GaetanMichelet commited on
Commit
f0dca5b
1 Parent(s): d60f7b2

Model save

Browse files
README.md ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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-1
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-1
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: 2.7460
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: 8
46
+ - total_train_batch_size: 8
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
+ | 2.1529 | 0.8696 | 5 | 2.0041 |
57
+ | 1.7276 | 1.9130 | 11 | 1.6163 |
58
+ | 1.5018 | 2.9565 | 17 | 1.4631 |
59
+ | 1.3727 | 4.0 | 23 | 1.3338 |
60
+ | 1.0912 | 4.8696 | 28 | 1.2853 |
61
+ | 0.8709 | 5.9130 | 34 | 1.2905 |
62
+ | 0.713 | 6.9565 | 40 | 1.4713 |
63
+ | 0.4007 | 8.0 | 46 | 1.6719 |
64
+ | 0.2182 | 8.8696 | 51 | 2.0390 |
65
+ | 0.0802 | 9.9130 | 57 | 2.5003 |
66
+ | 0.0304 | 10.9565 | 63 | 2.7328 |
67
+ | 0.0155 | 12.0 | 69 | 2.7460 |
68
+
69
+
70
+ ### Framework versions
71
+
72
+ - PEFT 0.12.0
73
+ - Transformers 4.44.0
74
+ - Pytorch 2.1.2+cu121
75
+ - Datasets 2.20.0
76
+ - Tokenizers 0.19.1
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:14fadf819f0d79152385e52c8f6dff9a9c382f8a298145a658e5ca8860a54fd6
3
  size 167832240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47e20718651ee7c17c9653c22c3fe88b77bc9e845a90cb50a94b64296844e60c
3
  size 167832240
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.0,
3
+ "total_flos": 8262552305795072.0,
4
+ "train_loss": 0.8597148116678,
5
+ "train_runtime": 650.8523,
6
+ "train_samples": 46,
7
+ "train_samples_per_second": 3.534,
8
+ "train_steps_per_second": 0.384
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.0,
3
+ "total_flos": 8262552305795072.0,
4
+ "train_loss": 0.8597148116678,
5
+ "train_runtime": 650.8523,
6
+ "train_samples": 46,
7
+ "train_samples_per_second": 3.534,
8
+ "train_steps_per_second": 0.384
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,392 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.2853014469146729,
3
+ "best_model_checkpoint": "data/Llama-31-8B_task-1_60-samples_config-1/checkpoint-28",
4
+ "epoch": 12.0,
5
+ "eval_steps": 500,
6
+ "global_step": 69,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.17391304347826086,
13
+ "grad_norm": 1.9813579320907593,
14
+ "learning_rate": 4.000000000000001e-06,
15
+ "loss": 2.073,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.34782608695652173,
20
+ "grad_norm": 1.8313475847244263,
21
+ "learning_rate": 8.000000000000001e-06,
22
+ "loss": 2.0852,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.6956521739130435,
27
+ "grad_norm": 1.7238192558288574,
28
+ "learning_rate": 1.6000000000000003e-05,
29
+ "loss": 2.1529,
30
+ "step": 4
31
+ },
32
+ {
33
+ "epoch": 0.8695652173913043,
34
+ "eval_loss": 2.0041069984436035,
35
+ "eval_runtime": 4.8643,
36
+ "eval_samples_per_second": 2.467,
37
+ "eval_steps_per_second": 2.467,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 1.0434782608695652,
42
+ "grad_norm": 1.8038513660430908,
43
+ "learning_rate": 2.4e-05,
44
+ "loss": 2.0483,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 1.391304347826087,
49
+ "grad_norm": 1.8350396156311035,
50
+ "learning_rate": 3.2000000000000005e-05,
51
+ "loss": 2.0287,
52
+ "step": 8
53
+ },
54
+ {
55
+ "epoch": 1.7391304347826086,
56
+ "grad_norm": 1.627516746520996,
57
+ "learning_rate": 4e-05,
58
+ "loss": 1.7276,
59
+ "step": 10
60
+ },
61
+ {
62
+ "epoch": 1.9130434782608696,
63
+ "eval_loss": 1.6163220405578613,
64
+ "eval_runtime": 4.8521,
65
+ "eval_samples_per_second": 2.473,
66
+ "eval_steps_per_second": 2.473,
67
+ "step": 11
68
+ },
69
+ {
70
+ "epoch": 2.0869565217391304,
71
+ "grad_norm": 0.9827105402946472,
72
+ "learning_rate": 4.8e-05,
73
+ "loss": 1.5965,
74
+ "step": 12
75
+ },
76
+ {
77
+ "epoch": 2.4347826086956523,
78
+ "grad_norm": 0.7503597140312195,
79
+ "learning_rate": 5.6000000000000006e-05,
80
+ "loss": 1.5403,
81
+ "step": 14
82
+ },
83
+ {
84
+ "epoch": 2.782608695652174,
85
+ "grad_norm": 1.0193320512771606,
86
+ "learning_rate": 6.400000000000001e-05,
87
+ "loss": 1.5018,
88
+ "step": 16
89
+ },
90
+ {
91
+ "epoch": 2.9565217391304346,
92
+ "eval_loss": 1.463134765625,
93
+ "eval_runtime": 4.9169,
94
+ "eval_samples_per_second": 2.441,
95
+ "eval_steps_per_second": 2.441,
96
+ "step": 17
97
+ },
98
+ {
99
+ "epoch": 3.130434782608696,
100
+ "grad_norm": 0.7495579719543457,
101
+ "learning_rate": 7.2e-05,
102
+ "loss": 1.4009,
103
+ "step": 18
104
+ },
105
+ {
106
+ "epoch": 3.4782608695652173,
107
+ "grad_norm": 0.8146828413009644,
108
+ "learning_rate": 8e-05,
109
+ "loss": 1.2842,
110
+ "step": 20
111
+ },
112
+ {
113
+ "epoch": 3.8260869565217392,
114
+ "grad_norm": 0.8759430050849915,
115
+ "learning_rate": 8.800000000000001e-05,
116
+ "loss": 1.3727,
117
+ "step": 22
118
+ },
119
+ {
120
+ "epoch": 4.0,
121
+ "eval_loss": 1.3337726593017578,
122
+ "eval_runtime": 4.9016,
123
+ "eval_samples_per_second": 2.448,
124
+ "eval_steps_per_second": 2.448,
125
+ "step": 23
126
+ },
127
+ {
128
+ "epoch": 4.173913043478261,
129
+ "grad_norm": 0.7865839004516602,
130
+ "learning_rate": 9.6e-05,
131
+ "loss": 1.2452,
132
+ "step": 24
133
+ },
134
+ {
135
+ "epoch": 4.521739130434782,
136
+ "grad_norm": 0.7563604116439819,
137
+ "learning_rate": 9.999512620046522e-05,
138
+ "loss": 1.1817,
139
+ "step": 26
140
+ },
141
+ {
142
+ "epoch": 4.869565217391305,
143
+ "grad_norm": 0.791088342666626,
144
+ "learning_rate": 9.995614150494293e-05,
145
+ "loss": 1.0912,
146
+ "step": 28
147
+ },
148
+ {
149
+ "epoch": 4.869565217391305,
150
+ "eval_loss": 1.2853014469146729,
151
+ "eval_runtime": 4.8827,
152
+ "eval_samples_per_second": 2.458,
153
+ "eval_steps_per_second": 2.458,
154
+ "step": 28
155
+ },
156
+ {
157
+ "epoch": 5.217391304347826,
158
+ "grad_norm": 0.8185644745826721,
159
+ "learning_rate": 9.987820251299122e-05,
160
+ "loss": 1.0722,
161
+ "step": 30
162
+ },
163
+ {
164
+ "epoch": 5.565217391304348,
165
+ "grad_norm": 0.9191321134567261,
166
+ "learning_rate": 9.976136999909156e-05,
167
+ "loss": 0.9712,
168
+ "step": 32
169
+ },
170
+ {
171
+ "epoch": 5.913043478260869,
172
+ "grad_norm": 1.0726240873336792,
173
+ "learning_rate": 9.96057350657239e-05,
174
+ "loss": 0.8709,
175
+ "step": 34
176
+ },
177
+ {
178
+ "epoch": 5.913043478260869,
179
+ "eval_loss": 1.2904514074325562,
180
+ "eval_runtime": 4.8494,
181
+ "eval_samples_per_second": 2.475,
182
+ "eval_steps_per_second": 2.475,
183
+ "step": 34
184
+ },
185
+ {
186
+ "epoch": 6.260869565217392,
187
+ "grad_norm": 0.9275307059288025,
188
+ "learning_rate": 9.941141907232765e-05,
189
+ "loss": 0.6996,
190
+ "step": 36
191
+ },
192
+ {
193
+ "epoch": 6.608695652173913,
194
+ "grad_norm": 1.382583498954773,
195
+ "learning_rate": 9.917857354066931e-05,
196
+ "loss": 0.7177,
197
+ "step": 38
198
+ },
199
+ {
200
+ "epoch": 6.956521739130435,
201
+ "grad_norm": 1.5849192142486572,
202
+ "learning_rate": 9.890738003669029e-05,
203
+ "loss": 0.713,
204
+ "step": 40
205
+ },
206
+ {
207
+ "epoch": 6.956521739130435,
208
+ "eval_loss": 1.4712576866149902,
209
+ "eval_runtime": 4.8441,
210
+ "eval_samples_per_second": 2.477,
211
+ "eval_steps_per_second": 2.477,
212
+ "step": 40
213
+ },
214
+ {
215
+ "epoch": 7.304347826086957,
216
+ "grad_norm": 1.2915489673614502,
217
+ "learning_rate": 9.859805002892732e-05,
218
+ "loss": 0.5017,
219
+ "step": 42
220
+ },
221
+ {
222
+ "epoch": 7.6521739130434785,
223
+ "grad_norm": 1.3434025049209595,
224
+ "learning_rate": 9.825082472361557e-05,
225
+ "loss": 0.4617,
226
+ "step": 44
227
+ },
228
+ {
229
+ "epoch": 8.0,
230
+ "grad_norm": 1.7452040910720825,
231
+ "learning_rate": 9.786597487660337e-05,
232
+ "loss": 0.4007,
233
+ "step": 46
234
+ },
235
+ {
236
+ "epoch": 8.0,
237
+ "eval_loss": 1.671922206878662,
238
+ "eval_runtime": 4.8483,
239
+ "eval_samples_per_second": 2.475,
240
+ "eval_steps_per_second": 2.475,
241
+ "step": 46
242
+ },
243
+ {
244
+ "epoch": 8.347826086956522,
245
+ "grad_norm": 1.627158761024475,
246
+ "learning_rate": 9.744380058222483e-05,
247
+ "loss": 0.2485,
248
+ "step": 48
249
+ },
250
+ {
251
+ "epoch": 8.695652173913043,
252
+ "grad_norm": 1.4861387014389038,
253
+ "learning_rate": 9.698463103929542e-05,
254
+ "loss": 0.2182,
255
+ "step": 50
256
+ },
257
+ {
258
+ "epoch": 8.869565217391305,
259
+ "eval_loss": 2.0389816761016846,
260
+ "eval_runtime": 4.8412,
261
+ "eval_samples_per_second": 2.479,
262
+ "eval_steps_per_second": 2.479,
263
+ "step": 51
264
+ },
265
+ {
266
+ "epoch": 9.043478260869565,
267
+ "grad_norm": 1.8841030597686768,
268
+ "learning_rate": 9.648882429441257e-05,
269
+ "loss": 0.1563,
270
+ "step": 52
271
+ },
272
+ {
273
+ "epoch": 9.391304347826088,
274
+ "grad_norm": 2.056680917739868,
275
+ "learning_rate": 9.595676696276172e-05,
276
+ "loss": 0.0866,
277
+ "step": 54
278
+ },
279
+ {
280
+ "epoch": 9.73913043478261,
281
+ "grad_norm": 2.4738428592681885,
282
+ "learning_rate": 9.538887392664544e-05,
283
+ "loss": 0.0802,
284
+ "step": 56
285
+ },
286
+ {
287
+ "epoch": 9.91304347826087,
288
+ "eval_loss": 2.5002787113189697,
289
+ "eval_runtime": 4.8542,
290
+ "eval_samples_per_second": 2.472,
291
+ "eval_steps_per_second": 2.472,
292
+ "step": 57
293
+ },
294
+ {
295
+ "epoch": 10.08695652173913,
296
+ "grad_norm": 1.4879595041275024,
297
+ "learning_rate": 9.478558801197065e-05,
298
+ "loss": 0.0638,
299
+ "step": 58
300
+ },
301
+ {
302
+ "epoch": 10.434782608695652,
303
+ "grad_norm": 2.2851953506469727,
304
+ "learning_rate": 9.414737964294636e-05,
305
+ "loss": 0.0328,
306
+ "step": 60
307
+ },
308
+ {
309
+ "epoch": 10.782608695652174,
310
+ "grad_norm": 1.1112951040267944,
311
+ "learning_rate": 9.347474647526095e-05,
312
+ "loss": 0.0304,
313
+ "step": 62
314
+ },
315
+ {
316
+ "epoch": 10.956521739130435,
317
+ "eval_loss": 2.7327919006347656,
318
+ "eval_runtime": 4.8404,
319
+ "eval_samples_per_second": 2.479,
320
+ "eval_steps_per_second": 2.479,
321
+ "step": 63
322
+ },
323
+ {
324
+ "epoch": 11.130434782608695,
325
+ "grad_norm": 1.1123360395431519,
326
+ "learning_rate": 9.276821300802534e-05,
327
+ "loss": 0.029,
328
+ "step": 64
329
+ },
330
+ {
331
+ "epoch": 11.478260869565217,
332
+ "grad_norm": 1.831784963607788,
333
+ "learning_rate": 9.202833017478422e-05,
334
+ "loss": 0.0298,
335
+ "step": 66
336
+ },
337
+ {
338
+ "epoch": 11.826086956521738,
339
+ "grad_norm": 1.1904350519180298,
340
+ "learning_rate": 9.125567491391476e-05,
341
+ "loss": 0.0155,
342
+ "step": 68
343
+ },
344
+ {
345
+ "epoch": 12.0,
346
+ "eval_loss": 2.7460126876831055,
347
+ "eval_runtime": 4.839,
348
+ "eval_samples_per_second": 2.48,
349
+ "eval_steps_per_second": 2.48,
350
+ "step": 69
351
+ },
352
+ {
353
+ "epoch": 12.0,
354
+ "step": 69,
355
+ "total_flos": 8262552305795072.0,
356
+ "train_loss": 0.8597148116678,
357
+ "train_runtime": 650.8523,
358
+ "train_samples_per_second": 3.534,
359
+ "train_steps_per_second": 0.384
360
+ }
361
+ ],
362
+ "logging_steps": 2,
363
+ "max_steps": 250,
364
+ "num_input_tokens_seen": 0,
365
+ "num_train_epochs": 50,
366
+ "save_steps": 25,
367
+ "stateful_callbacks": {
368
+ "EarlyStoppingCallback": {
369
+ "args": {
370
+ "early_stopping_patience": 7,
371
+ "early_stopping_threshold": 0.0
372
+ },
373
+ "attributes": {
374
+ "early_stopping_patience_counter": 0
375
+ }
376
+ },
377
+ "TrainerControl": {
378
+ "args": {
379
+ "should_epoch_stop": false,
380
+ "should_evaluate": false,
381
+ "should_log": false,
382
+ "should_save": true,
383
+ "should_training_stop": true
384
+ },
385
+ "attributes": {}
386
+ }
387
+ },
388
+ "total_flos": 8262552305795072.0,
389
+ "train_batch_size": 1,
390
+ "trial_name": null,
391
+ "trial_params": null
392
+ }