GaetanMichelet commited on
Commit
d01cc71
1 Parent(s): 5975d3e

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-3_180-samples_config-1_auto
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-3_180-samples_config-1_auto
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 the None dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.5623
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
+ | 1.7081 | 1.0 | 17 | 1.5542 |
57
+ | 0.509 | 2.0 | 34 | 0.4773 |
58
+ | 0.3144 | 3.0 | 51 | 0.3656 |
59
+ | 0.2899 | 4.0 | 68 | 0.3454 |
60
+ | 0.2162 | 5.0 | 85 | 0.3313 |
61
+ | 0.3465 | 6.0 | 102 | 0.3374 |
62
+ | 0.2498 | 7.0 | 119 | 0.3556 |
63
+ | 0.1023 | 8.0 | 136 | 0.4104 |
64
+ | 0.0862 | 9.0 | 153 | 0.4676 |
65
+ | 0.0549 | 10.0 | 170 | 0.4978 |
66
+ | 0.032 | 11.0 | 187 | 0.5856 |
67
+ | 0.0072 | 12.0 | 204 | 0.5623 |
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:d467a1107e31e5ca6d60e260952f9ec0d038b8b3258d9b8cf393b6d445638d0d
3
  size 167832240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e792adea19d48882a1b597cb94ba58b96d8b49cda71c167e110e5cc518df7bdd
3
  size 167832240
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.0,
3
+ "total_flos": 1.2893298226941133e+17,
4
+ "train_loss": 0.38738249295719845,
5
+ "train_runtime": 4719.6721,
6
+ "train_samples": 136,
7
+ "train_samples_per_second": 1.441,
8
+ "train_steps_per_second": 0.18
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.0,
3
+ "total_flos": 1.2893298226941133e+17,
4
+ "train_loss": 0.38738249295719845,
5
+ "train_runtime": 4719.6721,
6
+ "train_samples": 136,
7
+ "train_samples_per_second": 1.441,
8
+ "train_steps_per_second": 0.18
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,868 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.3313491940498352,
3
+ "best_model_checkpoint": "data/Llama-31-8B_task-3_180-samples_config-1_auto/checkpoint-85",
4
+ "epoch": 12.0,
5
+ "eval_steps": 500,
6
+ "global_step": 204,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.058823529411764705,
13
+ "grad_norm": 2.790146827697754,
14
+ "learning_rate": 1.1764705882352942e-06,
15
+ "loss": 2.4521,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.11764705882352941,
20
+ "grad_norm": 1.3707395792007446,
21
+ "learning_rate": 2.3529411764705885e-06,
22
+ "loss": 1.7077,
23
+ "step": 2
24
+ },
25
+ {
26
+ "epoch": 0.23529411764705882,
27
+ "grad_norm": 2.8815231323242188,
28
+ "learning_rate": 4.705882352941177e-06,
29
+ "loss": 2.6331,
30
+ "step": 4
31
+ },
32
+ {
33
+ "epoch": 0.35294117647058826,
34
+ "grad_norm": 1.3376513719558716,
35
+ "learning_rate": 7.058823529411765e-06,
36
+ "loss": 2.3661,
37
+ "step": 6
38
+ },
39
+ {
40
+ "epoch": 0.47058823529411764,
41
+ "grad_norm": 2.828890323638916,
42
+ "learning_rate": 9.411764705882354e-06,
43
+ "loss": 2.1508,
44
+ "step": 8
45
+ },
46
+ {
47
+ "epoch": 0.5882352941176471,
48
+ "grad_norm": 2.8286173343658447,
49
+ "learning_rate": 1.1764705882352942e-05,
50
+ "loss": 2.2319,
51
+ "step": 10
52
+ },
53
+ {
54
+ "epoch": 0.7058823529411765,
55
+ "grad_norm": 2.2810916900634766,
56
+ "learning_rate": 1.411764705882353e-05,
57
+ "loss": 1.808,
58
+ "step": 12
59
+ },
60
+ {
61
+ "epoch": 0.8235294117647058,
62
+ "grad_norm": 3.351749897003174,
63
+ "learning_rate": 1.647058823529412e-05,
64
+ "loss": 2.1354,
65
+ "step": 14
66
+ },
67
+ {
68
+ "epoch": 0.9411764705882353,
69
+ "grad_norm": 1.3132500648498535,
70
+ "learning_rate": 1.8823529411764708e-05,
71
+ "loss": 1.7081,
72
+ "step": 16
73
+ },
74
+ {
75
+ "epoch": 1.0,
76
+ "eval_loss": 1.5541911125183105,
77
+ "eval_runtime": 32.0115,
78
+ "eval_samples_per_second": 1.125,
79
+ "eval_steps_per_second": 1.125,
80
+ "step": 17
81
+ },
82
+ {
83
+ "epoch": 1.0588235294117647,
84
+ "grad_norm": 0.9471787810325623,
85
+ "learning_rate": 2.1176470588235296e-05,
86
+ "loss": 1.4844,
87
+ "step": 18
88
+ },
89
+ {
90
+ "epoch": 1.1764705882352942,
91
+ "grad_norm": 2.1010115146636963,
92
+ "learning_rate": 2.3529411764705884e-05,
93
+ "loss": 1.4457,
94
+ "step": 20
95
+ },
96
+ {
97
+ "epoch": 1.2941176470588236,
98
+ "grad_norm": 1.4327553510665894,
99
+ "learning_rate": 2.5882352941176475e-05,
100
+ "loss": 1.1948,
101
+ "step": 22
102
+ },
103
+ {
104
+ "epoch": 1.4117647058823528,
105
+ "grad_norm": 1.9643908739089966,
106
+ "learning_rate": 2.823529411764706e-05,
107
+ "loss": 0.9873,
108
+ "step": 24
109
+ },
110
+ {
111
+ "epoch": 1.5294117647058822,
112
+ "grad_norm": 1.6198688745498657,
113
+ "learning_rate": 3.058823529411765e-05,
114
+ "loss": 0.7604,
115
+ "step": 26
116
+ },
117
+ {
118
+ "epoch": 1.6470588235294117,
119
+ "grad_norm": 1.122072458267212,
120
+ "learning_rate": 3.294117647058824e-05,
121
+ "loss": 0.6013,
122
+ "step": 28
123
+ },
124
+ {
125
+ "epoch": 1.7647058823529411,
126
+ "grad_norm": 1.1838219165802002,
127
+ "learning_rate": 3.529411764705883e-05,
128
+ "loss": 0.7327,
129
+ "step": 30
130
+ },
131
+ {
132
+ "epoch": 1.8823529411764706,
133
+ "grad_norm": 1.0292763710021973,
134
+ "learning_rate": 3.7647058823529415e-05,
135
+ "loss": 0.4018,
136
+ "step": 32
137
+ },
138
+ {
139
+ "epoch": 2.0,
140
+ "grad_norm": 1.8941810131072998,
141
+ "learning_rate": 4e-05,
142
+ "loss": 0.509,
143
+ "step": 34
144
+ },
145
+ {
146
+ "epoch": 2.0,
147
+ "eval_loss": 0.47733280062675476,
148
+ "eval_runtime": 32.0078,
149
+ "eval_samples_per_second": 1.125,
150
+ "eval_steps_per_second": 1.125,
151
+ "step": 34
152
+ },
153
+ {
154
+ "epoch": 2.1176470588235294,
155
+ "grad_norm": 0.6447968482971191,
156
+ "learning_rate": 4.235294117647059e-05,
157
+ "loss": 0.4526,
158
+ "step": 36
159
+ },
160
+ {
161
+ "epoch": 2.235294117647059,
162
+ "grad_norm": 0.7567365169525146,
163
+ "learning_rate": 4.470588235294118e-05,
164
+ "loss": 0.4536,
165
+ "step": 38
166
+ },
167
+ {
168
+ "epoch": 2.3529411764705883,
169
+ "grad_norm": 0.8407445549964905,
170
+ "learning_rate": 4.705882352941177e-05,
171
+ "loss": 0.4453,
172
+ "step": 40
173
+ },
174
+ {
175
+ "epoch": 2.4705882352941178,
176
+ "grad_norm": 0.3435295522212982,
177
+ "learning_rate": 4.9411764705882355e-05,
178
+ "loss": 0.325,
179
+ "step": 42
180
+ },
181
+ {
182
+ "epoch": 2.588235294117647,
183
+ "grad_norm": 0.3069871664047241,
184
+ "learning_rate": 5.176470588235295e-05,
185
+ "loss": 0.3309,
186
+ "step": 44
187
+ },
188
+ {
189
+ "epoch": 2.7058823529411766,
190
+ "grad_norm": 0.37769752740859985,
191
+ "learning_rate": 5.411764705882353e-05,
192
+ "loss": 0.3642,
193
+ "step": 46
194
+ },
195
+ {
196
+ "epoch": 2.8235294117647056,
197
+ "grad_norm": 0.37647154927253723,
198
+ "learning_rate": 5.647058823529412e-05,
199
+ "loss": 0.3983,
200
+ "step": 48
201
+ },
202
+ {
203
+ "epoch": 2.9411764705882355,
204
+ "grad_norm": 0.4306583106517792,
205
+ "learning_rate": 5.882352941176471e-05,
206
+ "loss": 0.3144,
207
+ "step": 50
208
+ },
209
+ {
210
+ "epoch": 3.0,
211
+ "eval_loss": 0.3655799627304077,
212
+ "eval_runtime": 32.007,
213
+ "eval_samples_per_second": 1.125,
214
+ "eval_steps_per_second": 1.125,
215
+ "step": 51
216
+ },
217
+ {
218
+ "epoch": 3.0588235294117645,
219
+ "grad_norm": 0.3212827742099762,
220
+ "learning_rate": 6.11764705882353e-05,
221
+ "loss": 0.2755,
222
+ "step": 52
223
+ },
224
+ {
225
+ "epoch": 3.176470588235294,
226
+ "grad_norm": 1.307051658630371,
227
+ "learning_rate": 6.352941176470588e-05,
228
+ "loss": 0.448,
229
+ "step": 54
230
+ },
231
+ {
232
+ "epoch": 3.2941176470588234,
233
+ "grad_norm": 0.37125203013420105,
234
+ "learning_rate": 6.588235294117648e-05,
235
+ "loss": 0.312,
236
+ "step": 56
237
+ },
238
+ {
239
+ "epoch": 3.411764705882353,
240
+ "grad_norm": 0.23212695121765137,
241
+ "learning_rate": 6.823529411764707e-05,
242
+ "loss": 0.4181,
243
+ "step": 58
244
+ },
245
+ {
246
+ "epoch": 3.5294117647058822,
247
+ "grad_norm": 5.962123870849609,
248
+ "learning_rate": 7.058823529411765e-05,
249
+ "loss": 0.258,
250
+ "step": 60
251
+ },
252
+ {
253
+ "epoch": 3.6470588235294117,
254
+ "grad_norm": 0.6083526611328125,
255
+ "learning_rate": 7.294117647058823e-05,
256
+ "loss": 0.2953,
257
+ "step": 62
258
+ },
259
+ {
260
+ "epoch": 3.764705882352941,
261
+ "grad_norm": 0.3373955190181732,
262
+ "learning_rate": 7.529411764705883e-05,
263
+ "loss": 0.4026,
264
+ "step": 64
265
+ },
266
+ {
267
+ "epoch": 3.8823529411764706,
268
+ "grad_norm": 0.45089852809906006,
269
+ "learning_rate": 7.764705882352942e-05,
270
+ "loss": 0.2389,
271
+ "step": 66
272
+ },
273
+ {
274
+ "epoch": 4.0,
275
+ "grad_norm": 0.4510911703109741,
276
+ "learning_rate": 8e-05,
277
+ "loss": 0.2899,
278
+ "step": 68
279
+ },
280
+ {
281
+ "epoch": 4.0,
282
+ "eval_loss": 0.34543025493621826,
283
+ "eval_runtime": 32.0083,
284
+ "eval_samples_per_second": 1.125,
285
+ "eval_steps_per_second": 1.125,
286
+ "step": 68
287
+ },
288
+ {
289
+ "epoch": 4.117647058823529,
290
+ "grad_norm": 0.642993688583374,
291
+ "learning_rate": 8.23529411764706e-05,
292
+ "loss": 0.265,
293
+ "step": 70
294
+ },
295
+ {
296
+ "epoch": 4.235294117647059,
297
+ "grad_norm": 0.37745243310928345,
298
+ "learning_rate": 8.470588235294118e-05,
299
+ "loss": 0.2576,
300
+ "step": 72
301
+ },
302
+ {
303
+ "epoch": 4.352941176470588,
304
+ "grad_norm": 0.4048821032047272,
305
+ "learning_rate": 8.705882352941177e-05,
306
+ "loss": 0.3338,
307
+ "step": 74
308
+ },
309
+ {
310
+ "epoch": 4.470588235294118,
311
+ "grad_norm": 0.40954768657684326,
312
+ "learning_rate": 8.941176470588236e-05,
313
+ "loss": 0.3065,
314
+ "step": 76
315
+ },
316
+ {
317
+ "epoch": 4.588235294117647,
318
+ "grad_norm": 0.30904123187065125,
319
+ "learning_rate": 9.176470588235295e-05,
320
+ "loss": 0.3121,
321
+ "step": 78
322
+ },
323
+ {
324
+ "epoch": 4.705882352941177,
325
+ "grad_norm": 1.0630900859832764,
326
+ "learning_rate": 9.411764705882353e-05,
327
+ "loss": 0.2534,
328
+ "step": 80
329
+ },
330
+ {
331
+ "epoch": 4.823529411764706,
332
+ "grad_norm": 0.4571320712566376,
333
+ "learning_rate": 9.647058823529412e-05,
334
+ "loss": 0.3327,
335
+ "step": 82
336
+ },
337
+ {
338
+ "epoch": 4.9411764705882355,
339
+ "grad_norm": 0.42991459369659424,
340
+ "learning_rate": 9.882352941176471e-05,
341
+ "loss": 0.2162,
342
+ "step": 84
343
+ },
344
+ {
345
+ "epoch": 5.0,
346
+ "eval_loss": 0.3313491940498352,
347
+ "eval_runtime": 32.0012,
348
+ "eval_samples_per_second": 1.125,
349
+ "eval_steps_per_second": 1.125,
350
+ "step": 85
351
+ },
352
+ {
353
+ "epoch": 5.0588235294117645,
354
+ "grad_norm": 0.4001903831958771,
355
+ "learning_rate": 9.99995783847866e-05,
356
+ "loss": 0.2581,
357
+ "step": 86
358
+ },
359
+ {
360
+ "epoch": 5.176470588235294,
361
+ "grad_norm": 0.4010772705078125,
362
+ "learning_rate": 9.999620550574153e-05,
363
+ "loss": 0.2951,
364
+ "step": 88
365
+ },
366
+ {
367
+ "epoch": 5.294117647058823,
368
+ "grad_norm": 1.2183829545974731,
369
+ "learning_rate": 9.998945997517956e-05,
370
+ "loss": 0.2299,
371
+ "step": 90
372
+ },
373
+ {
374
+ "epoch": 5.411764705882353,
375
+ "grad_norm": 0.0022620977833867073,
376
+ "learning_rate": 9.997934224814173e-05,
377
+ "loss": 0.0719,
378
+ "step": 92
379
+ },
380
+ {
381
+ "epoch": 5.529411764705882,
382
+ "grad_norm": 0.49284520745277405,
383
+ "learning_rate": 9.996585300715116e-05,
384
+ "loss": 0.2196,
385
+ "step": 94
386
+ },
387
+ {
388
+ "epoch": 5.647058823529412,
389
+ "grad_norm": 0.47940441966056824,
390
+ "learning_rate": 9.994899316216708e-05,
391
+ "loss": 0.2111,
392
+ "step": 96
393
+ },
394
+ {
395
+ "epoch": 5.764705882352941,
396
+ "grad_norm": 0.5573323965072632,
397
+ "learning_rate": 9.992876385052345e-05,
398
+ "loss": 0.2654,
399
+ "step": 98
400
+ },
401
+ {
402
+ "epoch": 5.882352941176471,
403
+ "grad_norm": 0.4234023094177246,
404
+ "learning_rate": 9.990516643685222e-05,
405
+ "loss": 0.2256,
406
+ "step": 100
407
+ },
408
+ {
409
+ "epoch": 6.0,
410
+ "grad_norm": 0.678525984287262,
411
+ "learning_rate": 9.987820251299122e-05,
412
+ "loss": 0.3465,
413
+ "step": 102
414
+ },
415
+ {
416
+ "epoch": 6.0,
417
+ "eval_loss": 0.33739909529685974,
418
+ "eval_runtime": 32.0036,
419
+ "eval_samples_per_second": 1.125,
420
+ "eval_steps_per_second": 1.125,
421
+ "step": 102
422
+ },
423
+ {
424
+ "epoch": 6.117647058823529,
425
+ "grad_norm": 0.4869462251663208,
426
+ "learning_rate": 9.984787389787688e-05,
427
+ "loss": 0.1337,
428
+ "step": 104
429
+ },
430
+ {
431
+ "epoch": 6.235294117647059,
432
+ "grad_norm": 0.45321589708328247,
433
+ "learning_rate": 9.981418263742148e-05,
434
+ "loss": 0.2013,
435
+ "step": 106
436
+ },
437
+ {
438
+ "epoch": 6.352941176470588,
439
+ "grad_norm": 0.3684188425540924,
440
+ "learning_rate": 9.977713100437509e-05,
441
+ "loss": 0.2076,
442
+ "step": 108
443
+ },
444
+ {
445
+ "epoch": 6.470588235294118,
446
+ "grad_norm": 0.41488441824913025,
447
+ "learning_rate": 9.973672149817232e-05,
448
+ "loss": 0.1307,
449
+ "step": 110
450
+ },
451
+ {
452
+ "epoch": 6.588235294117647,
453
+ "grad_norm": 0.5205825567245483,
454
+ "learning_rate": 9.96929568447637e-05,
455
+ "loss": 0.1358,
456
+ "step": 112
457
+ },
458
+ {
459
+ "epoch": 6.705882352941177,
460
+ "grad_norm": 0.5574342608451843,
461
+ "learning_rate": 9.964583999643174e-05,
462
+ "loss": 0.1133,
463
+ "step": 114
464
+ },
465
+ {
466
+ "epoch": 6.823529411764706,
467
+ "grad_norm": 0.8607386350631714,
468
+ "learning_rate": 9.95953741315919e-05,
469
+ "loss": 0.2207,
470
+ "step": 116
471
+ },
472
+ {
473
+ "epoch": 6.9411764705882355,
474
+ "grad_norm": 0.7382214069366455,
475
+ "learning_rate": 9.954156265457801e-05,
476
+ "loss": 0.2498,
477
+ "step": 118
478
+ },
479
+ {
480
+ "epoch": 7.0,
481
+ "eval_loss": 0.35563692450523376,
482
+ "eval_runtime": 31.9979,
483
+ "eval_samples_per_second": 1.125,
484
+ "eval_steps_per_second": 1.125,
485
+ "step": 119
486
+ },
487
+ {
488
+ "epoch": 7.0588235294117645,
489
+ "grad_norm": 0.5253345966339111,
490
+ "learning_rate": 9.948440919541278e-05,
491
+ "loss": 0.2012,
492
+ "step": 120
493
+ },
494
+ {
495
+ "epoch": 7.176470588235294,
496
+ "grad_norm": 0.616706371307373,
497
+ "learning_rate": 9.942391760956277e-05,
498
+ "loss": 0.1718,
499
+ "step": 122
500
+ },
501
+ {
502
+ "epoch": 7.294117647058823,
503
+ "grad_norm": 0.5785579681396484,
504
+ "learning_rate": 9.936009197767845e-05,
505
+ "loss": 0.1278,
506
+ "step": 124
507
+ },
508
+ {
509
+ "epoch": 7.411764705882353,
510
+ "grad_norm": 0.8303074240684509,
511
+ "learning_rate": 9.929293660531888e-05,
512
+ "loss": 0.1065,
513
+ "step": 126
514
+ },
515
+ {
516
+ "epoch": 7.529411764705882,
517
+ "grad_norm": 0.8201299905776978,
518
+ "learning_rate": 9.922245602266118e-05,
519
+ "loss": 0.1027,
520
+ "step": 128
521
+ },
522
+ {
523
+ "epoch": 7.647058823529412,
524
+ "grad_norm": 0.5928418040275574,
525
+ "learning_rate": 9.91486549841951e-05,
526
+ "loss": 0.091,
527
+ "step": 130
528
+ },
529
+ {
530
+ "epoch": 7.764705882352941,
531
+ "grad_norm": 3.87727952003479,
532
+ "learning_rate": 9.90715384684021e-05,
533
+ "loss": 0.1228,
534
+ "step": 132
535
+ },
536
+ {
537
+ "epoch": 7.882352941176471,
538
+ "grad_norm": 0.5786693692207336,
539
+ "learning_rate": 9.899111167741966e-05,
540
+ "loss": 0.1583,
541
+ "step": 134
542
+ },
543
+ {
544
+ "epoch": 8.0,
545
+ "grad_norm": 0.5984454154968262,
546
+ "learning_rate": 9.890738003669029e-05,
547
+ "loss": 0.1023,
548
+ "step": 136
549
+ },
550
+ {
551
+ "epoch": 8.0,
552
+ "eval_loss": 0.41042014956474304,
553
+ "eval_runtime": 32.0113,
554
+ "eval_samples_per_second": 1.125,
555
+ "eval_steps_per_second": 1.125,
556
+ "step": 136
557
+ },
558
+ {
559
+ "epoch": 8.117647058823529,
560
+ "grad_norm": 0.5814339518547058,
561
+ "learning_rate": 9.882034919459555e-05,
562
+ "loss": 0.0595,
563
+ "step": 138
564
+ },
565
+ {
566
+ "epoch": 8.235294117647058,
567
+ "grad_norm": 0.7130110263824463,
568
+ "learning_rate": 9.873002502207503e-05,
569
+ "loss": 0.0864,
570
+ "step": 140
571
+ },
572
+ {
573
+ "epoch": 8.352941176470589,
574
+ "grad_norm": 0.7068061232566833,
575
+ "learning_rate": 9.863641361223024e-05,
576
+ "loss": 0.0951,
577
+ "step": 142
578
+ },
579
+ {
580
+ "epoch": 8.470588235294118,
581
+ "grad_norm": 0.0006599414627999067,
582
+ "learning_rate": 9.853952127991372e-05,
583
+ "loss": 0.0686,
584
+ "step": 144
585
+ },
586
+ {
587
+ "epoch": 8.588235294117647,
588
+ "grad_norm": 0.8574017882347107,
589
+ "learning_rate": 9.843935456130295e-05,
590
+ "loss": 0.0731,
591
+ "step": 146
592
+ },
593
+ {
594
+ "epoch": 8.705882352941176,
595
+ "grad_norm": 0.38401010632514954,
596
+ "learning_rate": 9.833592021345937e-05,
597
+ "loss": 0.0227,
598
+ "step": 148
599
+ },
600
+ {
601
+ "epoch": 8.823529411764707,
602
+ "grad_norm": 0.6860135793685913,
603
+ "learning_rate": 9.822922521387276e-05,
604
+ "loss": 0.0732,
605
+ "step": 150
606
+ },
607
+ {
608
+ "epoch": 8.941176470588236,
609
+ "grad_norm": 1.105839490890503,
610
+ "learning_rate": 9.811927675999036e-05,
611
+ "loss": 0.0862,
612
+ "step": 152
613
+ },
614
+ {
615
+ "epoch": 9.0,
616
+ "eval_loss": 0.4675668179988861,
617
+ "eval_runtime": 32.0127,
618
+ "eval_samples_per_second": 1.125,
619
+ "eval_steps_per_second": 1.125,
620
+ "step": 153
621
+ },
622
+ {
623
+ "epoch": 9.058823529411764,
624
+ "grad_norm": 0.5514603853225708,
625
+ "learning_rate": 9.800608226873142e-05,
626
+ "loss": 0.0442,
627
+ "step": 154
628
+ },
629
+ {
630
+ "epoch": 9.176470588235293,
631
+ "grad_norm": 0.3878594636917114,
632
+ "learning_rate": 9.788964937598689e-05,
633
+ "loss": 0.0347,
634
+ "step": 156
635
+ },
636
+ {
637
+ "epoch": 9.294117647058824,
638
+ "grad_norm": 0.5091602802276611,
639
+ "learning_rate": 9.776998593610428e-05,
640
+ "loss": 0.0354,
641
+ "step": 158
642
+ },
643
+ {
644
+ "epoch": 9.411764705882353,
645
+ "grad_norm": 0.5891442894935608,
646
+ "learning_rate": 9.764710002135784e-05,
647
+ "loss": 0.0444,
648
+ "step": 160
649
+ },
650
+ {
651
+ "epoch": 9.529411764705882,
652
+ "grad_norm": 0.6681481003761292,
653
+ "learning_rate": 9.752099992140399e-05,
654
+ "loss": 0.0281,
655
+ "step": 162
656
+ },
657
+ {
658
+ "epoch": 9.647058823529411,
659
+ "grad_norm": 0.5979675650596619,
660
+ "learning_rate": 9.739169414272217e-05,
661
+ "loss": 0.0385,
662
+ "step": 164
663
+ },
664
+ {
665
+ "epoch": 9.764705882352942,
666
+ "grad_norm": 0.8215705156326294,
667
+ "learning_rate": 9.725919140804099e-05,
668
+ "loss": 0.0464,
669
+ "step": 166
670
+ },
671
+ {
672
+ "epoch": 9.882352941176471,
673
+ "grad_norm": 0.4677208364009857,
674
+ "learning_rate": 9.71235006557497e-05,
675
+ "loss": 0.0284,
676
+ "step": 168
677
+ },
678
+ {
679
+ "epoch": 10.0,
680
+ "grad_norm": 0.5602480173110962,
681
+ "learning_rate": 9.698463103929542e-05,
682
+ "loss": 0.0549,
683
+ "step": 170
684
+ },
685
+ {
686
+ "epoch": 10.0,
687
+ "eval_loss": 0.49783653020858765,
688
+ "eval_runtime": 32.0118,
689
+ "eval_samples_per_second": 1.125,
690
+ "eval_steps_per_second": 1.125,
691
+ "step": 170
692
+ },
693
+ {
694
+ "epoch": 10.117647058823529,
695
+ "grad_norm": 0.38908106088638306,
696
+ "learning_rate": 9.684259192656553e-05,
697
+ "loss": 0.0192,
698
+ "step": 172
699
+ },
700
+ {
701
+ "epoch": 10.235294117647058,
702
+ "grad_norm": 0.6568523049354553,
703
+ "learning_rate": 9.669739289925577e-05,
704
+ "loss": 0.0176,
705
+ "step": 174
706
+ },
707
+ {
708
+ "epoch": 10.352941176470589,
709
+ "grad_norm": 0.6142391562461853,
710
+ "learning_rate": 9.654904375222385e-05,
711
+ "loss": 0.0137,
712
+ "step": 176
713
+ },
714
+ {
715
+ "epoch": 10.470588235294118,
716
+ "grad_norm": 0.6784031391143799,
717
+ "learning_rate": 9.639755449282875e-05,
718
+ "loss": 0.0305,
719
+ "step": 178
720
+ },
721
+ {
722
+ "epoch": 10.588235294117647,
723
+ "grad_norm": 0.5321065783500671,
724
+ "learning_rate": 9.62429353402556e-05,
725
+ "loss": 0.0226,
726
+ "step": 180
727
+ },
728
+ {
729
+ "epoch": 10.705882352941176,
730
+ "grad_norm": 0.5628299117088318,
731
+ "learning_rate": 9.608519672482636e-05,
732
+ "loss": 0.0551,
733
+ "step": 182
734
+ },
735
+ {
736
+ "epoch": 10.823529411764707,
737
+ "grad_norm": 0.4615156054496765,
738
+ "learning_rate": 9.592434928729616e-05,
739
+ "loss": 0.0304,
740
+ "step": 184
741
+ },
742
+ {
743
+ "epoch": 10.941176470588236,
744
+ "grad_norm": 0.7079306244850159,
745
+ "learning_rate": 9.576040387813552e-05,
746
+ "loss": 0.032,
747
+ "step": 186
748
+ },
749
+ {
750
+ "epoch": 11.0,
751
+ "eval_loss": 0.5855510234832764,
752
+ "eval_runtime": 32.0018,
753
+ "eval_samples_per_second": 1.125,
754
+ "eval_steps_per_second": 1.125,
755
+ "step": 187
756
+ },
757
+ {
758
+ "epoch": 11.058823529411764,
759
+ "grad_norm": 0.2959321439266205,
760
+ "learning_rate": 9.559337155679842e-05,
761
+ "loss": 0.0096,
762
+ "step": 188
763
+ },
764
+ {
765
+ "epoch": 11.176470588235293,
766
+ "grad_norm": 0.6475893259048462,
767
+ "learning_rate": 9.542326359097619e-05,
768
+ "loss": 0.0203,
769
+ "step": 190
770
+ },
771
+ {
772
+ "epoch": 11.294117647058824,
773
+ "grad_norm": 1.12827730178833,
774
+ "learning_rate": 9.525009145583745e-05,
775
+ "loss": 0.0221,
776
+ "step": 192
777
+ },
778
+ {
779
+ "epoch": 11.411764705882353,
780
+ "grad_norm": 0.5682121515274048,
781
+ "learning_rate": 9.507386683325404e-05,
782
+ "loss": 0.0179,
783
+ "step": 194
784
+ },
785
+ {
786
+ "epoch": 11.529411764705882,
787
+ "grad_norm": 0.4497404396533966,
788
+ "learning_rate": 9.489460161101291e-05,
789
+ "loss": 0.0236,
790
+ "step": 196
791
+ },
792
+ {
793
+ "epoch": 11.647058823529411,
794
+ "grad_norm": 0.08139002323150635,
795
+ "learning_rate": 9.471230788201429e-05,
796
+ "loss": 0.009,
797
+ "step": 198
798
+ },
799
+ {
800
+ "epoch": 11.764705882352942,
801
+ "grad_norm": 0.5133498311042786,
802
+ "learning_rate": 9.452699794345581e-05,
803
+ "loss": 0.0204,
804
+ "step": 200
805
+ },
806
+ {
807
+ "epoch": 11.882352941176471,
808
+ "grad_norm": 0.38405758142471313,
809
+ "learning_rate": 9.43386842960031e-05,
810
+ "loss": 0.011,
811
+ "step": 202
812
+ },
813
+ {
814
+ "epoch": 12.0,
815
+ "grad_norm": 0.24333085119724274,
816
+ "learning_rate": 9.414737964294636e-05,
817
+ "loss": 0.0072,
818
+ "step": 204
819
+ },
820
+ {
821
+ "epoch": 12.0,
822
+ "eval_loss": 0.5622949600219727,
823
+ "eval_runtime": 32.0008,
824
+ "eval_samples_per_second": 1.125,
825
+ "eval_steps_per_second": 1.125,
826
+ "step": 204
827
+ },
828
+ {
829
+ "epoch": 12.0,
830
+ "step": 204,
831
+ "total_flos": 1.2893298226941133e+17,
832
+ "train_loss": 0.38738249295719845,
833
+ "train_runtime": 4719.6721,
834
+ "train_samples_per_second": 1.441,
835
+ "train_steps_per_second": 0.18
836
+ }
837
+ ],
838
+ "logging_steps": 2,
839
+ "max_steps": 850,
840
+ "num_input_tokens_seen": 0,
841
+ "num_train_epochs": 50,
842
+ "save_steps": 25,
843
+ "stateful_callbacks": {
844
+ "EarlyStoppingCallback": {
845
+ "args": {
846
+ "early_stopping_patience": 7,
847
+ "early_stopping_threshold": 0.0
848
+ },
849
+ "attributes": {
850
+ "early_stopping_patience_counter": 0
851
+ }
852
+ },
853
+ "TrainerControl": {
854
+ "args": {
855
+ "should_epoch_stop": false,
856
+ "should_evaluate": false,
857
+ "should_log": false,
858
+ "should_save": true,
859
+ "should_training_stop": true
860
+ },
861
+ "attributes": {}
862
+ }
863
+ },
864
+ "total_flos": 1.2893298226941133e+17,
865
+ "train_batch_size": 1,
866
+ "trial_name": null,
867
+ "trial_params": null
868
+ }