | 0.8815 | 10.7826 | 31 | 0.9331 |
67
+ | 0.8229 | 11.8261 | 34 | 0.9216 |
68
+ | 0.8136 | 12.8696 | 37 | 0.9111 |
69
+ | 0.7507 | 13.9130 | 40 | 0.9021 |
70
+ | 0.7373 | 14.9565 | 43 | 0.8982 |
71
+ | 0.6959 | 16.0 | 46 | 0.9020 |
72
+ | 0.6651 | 16.6957 | 48 | 0.9060 |
73
+ | 0.6589 | 17.7391 | 51 | 0.9119 |
74
+ | 0.5782 | 18.7826 | 54 | 0.9264 |
75
+ | 0.585 | 19.8261 | 57 | 0.9372 |
76
+ | 0.511 | 20.8696 | 60 | 0.9604 |
77
+ | 0.4767 | 21.9130 | 63 | 0.9806 |
78
+
79
+
80
+ ### Framework versions
81
+
82
+ - PEFT 0.12.0
83
+ - Transformers 4.44.0
84
+ - Pytorch 2.1.2+cu121
85
+ - Datasets 2.20.0
86
+ - Tokenizers 0.19.1
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b9a74cee8a91c70d5c243dda11701640f8caafd5937c9f26ee82911e945d30d1
3
  size 167832240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8141e12f0523a36ea5f1bbe01e7d822b703ce3b2e9ea68ab11d61a3a32d5a560
3
  size 167832240
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 21.91304347826087,
3
+ "total_flos": 1.5148012584042496e+16,
4
+ "train_loss": 1.081045771402026,
5
+ "train_runtime": 1186.9402,
6
+ "train_samples": 46,
7
+ "train_samples_per_second": 1.938,
8
+ "train_steps_per_second": 0.084
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 21.91304347826087,
3
+ "total_flos": 1.5148012584042496e+16,
4
+ "train_loss": 1.081045771402026,
5
+ "train_runtime": 1186.9402,
6
+ "train_samples": 46,
7
+ "train_samples_per_second": 1.938,
8
+ "train_steps_per_second": 0.084
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,451 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8982431888580322,
3
+ "best_model_checkpoint": "data/Llama-31-8B_task-1_60-samples_config-2_full/checkpoint-43",
4
+ "epoch": 21.91304347826087,
5
+ "eval_steps": 500,
6
+ "global_step": 63,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.34782608695652173,
13
+ "grad_norm": 2.058223247528076,
14
+ "learning_rate": 1e-05,
15
+ "loss": 2.5057,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.6956521739130435,
20
+ "grad_norm": 2.0094869136810303,
21
+ "learning_rate": 2e-05,
22
+ "loss": 2.5391,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.6956521739130435,
27
+ "eval_loss": 2.391584873199463,
28
+ "eval_runtime": 4.8261,
29
+ "eval_samples_per_second": 2.486,
30
+ "eval_steps_per_second": 2.486,
31
+ "step": 2
32
+ },
33
+ {
34
+ "epoch": 1.391304347826087,
35
+ "grad_norm": 1.6555781364440918,
36
+ "learning_rate": 4e-05,
37
+ "loss": 2.4703,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 1.7391304347826086,
42
+ "eval_loss": 2.20740008354187,
43
+ "eval_runtime": 4.8274,
44
+ "eval_samples_per_second": 2.486,
45
+ "eval_steps_per_second": 2.486,
46
+ "step": 5
47
+ },
48
+ {
49
+ "epoch": 2.0869565217391304,
50
+ "grad_norm": 1.380189299583435,
51
+ "learning_rate": 6e-05,
52
+ "loss": 2.3343,
53
+ "step": 6
54
+ },
55
+ {
56
+ "epoch": 2.782608695652174,
57
+ "grad_norm": 1.1684484481811523,
58
+ "learning_rate": 8e-05,
59
+ "loss": 2.1328,
60
+ "step": 8
61
+ },
62
+ {
63
+ "epoch": 2.782608695652174,
64
+ "eval_loss": 1.9536972045898438,
65
+ "eval_runtime": 4.8279,
66
+ "eval_samples_per_second": 2.486,
67
+ "eval_steps_per_second": 2.486,
68
+ "step": 8
69
+ },
70
+ {
71
+ "epoch": 3.4782608695652173,
72
+ "grad_norm": 1.248038649559021,
73
+ "learning_rate": 0.0001,
74
+ "loss": 1.9709,
75
+ "step": 10
76
+ },
77
+ {
78
+ "epoch": 3.8260869565217392,
79
+ "eval_loss": 1.6768689155578613,
80
+ "eval_runtime": 4.8282,
81
+ "eval_samples_per_second": 2.485,
82
+ "eval_steps_per_second": 2.485,
83
+ "step": 11
84
+ },
85
+ {
86
+ "epoch": 4.173913043478261,
87
+ "grad_norm": 0.9004043936729431,
88
+ "learning_rate": 9.987820251299122e-05,
89
+ "loss": 1.7679,
90
+ "step": 12
91
+ },
92
+ {
93
+ "epoch": 4.869565217391305,
94
+ "grad_norm": 0.9594781994819641,
95
+ "learning_rate": 9.951340343707852e-05,
96
+ "loss": 1.5704,
97
+ "step": 14
98
+ },
99
+ {
100
+ "epoch": 4.869565217391305,
101
+ "eval_loss": 1.3962249755859375,
102
+ "eval_runtime": 4.8269,
103
+ "eval_samples_per_second": 2.486,
104
+ "eval_steps_per_second": 2.486,
105
+ "step": 14
106
+ },
107
+ {
108
+ "epoch": 5.565217391304348,
109
+ "grad_norm": 1.0566002130508423,
110
+ "learning_rate": 9.890738003669029e-05,
111
+ "loss": 1.3765,
112
+ "step": 16
113
+ },
114
+ {
115
+ "epoch": 5.913043478260869,
116
+ "eval_loss": 1.1357998847961426,
117
+ "eval_runtime": 4.8302,
118
+ "eval_samples_per_second": 2.484,
119
+ "eval_steps_per_second": 2.484,
120
+ "step": 17
121
+ },
122
+ {
123
+ "epoch": 6.260869565217392,
124
+ "grad_norm": 0.7590504288673401,
125
+ "learning_rate": 9.806308479691595e-05,
126
+ "loss": 1.131,
127
+ "step": 18
128
+ },
129
+ {
130
+ "epoch": 6.956521739130435,
131
+ "grad_norm": 0.5426605343818665,
132
+ "learning_rate": 9.698463103929542e-05,
133
+ "loss": 1.0594,
134
+ "step": 20
135
+ },
136
+ {
137
+ "epoch": 6.956521739130435,
138
+ "eval_loss": 1.0274940729141235,
139
+ "eval_runtime": 4.825,
140
+ "eval_samples_per_second": 2.487,
141
+ "eval_steps_per_second": 2.487,
142
+ "step": 20
143
+ },
144
+ {
145
+ "epoch": 7.6521739130434785,
146
+ "grad_norm": 0.3678481876850128,
147
+ "learning_rate": 9.567727288213005e-05,
148
+ "loss": 0.9969,
149
+ "step": 22
150
+ },
151
+ {
152
+ "epoch": 8.0,
153
+ "eval_loss": 0.9876609444618225,
154
+ "eval_runtime": 4.821,
155
+ "eval_samples_per_second": 2.489,
156
+ "eval_steps_per_second": 2.489,
157
+ "step": 23
158
+ },
159
+ {
160
+ "epoch": 8.347826086956522,
161
+ "grad_norm": 0.33574286103248596,
162
+ "learning_rate": 9.414737964294636e-05,
163
+ "loss": 0.9485,
164
+ "step": 24
165
+ },
166
+ {
167
+ "epoch": 8.695652173913043,
168
+ "eval_loss": 0.9700170159339905,
169
+ "eval_runtime": 4.8279,
170
+ "eval_samples_per_second": 2.486,
171
+ "eval_steps_per_second": 2.486,
172
+ "step": 25
173
+ },
174
+ {
175
+ "epoch": 9.043478260869565,
176
+ "grad_norm": 0.3618321716785431,
177
+ "learning_rate": 9.24024048078213e-05,
178
+ "loss": 0.9357,
179
+ "step": 26
180
+ },
181
+ {
182
+ "epoch": 9.73913043478261,
183
+ "grad_norm": 0.3733964264392853,
184
+ "learning_rate": 9.045084971874738e-05,
185
+ "loss": 0.8932,
186
+ "step": 28
187
+ },
188
+ {
189
+ "epoch": 9.73913043478261,
190
+ "eval_loss": 0.9503232836723328,
191
+ "eval_runtime": 4.8207,
192
+ "eval_samples_per_second": 2.489,
193
+ "eval_steps_per_second": 2.489,
194
+ "step": 28
195
+ },
196
+ {
197
+ "epoch": 10.434782608695652,
198
+ "grad_norm": 0.34359118342399597,
199
+ "learning_rate": 8.83022221559489e-05,
200
+ "loss": 0.8815,
201
+ "step": 30
202
+ },
203
+ {
204
+ "epoch": 10.782608695652174,
205
+ "eval_loss": 0.9331014752388,
206
+ "eval_runtime": 4.8274,
207
+ "eval_samples_per_second": 2.486,
208
+ "eval_steps_per_second": 2.486,
209
+ "step": 31
210
+ },
211
+ {
212
+ "epoch": 11.130434782608695,
213
+ "grad_norm": 0.3720068037509918,
214
+ "learning_rate": 8.596699001693255e-05,
215
+ "loss": 0.8892,
216
+ "step": 32
217
+ },
218
+ {
219
+ "epoch": 11.826086956521738,
220
+ "grad_norm": 0.3460403382778168,
221
+ "learning_rate": 8.345653031794292e-05,
222
+ "loss": 0.8229,
223
+ "step": 34
224
+ },
225
+ {
226
+ "epoch": 11.826086956521738,
227
+ "eval_loss": 0.9215609431266785,
228
+ "eval_runtime": 4.8299,
229
+ "eval_samples_per_second": 2.485,
230
+ "eval_steps_per_second": 2.485,
231
+ "step": 34
232
+ },
233
+ {
234
+ "epoch": 12.521739130434783,
235
+ "grad_norm": 0.37291157245635986,
236
+ "learning_rate": 8.07830737662829e-05,
237
+ "loss": 0.8136,
238
+ "step": 36
239
+ },
240
+ {
241
+ "epoch": 12.869565217391305,
242
+ "eval_loss": 0.9111137390136719,
243
+ "eval_runtime": 4.8323,
244
+ "eval_samples_per_second": 2.483,
245
+ "eval_steps_per_second": 2.483,
246
+ "step": 37
247
+ },
248
+ {
249
+ "epoch": 13.217391304347826,
250
+ "grad_norm": 0.44780686497688293,
251
+ "learning_rate": 7.795964517353735e-05,
252
+ "loss": 0.8011,
253
+ "step": 38
254
+ },
255
+ {
256
+ "epoch": 13.91304347826087,
257
+ "grad_norm": 0.4574354887008667,
258
+ "learning_rate": 7.500000000000001e-05,
259
+ "loss": 0.7507,
260
+ "step": 40
261
+ },
262
+ {
263
+ "epoch": 13.91304347826087,
264
+ "eval_loss": 0.9021480679512024,
265
+ "eval_runtime": 4.8292,
266
+ "eval_samples_per_second": 2.485,
267
+ "eval_steps_per_second": 2.485,
268
+ "step": 40
269
+ },
270
+ {
271
+ "epoch": 14.608695652173914,
272
+ "grad_norm": 0.4815647602081299,
273
+ "learning_rate": 7.191855733945387e-05,
274
+ "loss": 0.7373,
275
+ "step": 42
276
+ },
277
+ {
278
+ "epoch": 14.956521739130435,
279
+ "eval_loss": 0.8982431888580322,
280
+ "eval_runtime": 4.8279,
281
+ "eval_samples_per_second": 2.486,
282
+ "eval_steps_per_second": 2.486,
283
+ "step": 43
284
+ },
285
+ {
286
+ "epoch": 15.304347826086957,
287
+ "grad_norm": 0.4997287690639496,
288
+ "learning_rate": 6.873032967079561e-05,
289
+ "loss": 0.7148,
290
+ "step": 44
291
+ },
292
+ {
293
+ "epoch": 16.0,
294
+ "grad_norm": 0.5747138857841492,
295
+ "learning_rate": 6.545084971874738e-05,
296
+ "loss": 0.6959,
297
+ "step": 46
298
+ },
299
+ {
300
+ "epoch": 16.0,
301
+ "eval_loss": 0.9020419716835022,
302
+ "eval_runtime": 4.8185,
303
+ "eval_samples_per_second": 2.49,
304
+ "eval_steps_per_second": 2.49,
305
+ "step": 46
306
+ },
307
+ {
308
+ "epoch": 16.695652173913043,
309
+ "grad_norm": 0.672699511051178,
310
+ "learning_rate": 6.209609477998338e-05,
311
+ "loss": 0.6651,
312
+ "step": 48
313
+ },
314
+ {
315
+ "epoch": 16.695652173913043,
316
+ "eval_loss": 0.9060425758361816,
317
+ "eval_runtime": 4.821,
318
+ "eval_samples_per_second": 2.489,
319
+ "eval_steps_per_second": 2.489,
320
+ "step": 48
321
+ },
322
+ {
323
+ "epoch": 17.391304347826086,
324
+ "grad_norm": 0.6403476595878601,
325
+ "learning_rate": 5.868240888334653e-05,
326
+ "loss": 0.6589,
327
+ "step": 50
328
+ },
329
+ {
330
+ "epoch": 17.73913043478261,
331
+ "eval_loss": 0.9118922352790833,
332
+ "eval_runtime": 4.8216,
333
+ "eval_samples_per_second": 2.489,
334
+ "eval_steps_per_second": 2.489,
335
+ "step": 51
336
+ },
337
+ {
338
+ "epoch": 18.08695652173913,
339
+ "grad_norm": 0.7273726463317871,
340
+ "learning_rate": 5.522642316338268e-05,
341
+ "loss": 0.5916,
342
+ "step": 52
343
+ },
344
+ {
345
+ "epoch": 18.782608695652176,
346
+ "grad_norm": 0.7468767166137695,
347
+ "learning_rate": 5.174497483512506e-05,
348
+ "loss": 0.5782,
349
+ "step": 54
350
+ },
351
+ {
352
+ "epoch": 18.782608695652176,
353
+ "eval_loss": 0.9264171123504639,
354
+ "eval_runtime": 4.824,
355
+ "eval_samples_per_second": 2.488,
356
+ "eval_steps_per_second": 2.488,
357
+ "step": 54
358
+ },
359
+ {
360
+ "epoch": 19.47826086956522,
361
+ "grad_norm": 0.7150223851203918,
362
+ "learning_rate": 4.825502516487497e-05,
363
+ "loss": 0.585,
364
+ "step": 56
365
+ },
366
+ {
367
+ "epoch": 19.82608695652174,
368
+ "eval_loss": 0.9372181296348572,
369
+ "eval_runtime": 4.8257,
370
+ "eval_samples_per_second": 2.487,
371
+ "eval_steps_per_second": 2.487,
372
+ "step": 57
373
+ },
374
+ {
375
+ "epoch": 20.17391304347826,
376
+ "grad_norm": 0.7036177515983582,
377
+ "learning_rate": 4.477357683661734e-05,
378
+ "loss": 0.4927,
379
+ "step": 58
380
+ },
381
+ {
382
+ "epoch": 20.869565217391305,
383
+ "grad_norm": 0.6790146231651306,
384
+ "learning_rate": 4.131759111665349e-05,
385
+ "loss": 0.511,
386
+ "step": 60
387
+ },
388
+ {
389
+ "epoch": 20.869565217391305,
390
+ "eval_loss": 0.9603924751281738,
391
+ "eval_runtime": 4.8248,
392
+ "eval_samples_per_second": 2.487,
393
+ "eval_steps_per_second": 2.487,
394
+ "step": 60
395
+ },
396
+ {
397
+ "epoch": 21.565217391304348,
398
+ "grad_norm": 0.5890304446220398,
399
+ "learning_rate": 3.790390522001662e-05,
400
+ "loss": 0.4767,
401
+ "step": 62
402
+ },
403
+ {
404
+ "epoch": 21.91304347826087,
405
+ "eval_loss": 0.9806472659111023,
406
+ "eval_runtime": 4.8246,
407
+ "eval_samples_per_second": 2.487,
408
+ "eval_steps_per_second": 2.487,
409
+ "step": 63
410
+ },
411
+ {
412
+ "epoch": 21.91304347826087,
413
+ "step": 63,
414
+ "total_flos": 1.5148012584042496e+16,
415
+ "train_loss": 1.081045771402026,
416
+ "train_runtime": 1186.9402,
417
+ "train_samples_per_second": 1.938,
418
+ "train_steps_per_second": 0.084
419
+ }
420
+ ],
421
+ "logging_steps": 2,
422
+ "max_steps": 100,
423
+ "num_input_tokens_seen": 0,
424
+ "num_train_epochs": 50,
425
+ "save_steps": 25,
426
+ "stateful_callbacks": {
427
+ "EarlyStoppingCallback": {
428
+ "args": {
429
+ "early_stopping_patience": 7,
430
+ "early_stopping_threshold": 0.0
431
+ },
432
+ "attributes": {
433
+ "early_stopping_patience_counter": 0
434
+ }
435
+ },
436
+ "TrainerControl": {
437
+ "args": {
438
+ "should_epoch_stop": false,
439
+ "should_evaluate": false,
440
+ "should_log": false,
441
+ "should_save": true,
442
+ "should_training_stop": true
443
+ },
444
+ "attributes": {}
445
+ }
446
+ },
447
+ "total_flos": 1.5148012584042496e+16,
448
+ "train_batch_size": 1,
449
+ "trial_name": null,
450
+ "trial_params": null
451
+ }