chansung commited on
Commit
b0bd756
·
verified ·
1 Parent(s): 9855f6d

Model save

Browse files
Files changed (4) hide show
  1. README.md +78 -0
  2. all_results.json +9 -0
  3. train_results.json +9 -0
  4. trainer_state.json +1655 -0
README.md ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Meta-Llama-3-8B
3
+ datasets:
4
+ - generator
5
+ library_name: peft
6
+ license: llama3
7
+ tags:
8
+ - trl
9
+ - sft
10
+ - generated_from_trainer
11
+ model-index:
12
+ - name: llama3-8b-summarize-gpt4o-128k
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
+ # llama3-8b-summarize-gpt4o-128k
20
+
21
+ This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 2.2606
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: 0.0002
43
+ - train_batch_size: 8
44
+ - eval_batch_size: 4
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - num_devices: 8
48
+ - gradient_accumulation_steps: 2
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: 10
55
+
56
+ ### Training results
57
+
58
+ | Training Loss | Epoch | Step | Validation Loss |
59
+ |:-------------:|:------:|:----:|:---------------:|
60
+ | 0.8176 | 0.9954 | 109 | 2.1150 |
61
+ | 0.7464 | 2.0 | 219 | 2.1313 |
62
+ | 0.7128 | 2.9954 | 328 | 2.1444 |
63
+ | 0.6924 | 4.0 | 438 | 2.1631 |
64
+ | 0.6777 | 4.9954 | 547 | 2.1823 |
65
+ | 0.6526 | 6.0 | 657 | 2.2078 |
66
+ | 0.6326 | 6.9954 | 766 | 2.2296 |
67
+ | 0.6311 | 8.0 | 876 | 2.2485 |
68
+ | 0.6233 | 8.9954 | 985 | 2.2587 |
69
+ | 0.6194 | 9.9543 | 1090 | 2.2606 |
70
+
71
+
72
+ ### Framework versions
73
+
74
+ - PEFT 0.12.0
75
+ - Transformers 4.44.0
76
+ - Pytorch 2.2.0+cu121
77
+ - Datasets 2.20.0
78
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.954337899543379,
3
+ "total_flos": 6.456679991336763e+18,
4
+ "train_loss": 0.7130365929472338,
5
+ "train_runtime": 9668.9725,
6
+ "train_samples": 129221,
7
+ "train_samples_per_second": 14.452,
8
+ "train_steps_per_second": 0.113
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.954337899543379,
3
+ "total_flos": 6.456679991336763e+18,
4
+ "train_loss": 0.7130365929472338,
5
+ "train_runtime": 9668.9725,
6
+ "train_samples": 129221,
7
+ "train_samples_per_second": 14.452,
8
+ "train_steps_per_second": 0.113
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,1655 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 9.954337899543379,
5
+ "eval_steps": 500,
6
+ "global_step": 1090,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0091324200913242,
13
+ "grad_norm": 3.1570751667022705,
14
+ "learning_rate": 1.8348623853211011e-06,
15
+ "loss": 2.2349,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.045662100456621,
20
+ "grad_norm": 3.327242374420166,
21
+ "learning_rate": 9.174311926605506e-06,
22
+ "loss": 2.2947,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.091324200913242,
27
+ "grad_norm": 3.6639091968536377,
28
+ "learning_rate": 1.834862385321101e-05,
29
+ "loss": 2.1754,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.136986301369863,
34
+ "grad_norm": 6.006739616394043,
35
+ "learning_rate": 2.7522935779816515e-05,
36
+ "loss": 1.8388,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.182648401826484,
41
+ "grad_norm": 1.5798718929290771,
42
+ "learning_rate": 3.669724770642202e-05,
43
+ "loss": 1.4696,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.228310502283105,
48
+ "grad_norm": 1.038404107093811,
49
+ "learning_rate": 4.587155963302753e-05,
50
+ "loss": 1.2583,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.273972602739726,
55
+ "grad_norm": 0.5968286395072937,
56
+ "learning_rate": 5.504587155963303e-05,
57
+ "loss": 1.1721,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.319634703196347,
62
+ "grad_norm": 0.42477676272392273,
63
+ "learning_rate": 6.422018348623854e-05,
64
+ "loss": 1.0966,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.365296803652968,
69
+ "grad_norm": 0.46538373827934265,
70
+ "learning_rate": 7.339449541284404e-05,
71
+ "loss": 1.0378,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.410958904109589,
76
+ "grad_norm": 0.32725027203559875,
77
+ "learning_rate": 8.256880733944955e-05,
78
+ "loss": 0.996,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.45662100456621,
83
+ "grad_norm": 0.27102047204971313,
84
+ "learning_rate": 9.174311926605506e-05,
85
+ "loss": 0.9549,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.502283105022831,
90
+ "grad_norm": 0.20055657625198364,
91
+ "learning_rate": 0.00010091743119266055,
92
+ "loss": 0.9426,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.547945205479452,
97
+ "grad_norm": 0.24150966107845306,
98
+ "learning_rate": 0.00011009174311926606,
99
+ "loss": 0.918,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.593607305936073,
104
+ "grad_norm": 0.20545370876789093,
105
+ "learning_rate": 0.00011926605504587157,
106
+ "loss": 0.8995,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.639269406392694,
111
+ "grad_norm": 0.24165408313274384,
112
+ "learning_rate": 0.00012844036697247707,
113
+ "loss": 0.8851,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.684931506849315,
118
+ "grad_norm": 0.22626952826976776,
119
+ "learning_rate": 0.00013761467889908258,
120
+ "loss": 0.8729,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.730593607305936,
125
+ "grad_norm": 0.2153974175453186,
126
+ "learning_rate": 0.0001467889908256881,
127
+ "loss": 0.8588,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.776255707762557,
132
+ "grad_norm": 0.22552534937858582,
133
+ "learning_rate": 0.0001559633027522936,
134
+ "loss": 0.8453,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.821917808219178,
139
+ "grad_norm": 0.2299458086490631,
140
+ "learning_rate": 0.0001651376146788991,
141
+ "loss": 0.8491,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.867579908675799,
146
+ "grad_norm": 0.38727596402168274,
147
+ "learning_rate": 0.00017431192660550458,
148
+ "loss": 0.839,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.91324200913242,
153
+ "grad_norm": 0.5172630548477173,
154
+ "learning_rate": 0.00018348623853211012,
155
+ "loss": 0.8359,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.958904109589041,
160
+ "grad_norm": 0.3504508137702942,
161
+ "learning_rate": 0.0001926605504587156,
162
+ "loss": 0.8176,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.9954337899543378,
167
+ "eval_loss": 2.11497163772583,
168
+ "eval_runtime": 0.7274,
169
+ "eval_samples_per_second": 13.748,
170
+ "eval_steps_per_second": 1.375,
171
+ "step": 109
172
+ },
173
+ {
174
+ "epoch": 1.004566210045662,
175
+ "grad_norm": 0.21483348309993744,
176
+ "learning_rate": 0.00019999948721966259,
177
+ "loss": 0.812,
178
+ "step": 110
179
+ },
180
+ {
181
+ "epoch": 1.0502283105022832,
182
+ "grad_norm": 0.25326865911483765,
183
+ "learning_rate": 0.00019998154046002822,
184
+ "loss": 0.8028,
185
+ "step": 115
186
+ },
187
+ {
188
+ "epoch": 1.095890410958904,
189
+ "grad_norm": 0.23645517230033875,
190
+ "learning_rate": 0.0001999379599421534,
191
+ "loss": 0.8003,
192
+ "step": 120
193
+ },
194
+ {
195
+ "epoch": 1.1415525114155252,
196
+ "grad_norm": 0.22504329681396484,
197
+ "learning_rate": 0.00019986875683942535,
198
+ "loss": 0.7939,
199
+ "step": 125
200
+ },
201
+ {
202
+ "epoch": 1.187214611872146,
203
+ "grad_norm": 0.2312326580286026,
204
+ "learning_rate": 0.00019977394889447524,
205
+ "loss": 0.7847,
206
+ "step": 130
207
+ },
208
+ {
209
+ "epoch": 1.2328767123287672,
210
+ "grad_norm": 0.21880114078521729,
211
+ "learning_rate": 0.00019965356041462955,
212
+ "loss": 0.7856,
213
+ "step": 135
214
+ },
215
+ {
216
+ "epoch": 1.278538812785388,
217
+ "grad_norm": 0.19915428757667542,
218
+ "learning_rate": 0.00019950762226567781,
219
+ "loss": 0.7772,
220
+ "step": 140
221
+ },
222
+ {
223
+ "epoch": 1.3242009132420092,
224
+ "grad_norm": 0.21661226451396942,
225
+ "learning_rate": 0.00019933617186395917,
226
+ "loss": 0.7698,
227
+ "step": 145
228
+ },
229
+ {
230
+ "epoch": 1.36986301369863,
231
+ "grad_norm": 0.23418781161308289,
232
+ "learning_rate": 0.00019913925316676945,
233
+ "loss": 0.771,
234
+ "step": 150
235
+ },
236
+ {
237
+ "epoch": 1.4155251141552512,
238
+ "grad_norm": 0.2712844908237457,
239
+ "learning_rate": 0.00019891691666109113,
240
+ "loss": 0.7754,
241
+ "step": 155
242
+ },
243
+ {
244
+ "epoch": 1.461187214611872,
245
+ "grad_norm": 0.20658765733242035,
246
+ "learning_rate": 0.00019866921935064906,
247
+ "loss": 0.772,
248
+ "step": 160
249
+ },
250
+ {
251
+ "epoch": 1.5068493150684932,
252
+ "grad_norm": 0.2858370542526245,
253
+ "learning_rate": 0.00019839622474129596,
254
+ "loss": 0.7648,
255
+ "step": 165
256
+ },
257
+ {
258
+ "epoch": 1.5525114155251143,
259
+ "grad_norm": 0.2993509769439697,
260
+ "learning_rate": 0.00019809800282473013,
261
+ "loss": 0.7553,
262
+ "step": 170
263
+ },
264
+ {
265
+ "epoch": 1.5981735159817352,
266
+ "grad_norm": 0.19971656799316406,
267
+ "learning_rate": 0.0001977746300605507,
268
+ "loss": 0.7591,
269
+ "step": 175
270
+ },
271
+ {
272
+ "epoch": 1.643835616438356,
273
+ "grad_norm": 0.2134828269481659,
274
+ "learning_rate": 0.00019742618935665476,
275
+ "loss": 0.7549,
276
+ "step": 180
277
+ },
278
+ {
279
+ "epoch": 1.6894977168949772,
280
+ "grad_norm": 0.18528909981250763,
281
+ "learning_rate": 0.00019705277004798073,
282
+ "loss": 0.7511,
283
+ "step": 185
284
+ },
285
+ {
286
+ "epoch": 1.7351598173515983,
287
+ "grad_norm": 0.22641536593437195,
288
+ "learning_rate": 0.0001966544678736044,
289
+ "loss": 0.754,
290
+ "step": 190
291
+ },
292
+ {
293
+ "epoch": 1.7808219178082192,
294
+ "grad_norm": 0.23213805258274078,
295
+ "learning_rate": 0.00019623138495219292,
296
+ "loss": 0.7476,
297
+ "step": 195
298
+ },
299
+ {
300
+ "epoch": 1.82648401826484,
301
+ "grad_norm": 0.28454649448394775,
302
+ "learning_rate": 0.00019578362975582292,
303
+ "loss": 0.7535,
304
+ "step": 200
305
+ },
306
+ {
307
+ "epoch": 1.8721461187214612,
308
+ "grad_norm": 0.3238324224948883,
309
+ "learning_rate": 0.00019531131708217005,
310
+ "loss": 0.7489,
311
+ "step": 205
312
+ },
313
+ {
314
+ "epoch": 1.9178082191780823,
315
+ "grad_norm": 0.23858831822872162,
316
+ "learning_rate": 0.0001948145680250766,
317
+ "loss": 0.7514,
318
+ "step": 210
319
+ },
320
+ {
321
+ "epoch": 1.9634703196347032,
322
+ "grad_norm": 0.21854574978351593,
323
+ "learning_rate": 0.00019429350994350483,
324
+ "loss": 0.7464,
325
+ "step": 215
326
+ },
327
+ {
328
+ "epoch": 2.0,
329
+ "eval_loss": 2.1313483715057373,
330
+ "eval_runtime": 0.7223,
331
+ "eval_samples_per_second": 13.845,
332
+ "eval_steps_per_second": 1.384,
333
+ "step": 219
334
+ },
335
+ {
336
+ "epoch": 2.009132420091324,
337
+ "grad_norm": 0.24195519089698792,
338
+ "learning_rate": 0.00019374827642888398,
339
+ "loss": 0.7367,
340
+ "step": 220
341
+ },
342
+ {
343
+ "epoch": 2.0547945205479454,
344
+ "grad_norm": 0.2783088684082031,
345
+ "learning_rate": 0.0001931790072708596,
346
+ "loss": 0.7283,
347
+ "step": 225
348
+ },
349
+ {
350
+ "epoch": 2.1004566210045663,
351
+ "grad_norm": 0.25543445348739624,
352
+ "learning_rate": 0.00019258584842145343,
353
+ "loss": 0.7298,
354
+ "step": 230
355
+ },
356
+ {
357
+ "epoch": 2.146118721461187,
358
+ "grad_norm": 0.23951423168182373,
359
+ "learning_rate": 0.00019196895195764362,
360
+ "loss": 0.7296,
361
+ "step": 235
362
+ },
363
+ {
364
+ "epoch": 2.191780821917808,
365
+ "grad_norm": 0.2517695426940918,
366
+ "learning_rate": 0.0001913284760423745,
367
+ "loss": 0.719,
368
+ "step": 240
369
+ },
370
+ {
371
+ "epoch": 2.237442922374429,
372
+ "grad_norm": 0.22162802517414093,
373
+ "learning_rate": 0.00019066458488400584,
374
+ "loss": 0.7252,
375
+ "step": 245
376
+ },
377
+ {
378
+ "epoch": 2.2831050228310503,
379
+ "grad_norm": 0.2749063968658447,
380
+ "learning_rate": 0.00018997744869421246,
381
+ "loss": 0.7253,
382
+ "step": 250
383
+ },
384
+ {
385
+ "epoch": 2.328767123287671,
386
+ "grad_norm": 0.2638462483882904,
387
+ "learning_rate": 0.00018926724364434446,
388
+ "loss": 0.718,
389
+ "step": 255
390
+ },
391
+ {
392
+ "epoch": 2.374429223744292,
393
+ "grad_norm": 0.23074951767921448,
394
+ "learning_rate": 0.0001885341518202595,
395
+ "loss": 0.7203,
396
+ "step": 260
397
+ },
398
+ {
399
+ "epoch": 2.4200913242009134,
400
+ "grad_norm": 0.18681403994560242,
401
+ "learning_rate": 0.00018777836117563892,
402
+ "loss": 0.7253,
403
+ "step": 265
404
+ },
405
+ {
406
+ "epoch": 2.4657534246575343,
407
+ "grad_norm": 0.19247153401374817,
408
+ "learning_rate": 0.00018700006548379898,
409
+ "loss": 0.7175,
410
+ "step": 270
411
+ },
412
+ {
413
+ "epoch": 2.5114155251141552,
414
+ "grad_norm": 0.18738825619220734,
415
+ "learning_rate": 0.0001861994642880105,
416
+ "loss": 0.7206,
417
+ "step": 275
418
+ },
419
+ {
420
+ "epoch": 2.557077625570776,
421
+ "grad_norm": 0.2461051046848297,
422
+ "learning_rate": 0.00018537676285033887,
423
+ "loss": 0.7134,
424
+ "step": 280
425
+ },
426
+ {
427
+ "epoch": 2.602739726027397,
428
+ "grad_norm": 0.31902387738227844,
429
+ "learning_rate": 0.0001845321720990181,
430
+ "loss": 0.7135,
431
+ "step": 285
432
+ },
433
+ {
434
+ "epoch": 2.6484018264840183,
435
+ "grad_norm": 0.19439548254013062,
436
+ "learning_rate": 0.00018366590857437184,
437
+ "loss": 0.7194,
438
+ "step": 290
439
+ },
440
+ {
441
+ "epoch": 2.6940639269406392,
442
+ "grad_norm": 0.18397387862205505,
443
+ "learning_rate": 0.00018277819437329576,
444
+ "loss": 0.7172,
445
+ "step": 295
446
+ },
447
+ {
448
+ "epoch": 2.73972602739726,
449
+ "grad_norm": 0.1935635656118393,
450
+ "learning_rate": 0.00018186925709231532,
451
+ "loss": 0.7188,
452
+ "step": 300
453
+ },
454
+ {
455
+ "epoch": 2.7853881278538815,
456
+ "grad_norm": 0.20888376235961914,
457
+ "learning_rate": 0.0001809393297692334,
458
+ "loss": 0.71,
459
+ "step": 305
460
+ },
461
+ {
462
+ "epoch": 2.8310502283105023,
463
+ "grad_norm": 0.20157787203788757,
464
+ "learning_rate": 0.0001799886508233829,
465
+ "loss": 0.7152,
466
+ "step": 310
467
+ },
468
+ {
469
+ "epoch": 2.8767123287671232,
470
+ "grad_norm": 0.201010599732399,
471
+ "learning_rate": 0.0001790174639944997,
472
+ "loss": 0.7147,
473
+ "step": 315
474
+ },
475
+ {
476
+ "epoch": 2.922374429223744,
477
+ "grad_norm": 0.20877932012081146,
478
+ "learning_rate": 0.00017802601828023138,
479
+ "loss": 0.7101,
480
+ "step": 320
481
+ },
482
+ {
483
+ "epoch": 2.968036529680365,
484
+ "grad_norm": 0.19995956122875214,
485
+ "learning_rate": 0.00017701456787229804,
486
+ "loss": 0.7128,
487
+ "step": 325
488
+ },
489
+ {
490
+ "epoch": 2.9954337899543377,
491
+ "eval_loss": 2.144402027130127,
492
+ "eval_runtime": 0.7303,
493
+ "eval_samples_per_second": 13.694,
494
+ "eval_steps_per_second": 1.369,
495
+ "step": 328
496
+ },
497
+ {
498
+ "epoch": 3.0136986301369864,
499
+ "grad_norm": 0.18146736919879913,
500
+ "learning_rate": 0.0001759833720913214,
501
+ "loss": 0.7071,
502
+ "step": 330
503
+ },
504
+ {
505
+ "epoch": 3.0593607305936072,
506
+ "grad_norm": 0.20878320932388306,
507
+ "learning_rate": 0.00017493269532033883,
508
+ "loss": 0.6943,
509
+ "step": 335
510
+ },
511
+ {
512
+ "epoch": 3.105022831050228,
513
+ "grad_norm": 0.24599188566207886,
514
+ "learning_rate": 0.0001738628069370195,
515
+ "loss": 0.6968,
516
+ "step": 340
517
+ },
518
+ {
519
+ "epoch": 3.1506849315068495,
520
+ "grad_norm": 0.22413116693496704,
521
+ "learning_rate": 0.00017277398124460023,
522
+ "loss": 0.6949,
523
+ "step": 345
524
+ },
525
+ {
526
+ "epoch": 3.1963470319634704,
527
+ "grad_norm": 0.22311843931674957,
528
+ "learning_rate": 0.000171666497401558,
529
+ "loss": 0.6988,
530
+ "step": 350
531
+ },
532
+ {
533
+ "epoch": 3.2420091324200913,
534
+ "grad_norm": 0.19590473175048828,
535
+ "learning_rate": 0.0001705406393500381,
536
+ "loss": 0.692,
537
+ "step": 355
538
+ },
539
+ {
540
+ "epoch": 3.287671232876712,
541
+ "grad_norm": 0.18890812993049622,
542
+ "learning_rate": 0.00016939669574305566,
543
+ "loss": 0.6923,
544
+ "step": 360
545
+ },
546
+ {
547
+ "epoch": 3.3333333333333335,
548
+ "grad_norm": 0.19763173162937164,
549
+ "learning_rate": 0.0001682349598704892,
550
+ "loss": 0.6899,
551
+ "step": 365
552
+ },
553
+ {
554
+ "epoch": 3.3789954337899544,
555
+ "grad_norm": 0.21059083938598633,
556
+ "learning_rate": 0.00016705572958388576,
557
+ "loss": 0.689,
558
+ "step": 370
559
+ },
560
+ {
561
+ "epoch": 3.4246575342465753,
562
+ "grad_norm": 0.2090144008398056,
563
+ "learning_rate": 0.00016585930722009601,
564
+ "loss": 0.6909,
565
+ "step": 375
566
+ },
567
+ {
568
+ "epoch": 3.470319634703196,
569
+ "grad_norm": 0.1937251091003418,
570
+ "learning_rate": 0.00016464599952375998,
571
+ "loss": 0.6943,
572
+ "step": 380
573
+ },
574
+ {
575
+ "epoch": 3.5159817351598175,
576
+ "grad_norm": 0.19087150692939758,
577
+ "learning_rate": 0.000163416117568662,
578
+ "loss": 0.6936,
579
+ "step": 385
580
+ },
581
+ {
582
+ "epoch": 3.5616438356164384,
583
+ "grad_norm": 0.21907366812229156,
584
+ "learning_rate": 0.0001621699766779763,
585
+ "loss": 0.6928,
586
+ "step": 390
587
+ },
588
+ {
589
+ "epoch": 3.6073059360730593,
590
+ "grad_norm": 0.1987326741218567,
591
+ "learning_rate": 0.00016090789634342278,
592
+ "loss": 0.6913,
593
+ "step": 395
594
+ },
595
+ {
596
+ "epoch": 3.65296803652968,
597
+ "grad_norm": 0.19414757192134857,
598
+ "learning_rate": 0.00015963020014335438,
599
+ "loss": 0.6873,
600
+ "step": 400
601
+ },
602
+ {
603
+ "epoch": 3.6986301369863015,
604
+ "grad_norm": 0.19000251591205597,
605
+ "learning_rate": 0.0001583372156597961,
606
+ "loss": 0.6895,
607
+ "step": 405
608
+ },
609
+ {
610
+ "epoch": 3.7442922374429224,
611
+ "grad_norm": 0.17563720047473907,
612
+ "learning_rate": 0.00015702927439445826,
613
+ "loss": 0.6905,
614
+ "step": 410
615
+ },
616
+ {
617
+ "epoch": 3.7899543378995433,
618
+ "grad_norm": 0.22648802399635315,
619
+ "learning_rate": 0.00015570671168374438,
620
+ "loss": 0.685,
621
+ "step": 415
622
+ },
623
+ {
624
+ "epoch": 3.8356164383561646,
625
+ "grad_norm": 0.1822802722454071,
626
+ "learning_rate": 0.00015436986661277577,
627
+ "loss": 0.6897,
628
+ "step": 420
629
+ },
630
+ {
631
+ "epoch": 3.8812785388127855,
632
+ "grad_norm": 0.18406091630458832,
633
+ "learning_rate": 0.0001530190819284555,
634
+ "loss": 0.6849,
635
+ "step": 425
636
+ },
637
+ {
638
+ "epoch": 3.9269406392694064,
639
+ "grad_norm": 0.1968332827091217,
640
+ "learning_rate": 0.00015165470395159313,
641
+ "loss": 0.6841,
642
+ "step": 430
643
+ },
644
+ {
645
+ "epoch": 3.9726027397260273,
646
+ "grad_norm": 0.2520104646682739,
647
+ "learning_rate": 0.0001502770824881133,
648
+ "loss": 0.6924,
649
+ "step": 435
650
+ },
651
+ {
652
+ "epoch": 4.0,
653
+ "eval_loss": 2.1630699634552,
654
+ "eval_runtime": 0.7164,
655
+ "eval_samples_per_second": 13.958,
656
+ "eval_steps_per_second": 1.396,
657
+ "step": 438
658
+ },
659
+ {
660
+ "epoch": 4.018264840182648,
661
+ "grad_norm": 0.20841118693351746,
662
+ "learning_rate": 0.00014888657073937076,
663
+ "loss": 0.6786,
664
+ "step": 440
665
+ },
666
+ {
667
+ "epoch": 4.063926940639269,
668
+ "grad_norm": 0.2335735559463501,
669
+ "learning_rate": 0.00014748352521159493,
670
+ "loss": 0.6687,
671
+ "step": 445
672
+ },
673
+ {
674
+ "epoch": 4.109589041095891,
675
+ "grad_norm": 0.23169922828674316,
676
+ "learning_rate": 0.0001460683056244869,
677
+ "loss": 0.6701,
678
+ "step": 450
679
+ },
680
+ {
681
+ "epoch": 4.155251141552512,
682
+ "grad_norm": 0.20742633938789368,
683
+ "learning_rate": 0.00014464127481899312,
684
+ "loss": 0.6678,
685
+ "step": 455
686
+ },
687
+ {
688
+ "epoch": 4.200913242009133,
689
+ "grad_norm": 0.233662411570549,
690
+ "learning_rate": 0.00014320279866427796,
691
+ "loss": 0.669,
692
+ "step": 460
693
+ },
694
+ {
695
+ "epoch": 4.2465753424657535,
696
+ "grad_norm": 0.22496378421783447,
697
+ "learning_rate": 0.00014175324596392075,
698
+ "loss": 0.6695,
699
+ "step": 465
700
+ },
701
+ {
702
+ "epoch": 4.292237442922374,
703
+ "grad_norm": 0.1961798071861267,
704
+ "learning_rate": 0.00014029298836135988,
705
+ "loss": 0.6694,
706
+ "step": 470
707
+ },
708
+ {
709
+ "epoch": 4.337899543378995,
710
+ "grad_norm": 0.21112968027591705,
711
+ "learning_rate": 0.00013882240024460927,
712
+ "loss": 0.6758,
713
+ "step": 475
714
+ },
715
+ {
716
+ "epoch": 4.383561643835616,
717
+ "grad_norm": 0.21480615437030792,
718
+ "learning_rate": 0.0001373418586502706,
719
+ "loss": 0.6685,
720
+ "step": 480
721
+ },
722
+ {
723
+ "epoch": 4.429223744292237,
724
+ "grad_norm": 0.24116870760917664,
725
+ "learning_rate": 0.0001358517431668672,
726
+ "loss": 0.6686,
727
+ "step": 485
728
+ },
729
+ {
730
+ "epoch": 4.474885844748858,
731
+ "grad_norm": 0.20268571376800537,
732
+ "learning_rate": 0.00013435243583752294,
733
+ "loss": 0.6642,
734
+ "step": 490
735
+ },
736
+ {
737
+ "epoch": 4.52054794520548,
738
+ "grad_norm": 0.19283969700336456,
739
+ "learning_rate": 0.00013284432106201233,
740
+ "loss": 0.6729,
741
+ "step": 495
742
+ },
743
+ {
744
+ "epoch": 4.566210045662101,
745
+ "grad_norm": 0.18919415771961212,
746
+ "learning_rate": 0.00013132778549820618,
747
+ "loss": 0.6713,
748
+ "step": 500
749
+ },
750
+ {
751
+ "epoch": 4.6118721461187215,
752
+ "grad_norm": 0.19388696551322937,
753
+ "learning_rate": 0.00012980321796293836,
754
+ "loss": 0.6698,
755
+ "step": 505
756
+ },
757
+ {
758
+ "epoch": 4.657534246575342,
759
+ "grad_norm": 0.19509977102279663,
760
+ "learning_rate": 0.00012827100933231905,
761
+ "loss": 0.6732,
762
+ "step": 510
763
+ },
764
+ {
765
+ "epoch": 4.703196347031963,
766
+ "grad_norm": 0.19472207129001617,
767
+ "learning_rate": 0.00012673155244151985,
768
+ "loss": 0.6689,
769
+ "step": 515
770
+ },
771
+ {
772
+ "epoch": 4.748858447488584,
773
+ "grad_norm": 0.1974049210548401,
774
+ "learning_rate": 0.000125185241984057,
775
+ "loss": 0.6661,
776
+ "step": 520
777
+ },
778
+ {
779
+ "epoch": 4.794520547945205,
780
+ "grad_norm": 0.18362760543823242,
781
+ "learning_rate": 0.00012363247441059776,
782
+ "loss": 0.6705,
783
+ "step": 525
784
+ },
785
+ {
786
+ "epoch": 4.840182648401827,
787
+ "grad_norm": 0.1863856315612793,
788
+ "learning_rate": 0.00012207364782731655,
789
+ "loss": 0.6663,
790
+ "step": 530
791
+ },
792
+ {
793
+ "epoch": 4.885844748858448,
794
+ "grad_norm": 0.19237390160560608,
795
+ "learning_rate": 0.00012050916189382646,
796
+ "loss": 0.6701,
797
+ "step": 535
798
+ },
799
+ {
800
+ "epoch": 4.931506849315069,
801
+ "grad_norm": 0.1915617734193802,
802
+ "learning_rate": 0.00011893941772071249,
803
+ "loss": 0.6781,
804
+ "step": 540
805
+ },
806
+ {
807
+ "epoch": 4.9771689497716896,
808
+ "grad_norm": 0.18157874047756195,
809
+ "learning_rate": 0.00011736481776669306,
810
+ "loss": 0.6777,
811
+ "step": 545
812
+ },
813
+ {
814
+ "epoch": 4.995433789954338,
815
+ "eval_loss": 2.182265520095825,
816
+ "eval_runtime": 0.7269,
817
+ "eval_samples_per_second": 13.757,
818
+ "eval_steps_per_second": 1.376,
819
+ "step": 547
820
+ },
821
+ {
822
+ "epoch": 5.0228310502283104,
823
+ "grad_norm": 0.19536516070365906,
824
+ "learning_rate": 0.0001157857657354354,
825
+ "loss": 0.6657,
826
+ "step": 550
827
+ },
828
+ {
829
+ "epoch": 5.068493150684931,
830
+ "grad_norm": 0.22191356122493744,
831
+ "learning_rate": 0.00011420266647205231,
832
+ "loss": 0.6547,
833
+ "step": 555
834
+ },
835
+ {
836
+ "epoch": 5.114155251141552,
837
+ "grad_norm": 0.2473655790090561,
838
+ "learning_rate": 0.00011261592585930576,
839
+ "loss": 0.649,
840
+ "step": 560
841
+ },
842
+ {
843
+ "epoch": 5.159817351598173,
844
+ "grad_norm": 0.21013621985912323,
845
+ "learning_rate": 0.00011102595071354472,
846
+ "loss": 0.6493,
847
+ "step": 565
848
+ },
849
+ {
850
+ "epoch": 5.205479452054795,
851
+ "grad_norm": 0.2081068456172943,
852
+ "learning_rate": 0.00010943314868040364,
853
+ "loss": 0.647,
854
+ "step": 570
855
+ },
856
+ {
857
+ "epoch": 5.251141552511416,
858
+ "grad_norm": 0.23985308408737183,
859
+ "learning_rate": 0.00010783792813028827,
860
+ "loss": 0.6505,
861
+ "step": 575
862
+ },
863
+ {
864
+ "epoch": 5.296803652968037,
865
+ "grad_norm": 0.19800134003162384,
866
+ "learning_rate": 0.00010624069805367559,
867
+ "loss": 0.6537,
868
+ "step": 580
869
+ },
870
+ {
871
+ "epoch": 5.342465753424658,
872
+ "grad_norm": 0.21391679346561432,
873
+ "learning_rate": 0.00010464186795625482,
874
+ "loss": 0.6558,
875
+ "step": 585
876
+ },
877
+ {
878
+ "epoch": 5.3881278538812785,
879
+ "grad_norm": 0.205993190407753,
880
+ "learning_rate": 0.00010304184775393642,
881
+ "loss": 0.6578,
882
+ "step": 590
883
+ },
884
+ {
885
+ "epoch": 5.433789954337899,
886
+ "grad_norm": 0.20510774850845337,
887
+ "learning_rate": 0.00010144104766775572,
888
+ "loss": 0.6565,
889
+ "step": 595
890
+ },
891
+ {
892
+ "epoch": 5.47945205479452,
893
+ "grad_norm": 0.1970670372247696,
894
+ "learning_rate": 9.983987811869862e-05,
895
+ "loss": 0.6515,
896
+ "step": 600
897
+ },
898
+ {
899
+ "epoch": 5.525114155251142,
900
+ "grad_norm": 0.18836742639541626,
901
+ "learning_rate": 9.823874962247564e-05,
902
+ "loss": 0.6557,
903
+ "step": 605
904
+ },
905
+ {
906
+ "epoch": 5.570776255707763,
907
+ "grad_norm": 0.20027703046798706,
908
+ "learning_rate": 9.663807268427198e-05,
909
+ "loss": 0.6522,
910
+ "step": 610
911
+ },
912
+ {
913
+ "epoch": 5.616438356164384,
914
+ "grad_norm": 0.2243603765964508,
915
+ "learning_rate": 9.503825769350017e-05,
916
+ "loss": 0.6496,
917
+ "step": 615
918
+ },
919
+ {
920
+ "epoch": 5.662100456621005,
921
+ "grad_norm": 0.18932883441448212,
922
+ "learning_rate": 9.343971481858246e-05,
923
+ "loss": 0.6532,
924
+ "step": 620
925
+ },
926
+ {
927
+ "epoch": 5.707762557077626,
928
+ "grad_norm": 0.19952723383903503,
929
+ "learning_rate": 9.184285390178978e-05,
930
+ "loss": 0.6511,
931
+ "step": 625
932
+ },
933
+ {
934
+ "epoch": 5.7534246575342465,
935
+ "grad_norm": 0.20302745699882507,
936
+ "learning_rate": 9.024808435416434e-05,
937
+ "loss": 0.6524,
938
+ "step": 630
939
+ },
940
+ {
941
+ "epoch": 5.799086757990867,
942
+ "grad_norm": 0.1958286166191101,
943
+ "learning_rate": 8.865581505055291e-05,
944
+ "loss": 0.6543,
945
+ "step": 635
946
+ },
947
+ {
948
+ "epoch": 5.844748858447488,
949
+ "grad_norm": 0.189736470580101,
950
+ "learning_rate": 8.706645422477739e-05,
951
+ "loss": 0.6529,
952
+ "step": 640
953
+ },
954
+ {
955
+ "epoch": 5.890410958904109,
956
+ "grad_norm": 0.2069847285747528,
957
+ "learning_rate": 8.548040936496989e-05,
958
+ "loss": 0.6542,
959
+ "step": 645
960
+ },
961
+ {
962
+ "epoch": 5.936073059360731,
963
+ "grad_norm": 0.2031773030757904,
964
+ "learning_rate": 8.389808710909881e-05,
965
+ "loss": 0.6515,
966
+ "step": 650
967
+ },
968
+ {
969
+ "epoch": 5.981735159817352,
970
+ "grad_norm": 0.20669801533222198,
971
+ "learning_rate": 8.231989314071317e-05,
972
+ "loss": 0.6526,
973
+ "step": 655
974
+ },
975
+ {
976
+ "epoch": 6.0,
977
+ "eval_loss": 2.207833766937256,
978
+ "eval_runtime": 0.7173,
979
+ "eval_samples_per_second": 13.941,
980
+ "eval_steps_per_second": 1.394,
981
+ "step": 657
982
+ },
983
+ {
984
+ "epoch": 6.027397260273973,
985
+ "grad_norm": 0.2104310542345047,
986
+ "learning_rate": 8.07462320849313e-05,
987
+ "loss": 0.6444,
988
+ "step": 660
989
+ },
990
+ {
991
+ "epoch": 6.073059360730594,
992
+ "grad_norm": 0.18349581956863403,
993
+ "learning_rate": 7.917750740470117e-05,
994
+ "loss": 0.6397,
995
+ "step": 665
996
+ },
997
+ {
998
+ "epoch": 6.1187214611872145,
999
+ "grad_norm": 0.20204049348831177,
1000
+ "learning_rate": 7.761412129735852e-05,
1001
+ "loss": 0.6439,
1002
+ "step": 670
1003
+ },
1004
+ {
1005
+ "epoch": 6.164383561643835,
1006
+ "grad_norm": 0.20316743850708008,
1007
+ "learning_rate": 7.605647459150961e-05,
1008
+ "loss": 0.6392,
1009
+ "step": 675
1010
+ },
1011
+ {
1012
+ "epoch": 6.210045662100456,
1013
+ "grad_norm": 0.2150379866361618,
1014
+ "learning_rate": 7.450496664426477e-05,
1015
+ "loss": 0.634,
1016
+ "step": 680
1017
+ },
1018
+ {
1019
+ "epoch": 6.255707762557078,
1020
+ "grad_norm": 0.19887958467006683,
1021
+ "learning_rate": 7.295999523884921e-05,
1022
+ "loss": 0.6442,
1023
+ "step": 685
1024
+ },
1025
+ {
1026
+ "epoch": 6.301369863013699,
1027
+ "grad_norm": 0.1974460333585739,
1028
+ "learning_rate": 7.142195648261747e-05,
1029
+ "loss": 0.6396,
1030
+ "step": 690
1031
+ },
1032
+ {
1033
+ "epoch": 6.34703196347032,
1034
+ "grad_norm": 0.21475858986377716,
1035
+ "learning_rate": 6.989124470549745e-05,
1036
+ "loss": 0.6354,
1037
+ "step": 695
1038
+ },
1039
+ {
1040
+ "epoch": 6.392694063926941,
1041
+ "grad_norm": 0.19873230159282684,
1042
+ "learning_rate": 6.83682523588902e-05,
1043
+ "loss": 0.64,
1044
+ "step": 700
1045
+ },
1046
+ {
1047
+ "epoch": 6.438356164383562,
1048
+ "grad_norm": 0.20658674836158752,
1049
+ "learning_rate": 6.685336991505122e-05,
1050
+ "loss": 0.6367,
1051
+ "step": 705
1052
+ },
1053
+ {
1054
+ "epoch": 6.4840182648401825,
1055
+ "grad_norm": 0.2072938233613968,
1056
+ "learning_rate": 6.534698576697939e-05,
1057
+ "loss": 0.6388,
1058
+ "step": 710
1059
+ },
1060
+ {
1061
+ "epoch": 6.529680365296803,
1062
+ "grad_norm": 0.20482133328914642,
1063
+ "learning_rate": 6.384948612883873e-05,
1064
+ "loss": 0.6397,
1065
+ "step": 715
1066
+ },
1067
+ {
1068
+ "epoch": 6.575342465753424,
1069
+ "grad_norm": 0.20856335759162903,
1070
+ "learning_rate": 6.2361254936939e-05,
1071
+ "loss": 0.6407,
1072
+ "step": 720
1073
+ },
1074
+ {
1075
+ "epoch": 6.621004566210045,
1076
+ "grad_norm": 0.21489538252353668,
1077
+ "learning_rate": 6.088267375130023e-05,
1078
+ "loss": 0.6414,
1079
+ "step": 725
1080
+ },
1081
+ {
1082
+ "epoch": 6.666666666666667,
1083
+ "grad_norm": 0.19792255759239197,
1084
+ "learning_rate": 5.941412165782645e-05,
1085
+ "loss": 0.634,
1086
+ "step": 730
1087
+ },
1088
+ {
1089
+ "epoch": 6.712328767123288,
1090
+ "grad_norm": 0.19678843021392822,
1091
+ "learning_rate": 5.79559751711138e-05,
1092
+ "loss": 0.6407,
1093
+ "step": 735
1094
+ },
1095
+ {
1096
+ "epoch": 6.757990867579909,
1097
+ "grad_norm": 0.18998436629772186,
1098
+ "learning_rate": 5.650860813791785e-05,
1099
+ "loss": 0.6346,
1100
+ "step": 740
1101
+ },
1102
+ {
1103
+ "epoch": 6.80365296803653,
1104
+ "grad_norm": 0.19847027957439423,
1105
+ "learning_rate": 5.507239164130501e-05,
1106
+ "loss": 0.6409,
1107
+ "step": 745
1108
+ },
1109
+ {
1110
+ "epoch": 6.8493150684931505,
1111
+ "grad_norm": 0.205197274684906,
1112
+ "learning_rate": 5.364769390551225e-05,
1113
+ "loss": 0.6404,
1114
+ "step": 750
1115
+ },
1116
+ {
1117
+ "epoch": 6.894977168949771,
1118
+ "grad_norm": 0.19521455466747284,
1119
+ "learning_rate": 5.2234880201540284e-05,
1120
+ "loss": 0.6386,
1121
+ "step": 755
1122
+ },
1123
+ {
1124
+ "epoch": 6.940639269406392,
1125
+ "grad_norm": 0.1955854445695877,
1126
+ "learning_rate": 5.0834312753503124e-05,
1127
+ "loss": 0.6349,
1128
+ "step": 760
1129
+ },
1130
+ {
1131
+ "epoch": 6.986301369863014,
1132
+ "grad_norm": 0.1964850276708603,
1133
+ "learning_rate": 4.9446350645759885e-05,
1134
+ "loss": 0.6326,
1135
+ "step": 765
1136
+ },
1137
+ {
1138
+ "epoch": 6.995433789954338,
1139
+ "eval_loss": 2.229551315307617,
1140
+ "eval_runtime": 0.7351,
1141
+ "eval_samples_per_second": 13.604,
1142
+ "eval_steps_per_second": 1.36,
1143
+ "step": 766
1144
+ },
1145
+ {
1146
+ "epoch": 7.031963470319635,
1147
+ "grad_norm": 0.1950986385345459,
1148
+ "learning_rate": 4.807134973085036e-05,
1149
+ "loss": 0.626,
1150
+ "step": 770
1151
+ },
1152
+ {
1153
+ "epoch": 7.077625570776256,
1154
+ "grad_norm": 0.20411163568496704,
1155
+ "learning_rate": 4.6709662538260267e-05,
1156
+ "loss": 0.6293,
1157
+ "step": 775
1158
+ },
1159
+ {
1160
+ "epoch": 7.123287671232877,
1161
+ "grad_norm": 0.21672751009464264,
1162
+ "learning_rate": 4.53616381840377e-05,
1163
+ "loss": 0.6255,
1164
+ "step": 780
1165
+ },
1166
+ {
1167
+ "epoch": 7.168949771689498,
1168
+ "grad_norm": 0.20474384725093842,
1169
+ "learning_rate": 4.402762228128531e-05,
1170
+ "loss": 0.6317,
1171
+ "step": 785
1172
+ },
1173
+ {
1174
+ "epoch": 7.2146118721461185,
1175
+ "grad_norm": 0.20330122113227844,
1176
+ "learning_rate": 4.2707956851550016e-05,
1177
+ "loss": 0.6297,
1178
+ "step": 790
1179
+ },
1180
+ {
1181
+ "epoch": 7.260273972602739,
1182
+ "grad_norm": 0.21286101639270782,
1183
+ "learning_rate": 4.140298023713416e-05,
1184
+ "loss": 0.6278,
1185
+ "step": 795
1186
+ },
1187
+ {
1188
+ "epoch": 7.30593607305936,
1189
+ "grad_norm": 0.20241223275661469,
1190
+ "learning_rate": 4.011302701434937e-05,
1191
+ "loss": 0.6223,
1192
+ "step": 800
1193
+ },
1194
+ {
1195
+ "epoch": 7.351598173515982,
1196
+ "grad_norm": 0.20820164680480957,
1197
+ "learning_rate": 3.8838427907736476e-05,
1198
+ "loss": 0.631,
1199
+ "step": 805
1200
+ },
1201
+ {
1202
+ "epoch": 7.397260273972603,
1203
+ "grad_norm": 0.2071073353290558,
1204
+ "learning_rate": 3.757950970527249e-05,
1205
+ "loss": 0.627,
1206
+ "step": 810
1207
+ },
1208
+ {
1209
+ "epoch": 7.442922374429224,
1210
+ "grad_norm": 0.20754040777683258,
1211
+ "learning_rate": 3.633659517458736e-05,
1212
+ "loss": 0.6284,
1213
+ "step": 815
1214
+ },
1215
+ {
1216
+ "epoch": 7.488584474885845,
1217
+ "grad_norm": 0.204667329788208,
1218
+ "learning_rate": 3.5110002980210975e-05,
1219
+ "loss": 0.6214,
1220
+ "step": 820
1221
+ },
1222
+ {
1223
+ "epoch": 7.534246575342466,
1224
+ "grad_norm": 0.204596146941185,
1225
+ "learning_rate": 3.3900047601872596e-05,
1226
+ "loss": 0.6241,
1227
+ "step": 825
1228
+ },
1229
+ {
1230
+ "epoch": 7.579908675799087,
1231
+ "grad_norm": 0.2120332568883896,
1232
+ "learning_rate": 3.270703925387279e-05,
1233
+ "loss": 0.6295,
1234
+ "step": 830
1235
+ },
1236
+ {
1237
+ "epoch": 7.6255707762557075,
1238
+ "grad_norm": 0.20006106793880463,
1239
+ "learning_rate": 3.153128380554941e-05,
1240
+ "loss": 0.6259,
1241
+ "step": 835
1242
+ },
1243
+ {
1244
+ "epoch": 7.671232876712329,
1245
+ "grad_norm": 0.19608016312122345,
1246
+ "learning_rate": 3.037308270285709e-05,
1247
+ "loss": 0.6234,
1248
+ "step": 840
1249
+ },
1250
+ {
1251
+ "epoch": 7.71689497716895,
1252
+ "grad_norm": 0.20698235929012299,
1253
+ "learning_rate": 2.923273289108115e-05,
1254
+ "loss": 0.6312,
1255
+ "step": 845
1256
+ },
1257
+ {
1258
+ "epoch": 7.762557077625571,
1259
+ "grad_norm": 0.20038080215454102,
1260
+ "learning_rate": 2.8110526738705344e-05,
1261
+ "loss": 0.6266,
1262
+ "step": 850
1263
+ },
1264
+ {
1265
+ "epoch": 7.808219178082192,
1266
+ "grad_norm": 0.2022303193807602,
1267
+ "learning_rate": 2.7006751962452882e-05,
1268
+ "loss": 0.631,
1269
+ "step": 855
1270
+ },
1271
+ {
1272
+ "epoch": 7.853881278538813,
1273
+ "grad_norm": 0.20605237782001495,
1274
+ "learning_rate": 2.592169155352031e-05,
1275
+ "loss": 0.6331,
1276
+ "step": 860
1277
+ },
1278
+ {
1279
+ "epoch": 7.899543378995434,
1280
+ "grad_norm": 0.2080409824848175,
1281
+ "learning_rate": 2.485562370502279e-05,
1282
+ "loss": 0.6313,
1283
+ "step": 865
1284
+ },
1285
+ {
1286
+ "epoch": 7.945205479452055,
1287
+ "grad_norm": 0.2000366896390915,
1288
+ "learning_rate": 2.3808821740669606e-05,
1289
+ "loss": 0.6285,
1290
+ "step": 870
1291
+ },
1292
+ {
1293
+ "epoch": 7.9908675799086755,
1294
+ "grad_norm": 0.20418736338615417,
1295
+ "learning_rate": 2.2781554044688015e-05,
1296
+ "loss": 0.6311,
1297
+ "step": 875
1298
+ },
1299
+ {
1300
+ "epoch": 8.0,
1301
+ "eval_loss": 2.2484524250030518,
1302
+ "eval_runtime": 0.7257,
1303
+ "eval_samples_per_second": 13.779,
1304
+ "eval_steps_per_second": 1.378,
1305
+ "step": 876
1306
+ },
1307
+ {
1308
+ "epoch": 8.036529680365296,
1309
+ "grad_norm": 0.2047678381204605,
1310
+ "learning_rate": 2.1774083993013718e-05,
1311
+ "loss": 0.6168,
1312
+ "step": 880
1313
+ },
1314
+ {
1315
+ "epoch": 8.082191780821917,
1316
+ "grad_norm": 0.20055288076400757,
1317
+ "learning_rate": 2.078666988576504e-05,
1318
+ "loss": 0.6163,
1319
+ "step": 885
1320
+ },
1321
+ {
1322
+ "epoch": 8.127853881278538,
1323
+ "grad_norm": 0.2002391219139099,
1324
+ "learning_rate": 1.9819564881018983e-05,
1325
+ "loss": 0.6216,
1326
+ "step": 890
1327
+ },
1328
+ {
1329
+ "epoch": 8.173515981735159,
1330
+ "grad_norm": 0.19824841618537903,
1331
+ "learning_rate": 1.887301692990494e-05,
1332
+ "loss": 0.6215,
1333
+ "step": 895
1334
+ },
1335
+ {
1336
+ "epoch": 8.219178082191782,
1337
+ "grad_norm": 0.20471996068954468,
1338
+ "learning_rate": 1.7947268713034127e-05,
1339
+ "loss": 0.6168,
1340
+ "step": 900
1341
+ },
1342
+ {
1343
+ "epoch": 8.264840182648403,
1344
+ "grad_norm": 0.20544229447841644,
1345
+ "learning_rate": 1.7042557578279626e-05,
1346
+ "loss": 0.6188,
1347
+ "step": 905
1348
+ },
1349
+ {
1350
+ "epoch": 8.310502283105023,
1351
+ "grad_norm": 0.208919957280159,
1352
+ "learning_rate": 1.6159115479924257e-05,
1353
+ "loss": 0.6221,
1354
+ "step": 910
1355
+ },
1356
+ {
1357
+ "epoch": 8.356164383561644,
1358
+ "grad_norm": 0.20433735847473145,
1359
+ "learning_rate": 1.529716891919074e-05,
1360
+ "loss": 0.6187,
1361
+ "step": 915
1362
+ },
1363
+ {
1364
+ "epoch": 8.401826484018265,
1365
+ "grad_norm": 0.20233039557933807,
1366
+ "learning_rate": 1.4456938886170412e-05,
1367
+ "loss": 0.6224,
1368
+ "step": 920
1369
+ },
1370
+ {
1371
+ "epoch": 8.447488584474886,
1372
+ "grad_norm": 0.2014380395412445,
1373
+ "learning_rate": 1.3638640803164516e-05,
1374
+ "loss": 0.6186,
1375
+ "step": 925
1376
+ },
1377
+ {
1378
+ "epoch": 8.493150684931507,
1379
+ "grad_norm": 0.20106054842472076,
1380
+ "learning_rate": 1.2842484469453365e-05,
1381
+ "loss": 0.6229,
1382
+ "step": 930
1383
+ },
1384
+ {
1385
+ "epoch": 8.538812785388128,
1386
+ "grad_norm": 0.20424200594425201,
1387
+ "learning_rate": 1.2068674007506786e-05,
1388
+ "loss": 0.6225,
1389
+ "step": 935
1390
+ },
1391
+ {
1392
+ "epoch": 8.584474885844749,
1393
+ "grad_norm": 0.19919848442077637,
1394
+ "learning_rate": 1.1317407810650372e-05,
1395
+ "loss": 0.621,
1396
+ "step": 940
1397
+ },
1398
+ {
1399
+ "epoch": 8.63013698630137,
1400
+ "grad_norm": 0.21022921800613403,
1401
+ "learning_rate": 1.058887849220026e-05,
1402
+ "loss": 0.6208,
1403
+ "step": 945
1404
+ },
1405
+ {
1406
+ "epoch": 8.67579908675799,
1407
+ "grad_norm": 0.2036212980747223,
1408
+ "learning_rate": 9.883272836080116e-06,
1409
+ "loss": 0.6184,
1410
+ "step": 950
1411
+ },
1412
+ {
1413
+ "epoch": 8.721461187214611,
1414
+ "grad_norm": 0.19858643412590027,
1415
+ "learning_rate": 9.200771748932513e-06,
1416
+ "loss": 0.6224,
1417
+ "step": 955
1418
+ },
1419
+ {
1420
+ "epoch": 8.767123287671232,
1421
+ "grad_norm": 0.2013389766216278,
1422
+ "learning_rate": 8.541550213737171e-06,
1423
+ "loss": 0.6175,
1424
+ "step": 960
1425
+ },
1426
+ {
1427
+ "epoch": 8.812785388127853,
1428
+ "grad_norm": 0.20145344734191895,
1429
+ "learning_rate": 7.905777244947954e-06,
1430
+ "loss": 0.6204,
1431
+ "step": 965
1432
+ },
1433
+ {
1434
+ "epoch": 8.858447488584474,
1435
+ "grad_norm": 0.20189128816127777,
1436
+ "learning_rate": 7.293615845160196e-06,
1437
+ "loss": 0.6225,
1438
+ "step": 970
1439
+ },
1440
+ {
1441
+ "epoch": 8.904109589041095,
1442
+ "grad_norm": 0.2002427577972412,
1443
+ "learning_rate": 6.705222963319191e-06,
1444
+ "loss": 0.6197,
1445
+ "step": 975
1446
+ },
1447
+ {
1448
+ "epoch": 8.949771689497716,
1449
+ "grad_norm": 0.21808940172195435,
1450
+ "learning_rate": 6.140749454480932e-06,
1451
+ "loss": 0.6228,
1452
+ "step": 980
1453
+ },
1454
+ {
1455
+ "epoch": 8.995433789954339,
1456
+ "grad_norm": 0.20092381536960602,
1457
+ "learning_rate": 5.6003400411351325e-06,
1458
+ "loss": 0.6233,
1459
+ "step": 985
1460
+ },
1461
+ {
1462
+ "epoch": 8.995433789954339,
1463
+ "eval_loss": 2.258711099624634,
1464
+ "eval_runtime": 0.7286,
1465
+ "eval_samples_per_second": 13.726,
1466
+ "eval_steps_per_second": 1.373,
1467
+ "step": 985
1468
+ },
1469
+ {
1470
+ "epoch": 9.04109589041096,
1471
+ "grad_norm": 0.19878605008125305,
1472
+ "learning_rate": 5.0841332761005e-06,
1473
+ "loss": 0.6193,
1474
+ "step": 990
1475
+ },
1476
+ {
1477
+ "epoch": 9.08675799086758,
1478
+ "grad_norm": 0.20306305587291718,
1479
+ "learning_rate": 4.592261507001993e-06,
1480
+ "loss": 0.6177,
1481
+ "step": 995
1482
+ },
1483
+ {
1484
+ "epoch": 9.132420091324201,
1485
+ "grad_norm": 0.2019263654947281,
1486
+ "learning_rate": 4.124850842338779e-06,
1487
+ "loss": 0.6133,
1488
+ "step": 1000
1489
+ },
1490
+ {
1491
+ "epoch": 9.178082191780822,
1492
+ "grad_norm": 0.2005142867565155,
1493
+ "learning_rate": 3.6820211191520125e-06,
1494
+ "loss": 0.6141,
1495
+ "step": 1005
1496
+ },
1497
+ {
1498
+ "epoch": 9.223744292237443,
1499
+ "grad_norm": 0.20048119127750397,
1500
+ "learning_rate": 3.263885872300343e-06,
1501
+ "loss": 0.6224,
1502
+ "step": 1010
1503
+ },
1504
+ {
1505
+ "epoch": 9.269406392694064,
1506
+ "grad_norm": 0.20118780434131622,
1507
+ "learning_rate": 2.8705523053513816e-06,
1508
+ "loss": 0.6158,
1509
+ "step": 1015
1510
+ },
1511
+ {
1512
+ "epoch": 9.315068493150685,
1513
+ "grad_norm": 0.19994989037513733,
1514
+ "learning_rate": 2.502121263096224e-06,
1515
+ "loss": 0.6133,
1516
+ "step": 1020
1517
+ },
1518
+ {
1519
+ "epoch": 9.360730593607306,
1520
+ "grad_norm": 0.19647841155529022,
1521
+ "learning_rate": 2.1586872056944428e-06,
1522
+ "loss": 0.6172,
1523
+ "step": 1025
1524
+ },
1525
+ {
1526
+ "epoch": 9.406392694063927,
1527
+ "grad_norm": 0.20165683329105377,
1528
+ "learning_rate": 1.840338184455881e-06,
1529
+ "loss": 0.6193,
1530
+ "step": 1030
1531
+ },
1532
+ {
1533
+ "epoch": 9.452054794520548,
1534
+ "grad_norm": 0.20188592374324799,
1535
+ "learning_rate": 1.5471558192656777e-06,
1536
+ "loss": 0.6216,
1537
+ "step": 1035
1538
+ },
1539
+ {
1540
+ "epoch": 9.497716894977168,
1541
+ "grad_norm": 0.20157510042190552,
1542
+ "learning_rate": 1.2792152776580968e-06,
1543
+ "loss": 0.616,
1544
+ "step": 1040
1545
+ },
1546
+ {
1547
+ "epoch": 9.54337899543379,
1548
+ "grad_norm": 0.20351751148700714,
1549
+ "learning_rate": 1.036585255544764e-06,
1550
+ "loss": 0.6149,
1551
+ "step": 1045
1552
+ },
1553
+ {
1554
+ "epoch": 9.58904109589041,
1555
+ "grad_norm": 0.1994985044002533,
1556
+ "learning_rate": 8.193279596020121e-07,
1557
+ "loss": 0.6184,
1558
+ "step": 1050
1559
+ },
1560
+ {
1561
+ "epoch": 9.634703196347033,
1562
+ "grad_norm": 0.2011667788028717,
1563
+ "learning_rate": 6.274990913221035e-07,
1564
+ "loss": 0.6149,
1565
+ "step": 1055
1566
+ },
1567
+ {
1568
+ "epoch": 9.680365296803654,
1569
+ "grad_norm": 0.2020387500524521,
1570
+ "learning_rate": 4.6114783273213393e-07,
1571
+ "loss": 0.619,
1572
+ "step": 1060
1573
+ },
1574
+ {
1575
+ "epoch": 9.726027397260275,
1576
+ "grad_norm": 0.19661925733089447,
1577
+ "learning_rate": 3.203168337845508e-07,
1578
+ "loss": 0.6193,
1579
+ "step": 1065
1580
+ },
1581
+ {
1582
+ "epoch": 9.771689497716896,
1583
+ "grad_norm": 0.20111018419265747,
1584
+ "learning_rate": 2.05042201422323e-07,
1585
+ "loss": 0.615,
1586
+ "step": 1070
1587
+ },
1588
+ {
1589
+ "epoch": 9.817351598173516,
1590
+ "grad_norm": 0.20135089755058289,
1591
+ "learning_rate": 1.1535349032167908e-07,
1592
+ "loss": 0.6158,
1593
+ "step": 1075
1594
+ },
1595
+ {
1596
+ "epoch": 9.863013698630137,
1597
+ "grad_norm": 0.20196588337421417,
1598
+ "learning_rate": 5.127369531473525e-08,
1599
+ "loss": 0.614,
1600
+ "step": 1080
1601
+ },
1602
+ {
1603
+ "epoch": 9.908675799086758,
1604
+ "grad_norm": 0.19875915348529816,
1605
+ "learning_rate": 1.2819245493955744e-08,
1606
+ "loss": 0.6174,
1607
+ "step": 1085
1608
+ },
1609
+ {
1610
+ "epoch": 9.954337899543379,
1611
+ "grad_norm": 0.19983670115470886,
1612
+ "learning_rate": 0.0,
1613
+ "loss": 0.6194,
1614
+ "step": 1090
1615
+ },
1616
+ {
1617
+ "epoch": 9.954337899543379,
1618
+ "eval_loss": 2.260620594024658,
1619
+ "eval_runtime": 0.6934,
1620
+ "eval_samples_per_second": 14.422,
1621
+ "eval_steps_per_second": 1.442,
1622
+ "step": 1090
1623
+ },
1624
+ {
1625
+ "epoch": 9.954337899543379,
1626
+ "step": 1090,
1627
+ "total_flos": 6.456679991336763e+18,
1628
+ "train_loss": 0.7130365929472338,
1629
+ "train_runtime": 9668.9725,
1630
+ "train_samples_per_second": 14.452,
1631
+ "train_steps_per_second": 0.113
1632
+ }
1633
+ ],
1634
+ "logging_steps": 5,
1635
+ "max_steps": 1090,
1636
+ "num_input_tokens_seen": 0,
1637
+ "num_train_epochs": 10,
1638
+ "save_steps": 100,
1639
+ "stateful_callbacks": {
1640
+ "TrainerControl": {
1641
+ "args": {
1642
+ "should_epoch_stop": false,
1643
+ "should_evaluate": false,
1644
+ "should_log": false,
1645
+ "should_save": true,
1646
+ "should_training_stop": true
1647
+ },
1648
+ "attributes": {}
1649
+ }
1650
+ },
1651
+ "total_flos": 6.456679991336763e+18,
1652
+ "train_batch_size": 8,
1653
+ "trial_name": null,
1654
+ "trial_params": null
1655
+ }