li-muyang commited on
Commit
2a50cf3
·
verified ·
1 Parent(s): b6d01d3

Model save

Browse files
README.md ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: gemma
4
+ base_model: google/gemma-7b
5
+ tags:
6
+ - trl
7
+ - sft
8
+ - generated_from_trainer
9
+ datasets:
10
+ - generator
11
+ model-index:
12
+ - name: zephyr-7b-gemma-sft
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # zephyr-7b-gemma-sft
20
+
21
+ This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 1.0878
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 2e-05
43
+ - train_batch_size: 4
44
+ - eval_batch_size: 4
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - num_devices: 8
48
+ - gradient_accumulation_steps: 4
49
+ - total_train_batch_size: 128
50
+ - total_eval_batch_size: 32
51
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
+ - lr_scheduler_type: cosine
53
+ - lr_scheduler_warmup_ratio: 0.1
54
+ - num_epochs: 3
55
+
56
+ ### Training results
57
+
58
+ | Training Loss | Epoch | Step | Validation Loss |
59
+ |:-------------:|:------:|:----:|:---------------:|
60
+ | 1.0647 | 0.3339 | 100 | 1.1037 |
61
+ | 0.9963 | 0.6678 | 200 | 1.0778 |
62
+ | 0.9491 | 1.0017 | 300 | 1.0657 |
63
+ | 0.7858 | 1.3356 | 400 | 1.0748 |
64
+ | 0.7589 | 1.6694 | 500 | 1.0504 |
65
+ | 0.6865 | 2.0033 | 600 | 1.1397 |
66
+ | 0.4598 | 2.3372 | 700 | 1.1034 |
67
+ | 0.4212 | 2.6711 | 800 | 1.0878 |
68
+
69
+
70
+ ### Framework versions
71
+
72
+ - Transformers 4.45.2
73
+ - Pytorch 2.5.1+rocm6.2
74
+ - Datasets 3.5.0
75
+ - Tokenizers 0.20.3
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.994991652754591,
3
+ "total_flos": 246978202042368.0,
4
+ "train_loss": 0.8698950265695153,
5
+ "train_runtime": 25956.5038,
6
+ "train_samples": 9500,
7
+ "train_samples_per_second": 4.43,
8
+ "train_steps_per_second": 0.035
9
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 2,
4
+ "eos_token_id": 1,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.45.2"
7
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.994991652754591,
3
+ "total_flos": 246978202042368.0,
4
+ "train_loss": 0.8698950265695153,
5
+ "train_runtime": 25956.5038,
6
+ "train_samples": 9500,
7
+ "train_samples_per_second": 4.43,
8
+ "train_steps_per_second": 0.035
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,1366 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.994991652754591,
5
+ "eval_steps": 100,
6
+ "global_step": 897,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.00333889816360601,
13
+ "grad_norm": 1633.0450710701864,
14
+ "learning_rate": 2.2222222222222224e-07,
15
+ "loss": 14.7496,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.01669449081803005,
20
+ "grad_norm": 1227.5400958571288,
21
+ "learning_rate": 1.111111111111111e-06,
22
+ "loss": 13.7443,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.0333889816360601,
27
+ "grad_norm": 141.7033550856239,
28
+ "learning_rate": 2.222222222222222e-06,
29
+ "loss": 6.5371,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.05008347245409015,
34
+ "grad_norm": 46.656322480877336,
35
+ "learning_rate": 3.3333333333333333e-06,
36
+ "loss": 2.4199,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.0667779632721202,
41
+ "grad_norm": 45.70329256375775,
42
+ "learning_rate": 4.444444444444444e-06,
43
+ "loss": 1.6902,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.08347245409015025,
48
+ "grad_norm": 27.869061679803135,
49
+ "learning_rate": 5.555555555555557e-06,
50
+ "loss": 1.5335,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.1001669449081803,
55
+ "grad_norm": 10.911664477936862,
56
+ "learning_rate": 6.666666666666667e-06,
57
+ "loss": 1.4446,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.11686143572621036,
62
+ "grad_norm": 32.806216575924225,
63
+ "learning_rate": 7.77777777777778e-06,
64
+ "loss": 1.366,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.1335559265442404,
69
+ "grad_norm": 28.435286688487835,
70
+ "learning_rate": 8.888888888888888e-06,
71
+ "loss": 1.2277,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.15025041736227046,
76
+ "grad_norm": 13.779394668165384,
77
+ "learning_rate": 1e-05,
78
+ "loss": 1.179,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.1669449081803005,
83
+ "grad_norm": 8.314121594335042,
84
+ "learning_rate": 1.1111111111111113e-05,
85
+ "loss": 1.0907,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.18363939899833054,
90
+ "grad_norm": 49.47347167605334,
91
+ "learning_rate": 1.2222222222222224e-05,
92
+ "loss": 1.1785,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.2003338898163606,
97
+ "grad_norm": 23.601550693922352,
98
+ "learning_rate": 1.3333333333333333e-05,
99
+ "loss": 1.1147,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.21702838063439064,
104
+ "grad_norm": 20.726802549882166,
105
+ "learning_rate": 1.4444444444444446e-05,
106
+ "loss": 1.1147,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.2337228714524207,
111
+ "grad_norm": 54.27959065632635,
112
+ "learning_rate": 1.555555555555556e-05,
113
+ "loss": 1.118,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.25041736227045075,
118
+ "grad_norm": 34.61698243154044,
119
+ "learning_rate": 1.6666666666666667e-05,
120
+ "loss": 1.0938,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.2671118530884808,
125
+ "grad_norm": 34.48269684581954,
126
+ "learning_rate": 1.7777777777777777e-05,
127
+ "loss": 1.0753,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.2838063439065108,
132
+ "grad_norm": 54.75235207297046,
133
+ "learning_rate": 1.888888888888889e-05,
134
+ "loss": 1.0606,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.3005008347245409,
139
+ "grad_norm": 21.595507981935597,
140
+ "learning_rate": 2e-05,
141
+ "loss": 1.0623,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.31719532554257096,
146
+ "grad_norm": 12.729114013961075,
147
+ "learning_rate": 1.9998105699049984e-05,
148
+ "loss": 1.0475,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.333889816360601,
153
+ "grad_norm": 32.21622997035922,
154
+ "learning_rate": 1.9992423513875158e-05,
155
+ "loss": 1.0647,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.333889816360601,
160
+ "eval_loss": 1.103668212890625,
161
+ "eval_runtime": 99.9993,
162
+ "eval_samples_per_second": 19.13,
163
+ "eval_steps_per_second": 0.6,
164
+ "step": 100
165
+ },
166
+ {
167
+ "epoch": 0.35058430717863104,
168
+ "grad_norm": 44.54084920744845,
169
+ "learning_rate": 1.9982955597229275e-05,
170
+ "loss": 1.0921,
171
+ "step": 105
172
+ },
173
+ {
174
+ "epoch": 0.3672787979966611,
175
+ "grad_norm": 43.035307559559726,
176
+ "learning_rate": 1.9969705536129033e-05,
177
+ "loss": 1.0968,
178
+ "step": 110
179
+ },
180
+ {
181
+ "epoch": 0.38397328881469117,
182
+ "grad_norm": 45.868091031446006,
183
+ "learning_rate": 1.9952678350495104e-05,
184
+ "loss": 1.0572,
185
+ "step": 115
186
+ },
187
+ {
188
+ "epoch": 0.4006677796327212,
189
+ "grad_norm": 22.909320808161933,
190
+ "learning_rate": 1.9931880491250263e-05,
191
+ "loss": 1.0449,
192
+ "step": 120
193
+ },
194
+ {
195
+ "epoch": 0.41736227045075125,
196
+ "grad_norm": 16.217674844349403,
197
+ "learning_rate": 1.990731983787542e-05,
198
+ "loss": 1.0252,
199
+ "step": 125
200
+ },
201
+ {
202
+ "epoch": 0.4340567612687813,
203
+ "grad_norm": 25.232131034909912,
204
+ "learning_rate": 1.987900569542438e-05,
205
+ "loss": 1.0121,
206
+ "step": 130
207
+ },
208
+ {
209
+ "epoch": 0.4507512520868113,
210
+ "grad_norm": 22.126534585692138,
211
+ "learning_rate": 1.9846948790998532e-05,
212
+ "loss": 0.9896,
213
+ "step": 135
214
+ },
215
+ {
216
+ "epoch": 0.4674457429048414,
217
+ "grad_norm": 49.37568755161384,
218
+ "learning_rate": 1.9811161269682776e-05,
219
+ "loss": 1.004,
220
+ "step": 140
221
+ },
222
+ {
223
+ "epoch": 0.48414023372287146,
224
+ "grad_norm": 20.476974911235143,
225
+ "learning_rate": 1.9771656689944238e-05,
226
+ "loss": 1.022,
227
+ "step": 145
228
+ },
229
+ {
230
+ "epoch": 0.5008347245409015,
231
+ "grad_norm": 35.23767882551736,
232
+ "learning_rate": 1.9728450018495506e-05,
233
+ "loss": 1.0309,
234
+ "step": 150
235
+ },
236
+ {
237
+ "epoch": 0.5175292153589316,
238
+ "grad_norm": 17.97928184956425,
239
+ "learning_rate": 1.968155762462433e-05,
240
+ "loss": 1.0587,
241
+ "step": 155
242
+ },
243
+ {
244
+ "epoch": 0.5342237061769616,
245
+ "grad_norm": 19.474927591460286,
246
+ "learning_rate": 1.9630997273991964e-05,
247
+ "loss": 1.0331,
248
+ "step": 160
249
+ },
250
+ {
251
+ "epoch": 0.5509181969949917,
252
+ "grad_norm": 30.437795495910674,
253
+ "learning_rate": 1.9576788121902457e-05,
254
+ "loss": 1.0342,
255
+ "step": 165
256
+ },
257
+ {
258
+ "epoch": 0.5676126878130217,
259
+ "grad_norm": 25.94951957994925,
260
+ "learning_rate": 1.951895070604547e-05,
261
+ "loss": 1.0379,
262
+ "step": 170
263
+ },
264
+ {
265
+ "epoch": 0.5843071786310517,
266
+ "grad_norm": 11.838240662207202,
267
+ "learning_rate": 1.9457506938715357e-05,
268
+ "loss": 1.0265,
269
+ "step": 175
270
+ },
271
+ {
272
+ "epoch": 0.6010016694490818,
273
+ "grad_norm": 12.568669597424641,
274
+ "learning_rate": 1.9392480098509488e-05,
275
+ "loss": 1.0036,
276
+ "step": 180
277
+ },
278
+ {
279
+ "epoch": 0.6176961602671118,
280
+ "grad_norm": 18.395945763500272,
281
+ "learning_rate": 1.93238948215089e-05,
282
+ "loss": 1.0224,
283
+ "step": 185
284
+ },
285
+ {
286
+ "epoch": 0.6343906510851419,
287
+ "grad_norm": 18.80316735007712,
288
+ "learning_rate": 1.9251777091944665e-05,
289
+ "loss": 1.0176,
290
+ "step": 190
291
+ },
292
+ {
293
+ "epoch": 0.6510851419031719,
294
+ "grad_norm": 5.216376439418974,
295
+ "learning_rate": 1.9176154232353513e-05,
296
+ "loss": 1.0245,
297
+ "step": 195
298
+ },
299
+ {
300
+ "epoch": 0.667779632721202,
301
+ "grad_norm": 14.987209529675992,
302
+ "learning_rate": 1.9097054893226395e-05,
303
+ "loss": 0.9963,
304
+ "step": 200
305
+ },
306
+ {
307
+ "epoch": 0.667779632721202,
308
+ "eval_loss": 1.0777970552444458,
309
+ "eval_runtime": 99.7386,
310
+ "eval_samples_per_second": 19.18,
311
+ "eval_steps_per_second": 0.602,
312
+ "step": 200
313
+ },
314
+ {
315
+ "epoch": 0.6844741235392321,
316
+ "grad_norm": 24.67821208903156,
317
+ "learning_rate": 1.9014509042153964e-05,
318
+ "loss": 1.0121,
319
+ "step": 205
320
+ },
321
+ {
322
+ "epoch": 0.7011686143572621,
323
+ "grad_norm": 5.5322384698835405,
324
+ "learning_rate": 1.8928547952473037e-05,
325
+ "loss": 1.0241,
326
+ "step": 210
327
+ },
328
+ {
329
+ "epoch": 0.7178631051752922,
330
+ "grad_norm": 43.27725714311262,
331
+ "learning_rate": 1.8839204191418386e-05,
332
+ "loss": 1.0105,
333
+ "step": 215
334
+ },
335
+ {
336
+ "epoch": 0.7345575959933222,
337
+ "grad_norm": 22.23252550578356,
338
+ "learning_rate": 1.8746511607784298e-05,
339
+ "loss": 1.0254,
340
+ "step": 220
341
+ },
342
+ {
343
+ "epoch": 0.7512520868113522,
344
+ "grad_norm": 22.351531092855776,
345
+ "learning_rate": 1.865050531910062e-05,
346
+ "loss": 1.0304,
347
+ "step": 225
348
+ },
349
+ {
350
+ "epoch": 0.7679465776293823,
351
+ "grad_norm": 35.58389631190103,
352
+ "learning_rate": 1.855122169832813e-05,
353
+ "loss": 1.0116,
354
+ "step": 230
355
+ },
356
+ {
357
+ "epoch": 0.7846410684474123,
358
+ "grad_norm": 30.55541013801255,
359
+ "learning_rate": 1.844869836007825e-05,
360
+ "loss": 1.0068,
361
+ "step": 235
362
+ },
363
+ {
364
+ "epoch": 0.8013355592654424,
365
+ "grad_norm": 30.491225866822745,
366
+ "learning_rate": 1.8342974146362397e-05,
367
+ "loss": 0.9866,
368
+ "step": 240
369
+ },
370
+ {
371
+ "epoch": 0.8180300500834724,
372
+ "grad_norm": 22.716730486325236,
373
+ "learning_rate": 1.8234089111876256e-05,
374
+ "loss": 0.9607,
375
+ "step": 245
376
+ },
377
+ {
378
+ "epoch": 0.8347245409015025,
379
+ "grad_norm": 26.492909425398352,
380
+ "learning_rate": 1.8122084508824692e-05,
381
+ "loss": 0.935,
382
+ "step": 250
383
+ },
384
+ {
385
+ "epoch": 0.8514190317195326,
386
+ "grad_norm": 24.80837932712638,
387
+ "learning_rate": 1.80070027712929e-05,
388
+ "loss": 0.9315,
389
+ "step": 255
390
+ },
391
+ {
392
+ "epoch": 0.8681135225375626,
393
+ "grad_norm": 12.290117913483828,
394
+ "learning_rate": 1.7888887499169816e-05,
395
+ "loss": 0.9636,
396
+ "step": 260
397
+ },
398
+ {
399
+ "epoch": 0.8848080133555927,
400
+ "grad_norm": 14.806034707312756,
401
+ "learning_rate": 1.7767783441629883e-05,
402
+ "loss": 0.9724,
403
+ "step": 265
404
+ },
405
+ {
406
+ "epoch": 0.9015025041736227,
407
+ "grad_norm": 9.947843882385062,
408
+ "learning_rate": 1.7643736480179353e-05,
409
+ "loss": 0.9669,
410
+ "step": 270
411
+ },
412
+ {
413
+ "epoch": 0.9181969949916527,
414
+ "grad_norm": 21.563424938124133,
415
+ "learning_rate": 1.7516793611273614e-05,
416
+ "loss": 0.996,
417
+ "step": 275
418
+ },
419
+ {
420
+ "epoch": 0.9348914858096828,
421
+ "grad_norm": 12.303876339729308,
422
+ "learning_rate": 1.7387002928512093e-05,
423
+ "loss": 0.9829,
424
+ "step": 280
425
+ },
426
+ {
427
+ "epoch": 0.9515859766277128,
428
+ "grad_norm": 12.562858401151901,
429
+ "learning_rate": 1.725441360441752e-05,
430
+ "loss": 0.956,
431
+ "step": 285
432
+ },
433
+ {
434
+ "epoch": 0.9682804674457429,
435
+ "grad_norm": 30.182398173108304,
436
+ "learning_rate": 1.711907587180642e-05,
437
+ "loss": 0.9889,
438
+ "step": 290
439
+ },
440
+ {
441
+ "epoch": 0.9849749582637729,
442
+ "grad_norm": 24.788311254785803,
443
+ "learning_rate": 1.698104100475788e-05,
444
+ "loss": 0.9364,
445
+ "step": 295
446
+ },
447
+ {
448
+ "epoch": 1.001669449081803,
449
+ "grad_norm": 8.024014669533784,
450
+ "learning_rate": 1.684036129918786e-05,
451
+ "loss": 0.9491,
452
+ "step": 300
453
+ },
454
+ {
455
+ "epoch": 1.001669449081803,
456
+ "eval_loss": 1.0656992197036743,
457
+ "eval_runtime": 100.1998,
458
+ "eval_samples_per_second": 19.092,
459
+ "eval_steps_per_second": 0.599,
460
+ "step": 300
461
+ },
462
+ {
463
+ "epoch": 1.018363939899833,
464
+ "grad_norm": 25.34694503716232,
465
+ "learning_rate": 1.6697090053036344e-05,
466
+ "loss": 0.7788,
467
+ "step": 305
468
+ },
469
+ {
470
+ "epoch": 1.0350584307178632,
471
+ "grad_norm": 16.449602651309245,
472
+ "learning_rate": 1.6551281546074863e-05,
473
+ "loss": 0.7857,
474
+ "step": 310
475
+ },
476
+ {
477
+ "epoch": 1.0517529215358932,
478
+ "grad_norm": 14.554716464174657,
479
+ "learning_rate": 1.6402991019342073e-05,
480
+ "loss": 0.7419,
481
+ "step": 315
482
+ },
483
+ {
484
+ "epoch": 1.0684474123539232,
485
+ "grad_norm": 22.643446542701334,
486
+ "learning_rate": 1.625227465421511e-05,
487
+ "loss": 0.7691,
488
+ "step": 320
489
+ },
490
+ {
491
+ "epoch": 1.0851419031719534,
492
+ "grad_norm": 6.658564735729156,
493
+ "learning_rate": 1.60991895511247e-05,
494
+ "loss": 0.7524,
495
+ "step": 325
496
+ },
497
+ {
498
+ "epoch": 1.1018363939899833,
499
+ "grad_norm": 6.00983494629181,
500
+ "learning_rate": 1.5943793707922086e-05,
501
+ "loss": 0.7315,
502
+ "step": 330
503
+ },
504
+ {
505
+ "epoch": 1.1185308848080133,
506
+ "grad_norm": 10.930428697425201,
507
+ "learning_rate": 1.5786145997905952e-05,
508
+ "loss": 0.7511,
509
+ "step": 335
510
+ },
511
+ {
512
+ "epoch": 1.1352253756260433,
513
+ "grad_norm": 5.67952839269914,
514
+ "learning_rate": 1.5626306147517665e-05,
515
+ "loss": 0.7984,
516
+ "step": 340
517
+ },
518
+ {
519
+ "epoch": 1.1519198664440735,
520
+ "grad_norm": 10.358147787261181,
521
+ "learning_rate": 1.5464334713713312e-05,
522
+ "loss": 0.7846,
523
+ "step": 345
524
+ },
525
+ {
526
+ "epoch": 1.1686143572621035,
527
+ "grad_norm": 3.5576580904625,
528
+ "learning_rate": 1.5300293061021084e-05,
529
+ "loss": 0.8156,
530
+ "step": 350
531
+ },
532
+ {
533
+ "epoch": 1.1853088480801335,
534
+ "grad_norm": 10.455272462106366,
535
+ "learning_rate": 1.5134243338292686e-05,
536
+ "loss": 0.8064,
537
+ "step": 355
538
+ },
539
+ {
540
+ "epoch": 1.2020033388981637,
541
+ "grad_norm": 12.388442133679995,
542
+ "learning_rate": 1.4966248455157622e-05,
543
+ "loss": 0.8139,
544
+ "step": 360
545
+ },
546
+ {
547
+ "epoch": 1.2186978297161937,
548
+ "grad_norm": 3.603930889839059,
549
+ "learning_rate": 1.4796372058189235e-05,
550
+ "loss": 0.7864,
551
+ "step": 365
552
+ },
553
+ {
554
+ "epoch": 1.2353923205342237,
555
+ "grad_norm": 7.703996427844556,
556
+ "learning_rate": 1.4624678506791556e-05,
557
+ "loss": 0.8157,
558
+ "step": 370
559
+ },
560
+ {
561
+ "epoch": 1.2520868113522536,
562
+ "grad_norm": 14.186334062276845,
563
+ "learning_rate": 1.445123284881609e-05,
564
+ "loss": 0.8096,
565
+ "step": 375
566
+ },
567
+ {
568
+ "epoch": 1.2687813021702838,
569
+ "grad_norm": 14.091515421765601,
570
+ "learning_rate": 1.4276100795917777e-05,
571
+ "loss": 0.8201,
572
+ "step": 380
573
+ },
574
+ {
575
+ "epoch": 1.2854757929883138,
576
+ "grad_norm": 12.582754871892645,
577
+ "learning_rate": 1.409934869865945e-05,
578
+ "loss": 0.8339,
579
+ "step": 385
580
+ },
581
+ {
582
+ "epoch": 1.302170283806344,
583
+ "grad_norm": 10.868230586622435,
584
+ "learning_rate": 1.392104352137426e-05,
585
+ "loss": 0.7967,
586
+ "step": 390
587
+ },
588
+ {
589
+ "epoch": 1.318864774624374,
590
+ "grad_norm": 10.238679608523467,
591
+ "learning_rate": 1.3741252816795552e-05,
592
+ "loss": 0.798,
593
+ "step": 395
594
+ },
595
+ {
596
+ "epoch": 1.335559265442404,
597
+ "grad_norm": 9.418642992437631,
598
+ "learning_rate": 1.3560044700463824e-05,
599
+ "loss": 0.7858,
600
+ "step": 400
601
+ },
602
+ {
603
+ "epoch": 1.335559265442404,
604
+ "eval_loss": 1.074806571006775,
605
+ "eval_runtime": 99.6284,
606
+ "eval_samples_per_second": 19.201,
607
+ "eval_steps_per_second": 0.602,
608
+ "step": 400
609
+ },
610
+ {
611
+ "epoch": 1.352253756260434,
612
+ "grad_norm": 15.25086544338894,
613
+ "learning_rate": 1.3377487824920459e-05,
614
+ "loss": 0.8023,
615
+ "step": 405
616
+ },
617
+ {
618
+ "epoch": 1.3689482470784642,
619
+ "grad_norm": 14.361720592844934,
620
+ "learning_rate": 1.3193651353698012e-05,
621
+ "loss": 0.7794,
622
+ "step": 410
623
+ },
624
+ {
625
+ "epoch": 1.3856427378964942,
626
+ "grad_norm": 13.1061570632802,
627
+ "learning_rate": 1.30086049351169e-05,
628
+ "loss": 0.7834,
629
+ "step": 415
630
+ },
631
+ {
632
+ "epoch": 1.4023372287145242,
633
+ "grad_norm": 8.35024330990966,
634
+ "learning_rate": 1.2822418675898428e-05,
635
+ "loss": 0.7734,
636
+ "step": 420
637
+ },
638
+ {
639
+ "epoch": 1.4190317195325544,
640
+ "grad_norm": 16.298895202602473,
641
+ "learning_rate": 1.2635163114604131e-05,
642
+ "loss": 0.7662,
643
+ "step": 425
644
+ },
645
+ {
646
+ "epoch": 1.4357262103505843,
647
+ "grad_norm": 17.1487607830759,
648
+ "learning_rate": 1.2446909194911552e-05,
649
+ "loss": 0.7827,
650
+ "step": 430
651
+ },
652
+ {
653
+ "epoch": 1.4524207011686143,
654
+ "grad_norm": 8.830734419080278,
655
+ "learning_rate": 1.2257728238736468e-05,
656
+ "loss": 0.7631,
657
+ "step": 435
658
+ },
659
+ {
660
+ "epoch": 1.4691151919866443,
661
+ "grad_norm": 9.492080719123614,
662
+ "learning_rate": 1.2067691919211879e-05,
663
+ "loss": 0.7434,
664
+ "step": 440
665
+ },
666
+ {
667
+ "epoch": 1.4858096828046745,
668
+ "grad_norm": 9.389190258847595,
669
+ "learning_rate": 1.1876872233533909e-05,
670
+ "loss": 0.7731,
671
+ "step": 445
672
+ },
673
+ {
674
+ "epoch": 1.5025041736227045,
675
+ "grad_norm": 13.015268805104956,
676
+ "learning_rate": 1.1685341475684935e-05,
677
+ "loss": 0.7625,
678
+ "step": 450
679
+ },
680
+ {
681
+ "epoch": 1.5191986644407347,
682
+ "grad_norm": 5.344015360578971,
683
+ "learning_rate": 1.1493172209044259e-05,
684
+ "loss": 0.7334,
685
+ "step": 455
686
+ },
687
+ {
688
+ "epoch": 1.5358931552587647,
689
+ "grad_norm": 4.338959535952978,
690
+ "learning_rate": 1.1300437238896758e-05,
691
+ "loss": 0.7443,
692
+ "step": 460
693
+ },
694
+ {
695
+ "epoch": 1.5525876460767947,
696
+ "grad_norm": 6.590358839089158,
697
+ "learning_rate": 1.1107209584849845e-05,
698
+ "loss": 0.764,
699
+ "step": 465
700
+ },
701
+ {
702
+ "epoch": 1.5692821368948247,
703
+ "grad_norm": 36.303021940666895,
704
+ "learning_rate": 1.0913562453169241e-05,
705
+ "loss": 0.7995,
706
+ "step": 470
707
+ },
708
+ {
709
+ "epoch": 1.5859766277128546,
710
+ "grad_norm": 22.707355631090273,
711
+ "learning_rate": 1.0719569209044047e-05,
712
+ "loss": 0.7731,
713
+ "step": 475
714
+ },
715
+ {
716
+ "epoch": 1.6026711185308848,
717
+ "grad_norm": 7.301747259827321,
718
+ "learning_rate": 1.0525303348791599e-05,
719
+ "loss": 0.7963,
720
+ "step": 480
721
+ },
722
+ {
723
+ "epoch": 1.6193656093489148,
724
+ "grad_norm": 19.393998907091724,
725
+ "learning_rate": 1.0330838472012617e-05,
726
+ "loss": 0.7913,
727
+ "step": 485
728
+ },
729
+ {
730
+ "epoch": 1.636060100166945,
731
+ "grad_norm": 15.229727817223862,
732
+ "learning_rate": 1.0136248253707267e-05,
733
+ "loss": 0.7723,
734
+ "step": 490
735
+ },
736
+ {
737
+ "epoch": 1.652754590984975,
738
+ "grad_norm": 12.20180252707649,
739
+ "learning_rate": 9.94160641636263e-06,
740
+ "loss": 0.7437,
741
+ "step": 495
742
+ },
743
+ {
744
+ "epoch": 1.669449081803005,
745
+ "grad_norm": 17.03672803886117,
746
+ "learning_rate": 9.74698670202218e-06,
747
+ "loss": 0.7589,
748
+ "step": 500
749
+ },
750
+ {
751
+ "epoch": 1.669449081803005,
752
+ "eval_loss": 1.0504279136657715,
753
+ "eval_runtime": 100.5762,
754
+ "eval_samples_per_second": 19.02,
755
+ "eval_steps_per_second": 0.597,
756
+ "step": 500
757
+ },
758
+ {
759
+ "epoch": 1.686143572621035,
760
+ "grad_norm": 19.130122192738536,
761
+ "learning_rate": 9.552462844347883e-06,
762
+ "loss": 0.7534,
763
+ "step": 505
764
+ },
765
+ {
766
+ "epoch": 1.702838063439065,
767
+ "grad_norm": 2.7398317735628934,
768
+ "learning_rate": 9.358108540685406e-06,
769
+ "loss": 0.7334,
770
+ "step": 510
771
+ },
772
+ {
773
+ "epoch": 1.7195325542570952,
774
+ "grad_norm": 2.773227448421825,
775
+ "learning_rate": 9.163997424143167e-06,
776
+ "loss": 0.7386,
777
+ "step": 515
778
+ },
779
+ {
780
+ "epoch": 1.7362270450751254,
781
+ "grad_norm": 3.5377228178855766,
782
+ "learning_rate": 8.970203035695662e-06,
783
+ "loss": 0.7494,
784
+ "step": 520
785
+ },
786
+ {
787
+ "epoch": 1.7529215358931554,
788
+ "grad_norm": 4.684609025076765,
789
+ "learning_rate": 8.776798796321715e-06,
790
+ "loss": 0.758,
791
+ "step": 525
792
+ },
793
+ {
794
+ "epoch": 1.7696160267111853,
795
+ "grad_norm": 3.7108648922370056,
796
+ "learning_rate": 8.583857979188203e-06,
797
+ "loss": 0.7618,
798
+ "step": 530
799
+ },
800
+ {
801
+ "epoch": 1.7863105175292153,
802
+ "grad_norm": 4.002429802468382,
803
+ "learning_rate": 8.391453681889772e-06,
804
+ "loss": 0.7798,
805
+ "step": 535
806
+ },
807
+ {
808
+ "epoch": 1.8030050083472453,
809
+ "grad_norm": 5.846086197454255,
810
+ "learning_rate": 8.199658798755048e-06,
811
+ "loss": 0.8123,
812
+ "step": 540
813
+ },
814
+ {
815
+ "epoch": 1.8196994991652755,
816
+ "grad_norm": 3.63187175327646,
817
+ "learning_rate": 8.008545993229897e-06,
818
+ "loss": 0.7597,
819
+ "step": 545
820
+ },
821
+ {
822
+ "epoch": 1.8363939899833055,
823
+ "grad_norm": 3.906677171647028,
824
+ "learning_rate": 7.818187670348133e-06,
825
+ "loss": 0.7439,
826
+ "step": 550
827
+ },
828
+ {
829
+ "epoch": 1.8530884808013357,
830
+ "grad_norm": 4.310021563042138,
831
+ "learning_rate": 7.628655949300133e-06,
832
+ "loss": 0.7874,
833
+ "step": 555
834
+ },
835
+ {
836
+ "epoch": 1.8697829716193657,
837
+ "grad_norm": 4.889792328402707,
838
+ "learning_rate": 7.440022636109742e-06,
839
+ "loss": 0.7859,
840
+ "step": 560
841
+ },
842
+ {
843
+ "epoch": 1.8864774624373957,
844
+ "grad_norm": 7.215890316853048,
845
+ "learning_rate": 7.2523591964298345e-06,
846
+ "loss": 0.7345,
847
+ "step": 565
848
+ },
849
+ {
850
+ "epoch": 1.9031719532554257,
851
+ "grad_norm": 5.301408965949358,
852
+ "learning_rate": 7.065736728466832e-06,
853
+ "loss": 0.7549,
854
+ "step": 570
855
+ },
856
+ {
857
+ "epoch": 1.9198664440734556,
858
+ "grad_norm": 7.773497463412357,
859
+ "learning_rate": 6.880225936044402e-06,
860
+ "loss": 0.7533,
861
+ "step": 575
862
+ },
863
+ {
864
+ "epoch": 1.9365609348914858,
865
+ "grad_norm": 6.2446358939825215,
866
+ "learning_rate": 6.695897101816606e-06,
867
+ "loss": 0.7389,
868
+ "step": 580
869
+ },
870
+ {
871
+ "epoch": 1.9532554257095158,
872
+ "grad_norm": 7.2686173097914795,
873
+ "learning_rate": 6.512820060640608e-06,
874
+ "loss": 0.77,
875
+ "step": 585
876
+ },
877
+ {
878
+ "epoch": 1.969949916527546,
879
+ "grad_norm": 5.1942830728763125,
880
+ "learning_rate": 6.331064173119008e-06,
881
+ "loss": 0.7448,
882
+ "step": 590
883
+ },
884
+ {
885
+ "epoch": 1.986644407345576,
886
+ "grad_norm": 6.78958078868122,
887
+ "learning_rate": 6.150698299321889e-06,
888
+ "loss": 0.7397,
889
+ "step": 595
890
+ },
891
+ {
892
+ "epoch": 2.003338898163606,
893
+ "grad_norm": 3.1972134110464983,
894
+ "learning_rate": 5.971790772698467e-06,
895
+ "loss": 0.6865,
896
+ "step": 600
897
+ },
898
+ {
899
+ "epoch": 2.003338898163606,
900
+ "eval_loss": 1.1396924257278442,
901
+ "eval_runtime": 99.6851,
902
+ "eval_samples_per_second": 19.19,
903
+ "eval_steps_per_second": 0.602,
904
+ "step": 600
905
+ },
906
+ {
907
+ "epoch": 2.020033388981636,
908
+ "grad_norm": 6.784157421991238,
909
+ "learning_rate": 5.794409374188272e-06,
910
+ "loss": 0.4715,
911
+ "step": 605
912
+ },
913
+ {
914
+ "epoch": 2.036727879799666,
915
+ "grad_norm": 3.852671208279914,
916
+ "learning_rate": 5.61862130654165e-06,
917
+ "loss": 0.4653,
918
+ "step": 610
919
+ },
920
+ {
921
+ "epoch": 2.053422370617696,
922
+ "grad_norm": 5.2904071685387235,
923
+ "learning_rate": 5.444493168859304e-06,
924
+ "loss": 0.4449,
925
+ "step": 615
926
+ },
927
+ {
928
+ "epoch": 2.0701168614357264,
929
+ "grad_norm": 3.179776819836151,
930
+ "learning_rate": 5.272090931360564e-06,
931
+ "loss": 0.4619,
932
+ "step": 620
933
+ },
934
+ {
935
+ "epoch": 2.0868113522537564,
936
+ "grad_norm": 4.305851121327948,
937
+ "learning_rate": 5.101479910389888e-06,
938
+ "loss": 0.4594,
939
+ "step": 625
940
+ },
941
+ {
942
+ "epoch": 2.1035058430717863,
943
+ "grad_norm": 4.502666240394055,
944
+ "learning_rate": 4.932724743671089e-06,
945
+ "loss": 0.4823,
946
+ "step": 630
947
+ },
948
+ {
949
+ "epoch": 2.1202003338898163,
950
+ "grad_norm": 4.247884777704646,
951
+ "learning_rate": 4.765889365818708e-06,
952
+ "loss": 0.4599,
953
+ "step": 635
954
+ },
955
+ {
956
+ "epoch": 2.1368948247078463,
957
+ "grad_norm": 3.14788698725782,
958
+ "learning_rate": 4.601036984115684e-06,
959
+ "loss": 0.4638,
960
+ "step": 640
961
+ },
962
+ {
963
+ "epoch": 2.1535893155258763,
964
+ "grad_norm": 2.0716891153315298,
965
+ "learning_rate": 4.438230054566678e-06,
966
+ "loss": 0.4389,
967
+ "step": 645
968
+ },
969
+ {
970
+ "epoch": 2.1702838063439067,
971
+ "grad_norm": 2.5090253257766233,
972
+ "learning_rate": 4.277530258235955e-06,
973
+ "loss": 0.4636,
974
+ "step": 650
975
+ },
976
+ {
977
+ "epoch": 2.1869782971619367,
978
+ "grad_norm": 2.033141055907998,
979
+ "learning_rate": 4.118998477878879e-06,
980
+ "loss": 0.4584,
981
+ "step": 655
982
+ },
983
+ {
984
+ "epoch": 2.2036727879799667,
985
+ "grad_norm": 2.3730277065558503,
986
+ "learning_rate": 3.96269477487588e-06,
987
+ "loss": 0.4769,
988
+ "step": 660
989
+ },
990
+ {
991
+ "epoch": 2.2203672787979967,
992
+ "grad_norm": 2.415059497973404,
993
+ "learning_rate": 3.8086783664775827e-06,
994
+ "loss": 0.4433,
995
+ "step": 665
996
+ },
997
+ {
998
+ "epoch": 2.2370617696160267,
999
+ "grad_norm": 2.5774448857421075,
1000
+ "learning_rate": 3.657007603369728e-06,
1001
+ "loss": 0.4585,
1002
+ "step": 670
1003
+ },
1004
+ {
1005
+ "epoch": 2.2537562604340566,
1006
+ "grad_norm": 2.0047006393140134,
1007
+ "learning_rate": 3.5077399475664474e-06,
1008
+ "loss": 0.4567,
1009
+ "step": 675
1010
+ },
1011
+ {
1012
+ "epoch": 2.2704507512520866,
1013
+ "grad_norm": 2.311110170562805,
1014
+ "learning_rate": 3.360931950640185e-06,
1015
+ "loss": 0.4741,
1016
+ "step": 680
1017
+ },
1018
+ {
1019
+ "epoch": 2.287145242070117,
1020
+ "grad_norm": 2.7544264361024116,
1021
+ "learning_rate": 3.2166392322965423e-06,
1022
+ "loss": 0.464,
1023
+ "step": 685
1024
+ },
1025
+ {
1026
+ "epoch": 2.303839732888147,
1027
+ "grad_norm": 2.4224353237379437,
1028
+ "learning_rate": 3.074916459302211e-06,
1029
+ "loss": 0.4434,
1030
+ "step": 690
1031
+ },
1032
+ {
1033
+ "epoch": 2.320534223706177,
1034
+ "grad_norm": 2.4571471461227916,
1035
+ "learning_rate": 2.935817324773893e-06,
1036
+ "loss": 0.4438,
1037
+ "step": 695
1038
+ },
1039
+ {
1040
+ "epoch": 2.337228714524207,
1041
+ "grad_norm": 2.3643126227587183,
1042
+ "learning_rate": 2.799394527836129e-06,
1043
+ "loss": 0.4598,
1044
+ "step": 700
1045
+ },
1046
+ {
1047
+ "epoch": 2.337228714524207,
1048
+ "eval_loss": 1.1034263372421265,
1049
+ "eval_runtime": 100.5283,
1050
+ "eval_samples_per_second": 19.029,
1051
+ "eval_steps_per_second": 0.597,
1052
+ "step": 700
1053
+ },
1054
+ {
1055
+ "epoch": 2.353923205342237,
1056
+ "grad_norm": 2.0528401010873862,
1057
+ "learning_rate": 2.665699753655684e-06,
1058
+ "loss": 0.4524,
1059
+ "step": 705
1060
+ },
1061
+ {
1062
+ "epoch": 2.370617696160267,
1063
+ "grad_norm": 2.5579878502862856,
1064
+ "learning_rate": 2.5347836538601113e-06,
1065
+ "loss": 0.4375,
1066
+ "step": 710
1067
+ },
1068
+ {
1069
+ "epoch": 2.3873121869782974,
1070
+ "grad_norm": 3.3751318304530016,
1071
+ "learning_rate": 2.406695827347848e-06,
1072
+ "loss": 0.4519,
1073
+ "step": 715
1074
+ },
1075
+ {
1076
+ "epoch": 2.4040066777963274,
1077
+ "grad_norm": 2.2219251853782773,
1078
+ "learning_rate": 2.281484801497186e-06,
1079
+ "loss": 0.4408,
1080
+ "step": 720
1081
+ },
1082
+ {
1083
+ "epoch": 2.4207011686143574,
1084
+ "grad_norm": 1.9197182288045105,
1085
+ "learning_rate": 2.1591980137811684e-06,
1086
+ "loss": 0.4517,
1087
+ "step": 725
1088
+ },
1089
+ {
1090
+ "epoch": 2.4373956594323873,
1091
+ "grad_norm": 2.0748089118705266,
1092
+ "learning_rate": 2.0398817937954275e-06,
1093
+ "loss": 0.4428,
1094
+ "step": 730
1095
+ },
1096
+ {
1097
+ "epoch": 2.4540901502504173,
1098
+ "grad_norm": 2.029427151010182,
1099
+ "learning_rate": 1.923581345705736e-06,
1100
+ "loss": 0.4316,
1101
+ "step": 735
1102
+ },
1103
+ {
1104
+ "epoch": 2.4707846410684473,
1105
+ "grad_norm": 1.8596809439641007,
1106
+ "learning_rate": 1.8103407311219523e-06,
1107
+ "loss": 0.4517,
1108
+ "step": 740
1109
+ },
1110
+ {
1111
+ "epoch": 2.4874791318864773,
1112
+ "grad_norm": 2.287602775089728,
1113
+ "learning_rate": 1.7002028524048354e-06,
1114
+ "loss": 0.4301,
1115
+ "step": 745
1116
+ },
1117
+ {
1118
+ "epoch": 2.5041736227045073,
1119
+ "grad_norm": 2.514456715654159,
1120
+ "learning_rate": 1.5932094364120453e-06,
1121
+ "loss": 0.4415,
1122
+ "step": 750
1123
+ },
1124
+ {
1125
+ "epoch": 2.5208681135225377,
1126
+ "grad_norm": 1.963821507045021,
1127
+ "learning_rate": 1.489401018689488e-06,
1128
+ "loss": 0.4315,
1129
+ "step": 755
1130
+ },
1131
+ {
1132
+ "epoch": 2.5375626043405677,
1133
+ "grad_norm": 2.015989543520363,
1134
+ "learning_rate": 1.3888169281140284e-06,
1135
+ "loss": 0.4283,
1136
+ "step": 760
1137
+ },
1138
+ {
1139
+ "epoch": 2.5542570951585977,
1140
+ "grad_norm": 1.7832121894558464,
1141
+ "learning_rate": 1.291495271993337e-06,
1142
+ "loss": 0.4376,
1143
+ "step": 765
1144
+ },
1145
+ {
1146
+ "epoch": 2.5709515859766277,
1147
+ "grad_norm": 1.7494948268844588,
1148
+ "learning_rate": 1.1974729216285386e-06,
1149
+ "loss": 0.4457,
1150
+ "step": 770
1151
+ },
1152
+ {
1153
+ "epoch": 2.5876460767946576,
1154
+ "grad_norm": 1.7570121673428833,
1155
+ "learning_rate": 1.1067854983451575e-06,
1156
+ "loss": 0.4213,
1157
+ "step": 775
1158
+ },
1159
+ {
1160
+ "epoch": 2.604340567612688,
1161
+ "grad_norm": 2.024952733389987,
1162
+ "learning_rate": 1.0194673599976134e-06,
1163
+ "loss": 0.4241,
1164
+ "step": 780
1165
+ },
1166
+ {
1167
+ "epoch": 2.621035058430718,
1168
+ "grad_norm": 1.8656506162126876,
1169
+ "learning_rate": 9.355515879523858e-07,
1170
+ "loss": 0.4163,
1171
+ "step": 785
1172
+ },
1173
+ {
1174
+ "epoch": 2.637729549248748,
1175
+ "grad_norm": 2.4796135500044536,
1176
+ "learning_rate": 8.550699745548196e-07,
1177
+ "loss": 0.4431,
1178
+ "step": 790
1179
+ },
1180
+ {
1181
+ "epoch": 2.654424040066778,
1182
+ "grad_norm": 1.9687379265836358,
1183
+ "learning_rate": 7.780530110842566e-07,
1184
+ "loss": 0.4573,
1185
+ "step": 795
1186
+ },
1187
+ {
1188
+ "epoch": 2.671118530884808,
1189
+ "grad_norm": 1.6950357769626416,
1190
+ "learning_rate": 7.045298762021125e-07,
1191
+ "loss": 0.4212,
1192
+ "step": 800
1193
+ },
1194
+ {
1195
+ "epoch": 2.671118530884808,
1196
+ "eval_loss": 1.0877529382705688,
1197
+ "eval_runtime": 99.55,
1198
+ "eval_samples_per_second": 19.216,
1199
+ "eval_steps_per_second": 0.603,
1200
+ "step": 800
1201
+ },
1202
+ {
1203
+ "epoch": 2.687813021702838,
1204
+ "grad_norm": 1.7541078654573505,
1205
+ "learning_rate": 6.345284248972383e-07,
1206
+ "loss": 0.427,
1207
+ "step": 805
1208
+ },
1209
+ {
1210
+ "epoch": 2.704507512520868,
1211
+ "grad_norm": 1.7383407123514758,
1212
+ "learning_rate": 5.680751779327742e-07,
1213
+ "loss": 0.4156,
1214
+ "step": 810
1215
+ },
1216
+ {
1217
+ "epoch": 2.721202003338898,
1218
+ "grad_norm": 1.6815769561893499,
1219
+ "learning_rate": 5.05195311798491e-07,
1220
+ "loss": 0.4225,
1221
+ "step": 815
1222
+ },
1223
+ {
1224
+ "epoch": 2.7378964941569284,
1225
+ "grad_norm": 1.6242201677795636,
1226
+ "learning_rate": 4.4591264917242195e-07,
1227
+ "loss": 0.4069,
1228
+ "step": 820
1229
+ },
1230
+ {
1231
+ "epoch": 2.7545909849749584,
1232
+ "grad_norm": 1.76260741262681,
1233
+ "learning_rate": 3.9024964989539227e-07,
1234
+ "loss": 0.4359,
1235
+ "step": 825
1236
+ },
1237
+ {
1238
+ "epoch": 2.7712854757929883,
1239
+ "grad_norm": 2.3855493539801143,
1240
+ "learning_rate": 3.3822740246188477e-07,
1241
+ "loss": 0.4179,
1242
+ "step": 830
1243
+ },
1244
+ {
1245
+ "epoch": 2.7879799666110183,
1246
+ "grad_norm": 1.7206678326179792,
1247
+ "learning_rate": 2.8986561603044694e-07,
1248
+ "loss": 0.4321,
1249
+ "step": 835
1250
+ },
1251
+ {
1252
+ "epoch": 2.8046744574290483,
1253
+ "grad_norm": 1.7021364894136357,
1254
+ "learning_rate": 2.4518261295667255e-07,
1255
+ "loss": 0.4328,
1256
+ "step": 840
1257
+ },
1258
+ {
1259
+ "epoch": 2.8213689482470787,
1260
+ "grad_norm": 1.641003975157378,
1261
+ "learning_rate": 2.0419532185159796e-07,
1262
+ "loss": 0.4344,
1263
+ "step": 845
1264
+ },
1265
+ {
1266
+ "epoch": 2.8380634390651087,
1267
+ "grad_norm": 1.770702981651617,
1268
+ "learning_rate": 1.6691927116812002e-07,
1269
+ "loss": 0.4197,
1270
+ "step": 850
1271
+ },
1272
+ {
1273
+ "epoch": 2.8547579298831387,
1274
+ "grad_norm": 1.6998058078972085,
1275
+ "learning_rate": 1.3336858331787993e-07,
1276
+ "loss": 0.4218,
1277
+ "step": 855
1278
+ },
1279
+ {
1280
+ "epoch": 2.8714524207011687,
1281
+ "grad_norm": 1.8485073174343698,
1282
+ "learning_rate": 1.0355596932085432e-07,
1283
+ "loss": 0.4397,
1284
+ "step": 860
1285
+ },
1286
+ {
1287
+ "epoch": 2.8881469115191987,
1288
+ "grad_norm": 1.6420438723831818,
1289
+ "learning_rate": 7.749272398964613e-08,
1290
+ "loss": 0.4053,
1291
+ "step": 865
1292
+ },
1293
+ {
1294
+ "epoch": 2.9048414023372287,
1295
+ "grad_norm": 1.6830498720838685,
1296
+ "learning_rate": 5.518872165033329e-08,
1297
+ "loss": 0.4137,
1298
+ "step": 870
1299
+ },
1300
+ {
1301
+ "epoch": 2.9215358931552586,
1302
+ "grad_norm": 1.6815426021448905,
1303
+ "learning_rate": 3.6652412401478875e-08,
1304
+ "loss": 0.4117,
1305
+ "step": 875
1306
+ },
1307
+ {
1308
+ "epoch": 2.9382303839732886,
1309
+ "grad_norm": 1.6703264964322584,
1310
+ "learning_rate": 2.1890818912728706e-08,
1311
+ "loss": 0.4217,
1312
+ "step": 880
1313
+ },
1314
+ {
1315
+ "epoch": 2.9549248747913186,
1316
+ "grad_norm": 1.7642691508988728,
1317
+ "learning_rate": 1.0909533764194013e-08,
1318
+ "loss": 0.4179,
1319
+ "step": 885
1320
+ },
1321
+ {
1322
+ "epoch": 2.971619365609349,
1323
+ "grad_norm": 1.852312064422515,
1324
+ "learning_rate": 3.7127173276563234e-09,
1325
+ "loss": 0.4167,
1326
+ "step": 890
1327
+ },
1328
+ {
1329
+ "epoch": 2.988313856427379,
1330
+ "grad_norm": 1.6438587562151536,
1331
+ "learning_rate": 3.0309619035495675e-10,
1332
+ "loss": 0.4259,
1333
+ "step": 895
1334
+ },
1335
+ {
1336
+ "epoch": 2.994991652754591,
1337
+ "step": 897,
1338
+ "total_flos": 246978202042368.0,
1339
+ "train_loss": 0.8698950265695153,
1340
+ "train_runtime": 25956.5038,
1341
+ "train_samples_per_second": 4.43,
1342
+ "train_steps_per_second": 0.035
1343
+ }
1344
+ ],
1345
+ "logging_steps": 5,
1346
+ "max_steps": 897,
1347
+ "num_input_tokens_seen": 0,
1348
+ "num_train_epochs": 3,
1349
+ "save_steps": 100,
1350
+ "stateful_callbacks": {
1351
+ "TrainerControl": {
1352
+ "args": {
1353
+ "should_epoch_stop": false,
1354
+ "should_evaluate": false,
1355
+ "should_log": false,
1356
+ "should_save": true,
1357
+ "should_training_stop": true
1358
+ },
1359
+ "attributes": {}
1360
+ }
1361
+ },
1362
+ "total_flos": 246978202042368.0,
1363
+ "train_batch_size": 4,
1364
+ "trial_name": null,
1365
+ "trial_params": null
1366
+ }