Yifan99 commited on
Commit
ef20ef8
·
verified ·
1 Parent(s): 4f342fe

Model save

Browse files
README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: mistralai/Mistral-7B-v0.1
3
+ library_name: transformers
4
+ model_name: zephyr-7b-sft-full
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for zephyr-7b-sft-full
13
+
14
+ This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="Yifan99/zephyr-7b-sft-full", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+
31
+
32
+ This model was trained with SFT.
33
+
34
+ ### Framework versions
35
+
36
+ - TRL: 0.13.0.dev0
37
+ - Transformers: 4.46.0
38
+ - Pytorch: 2.3.0+cu121
39
+ - Datasets: 3.6.0
40
+ - Tokenizers: 0.20.3
41
+
42
+ ## Citations
43
+
44
+
45
+
46
+ Cite TRL as:
47
+
48
+ ```bibtex
49
+ @misc{vonwerra2022trl,
50
+ title = {{TRL: Transformer Reinforcement Learning}},
51
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
52
+ year = 2020,
53
+ journal = {GitHub repository},
54
+ publisher = {GitHub},
55
+ howpublished = {\url{https://github.com/huggingface/trl}}
56
+ }
57
+ ```
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 453935093514240.0,
4
+ "train_loss": 0.982664321841349,
5
+ "train_runtime": 6266.1955,
6
+ "train_samples": 207864,
7
+ "train_samples_per_second": 22.133,
8
+ "train_steps_per_second": 0.173
9
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.46.0"
6
+ }
runs/May30_18-42-57_yifaner-gpu-iad/events.out.tfevents.1748630614.yifaner-gpu-iad.34892.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a4e1ead04e7b13a3ca80e589761b5e16b3859a21d87baf5674c190affcdd47
3
- size 51091
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e8b5d3c1696243a912f9bfef3b6b0c78c36ab275e40980151490058dfd221d9
3
+ size 51716
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 453935093514240.0,
4
+ "train_loss": 0.982664321841349,
5
+ "train_runtime": 6266.1955,
6
+ "train_samples": 207864,
7
+ "train_samples_per_second": 22.133,
8
+ "train_steps_per_second": 0.173
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,1569 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1084,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0009225092250922509,
13
+ "grad_norm": 9.870990956795538,
14
+ "learning_rate": 1.8348623853211012e-07,
15
+ "loss": 1.1607,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.004612546125461255,
20
+ "grad_norm": 7.654279412229945,
21
+ "learning_rate": 9.174311926605506e-07,
22
+ "loss": 1.1303,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.00922509225092251,
27
+ "grad_norm": 3.8667009276426327,
28
+ "learning_rate": 1.8348623853211011e-06,
29
+ "loss": 1.0569,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.013837638376383764,
34
+ "grad_norm": 2.534713969460468,
35
+ "learning_rate": 2.7522935779816517e-06,
36
+ "loss": 1.0237,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.01845018450184502,
41
+ "grad_norm": 2.136778830664697,
42
+ "learning_rate": 3.6697247706422022e-06,
43
+ "loss": 0.993,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.023062730627306273,
48
+ "grad_norm": 1.9089432945409115,
49
+ "learning_rate": 4.587155963302753e-06,
50
+ "loss": 0.9822,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.027675276752767528,
55
+ "grad_norm": 2.483736099993364,
56
+ "learning_rate": 5.504587155963303e-06,
57
+ "loss": 0.9799,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.03228782287822878,
62
+ "grad_norm": 2.7399782424767936,
63
+ "learning_rate": 6.422018348623854e-06,
64
+ "loss": 0.9811,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.03690036900369004,
69
+ "grad_norm": 2.5580299604649968,
70
+ "learning_rate": 7.3394495412844045e-06,
71
+ "loss": 0.9806,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.04151291512915129,
76
+ "grad_norm": 2.25360816928206,
77
+ "learning_rate": 8.256880733944956e-06,
78
+ "loss": 0.9867,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.046125461254612546,
83
+ "grad_norm": 2.225590488447417,
84
+ "learning_rate": 9.174311926605506e-06,
85
+ "loss": 0.9966,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.0507380073800738,
90
+ "grad_norm": 2.0775964396198394,
91
+ "learning_rate": 1.0091743119266055e-05,
92
+ "loss": 0.991,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.055350553505535055,
97
+ "grad_norm": 2.3870280823080825,
98
+ "learning_rate": 1.1009174311926607e-05,
99
+ "loss": 0.9927,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.05996309963099631,
104
+ "grad_norm": 1.9832001883028227,
105
+ "learning_rate": 1.1926605504587156e-05,
106
+ "loss": 1.0028,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.06457564575645756,
111
+ "grad_norm": 2.494647363270376,
112
+ "learning_rate": 1.2844036697247708e-05,
113
+ "loss": 0.9875,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.06918819188191883,
118
+ "grad_norm": 2.145757053008219,
119
+ "learning_rate": 1.3761467889908258e-05,
120
+ "loss": 1.004,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.07380073800738007,
125
+ "grad_norm": 2.4778285872304453,
126
+ "learning_rate": 1.4678899082568809e-05,
127
+ "loss": 1.0176,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.07841328413284133,
132
+ "grad_norm": 2.378650902592371,
133
+ "learning_rate": 1.559633027522936e-05,
134
+ "loss": 1.0029,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.08302583025830258,
139
+ "grad_norm": 2.6973951224982935,
140
+ "learning_rate": 1.6513761467889912e-05,
141
+ "loss": 0.9945,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.08763837638376384,
146
+ "grad_norm": 2.7900934520086405,
147
+ "learning_rate": 1.743119266055046e-05,
148
+ "loss": 0.9874,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.09225092250922509,
153
+ "grad_norm": 1.9716018629774277,
154
+ "learning_rate": 1.834862385321101e-05,
155
+ "loss": 1.0383,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.09686346863468635,
160
+ "grad_norm": 2.4571675806519564,
161
+ "learning_rate": 1.9266055045871563e-05,
162
+ "loss": 1.0108,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.1014760147601476,
167
+ "grad_norm": 2.1433576022986025,
168
+ "learning_rate": 1.9999948088910656e-05,
169
+ "loss": 1.014,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.10608856088560886,
174
+ "grad_norm": 2.2070794225590524,
175
+ "learning_rate": 1.9998131257372878e-05,
176
+ "loss": 1.0238,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.11070110701107011,
181
+ "grad_norm": 1.797407393598312,
182
+ "learning_rate": 1.999371941029485e-05,
183
+ "loss": 1.0143,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.11531365313653137,
188
+ "grad_norm": 1.9527368920731798,
189
+ "learning_rate": 1.9986713692771732e-05,
190
+ "loss": 1.0372,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.11992619926199262,
195
+ "grad_norm": 2.065891476564467,
196
+ "learning_rate": 1.9977115923137912e-05,
197
+ "loss": 1.03,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.12453874538745388,
202
+ "grad_norm": 1.8362510554036173,
203
+ "learning_rate": 1.9964928592495046e-05,
204
+ "loss": 1.0307,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.12915129151291513,
209
+ "grad_norm": 2.054566764209732,
210
+ "learning_rate": 1.9950154864065497e-05,
211
+ "loss": 1.0292,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.13376383763837638,
216
+ "grad_norm": 1.9408930683317647,
217
+ "learning_rate": 1.993279857237133e-05,
218
+ "loss": 1.0332,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.13837638376383765,
223
+ "grad_norm": 2.046730114559884,
224
+ "learning_rate": 1.9912864222239045e-05,
225
+ "loss": 1.0408,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.1429889298892989,
230
+ "grad_norm": 2.146975729863243,
231
+ "learning_rate": 1.9890356987630362e-05,
232
+ "loss": 1.0212,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.14760147601476015,
237
+ "grad_norm": 1.8119879337927547,
238
+ "learning_rate": 1.986528271029931e-05,
239
+ "loss": 1.0241,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.1522140221402214,
244
+ "grad_norm": 1.812857293149788,
245
+ "learning_rate": 1.9837647898276008e-05,
246
+ "loss": 1.0248,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.15682656826568267,
251
+ "grad_norm": 2.148000093226172,
252
+ "learning_rate": 1.9807459724177497e-05,
253
+ "loss": 1.0268,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.16143911439114392,
258
+ "grad_norm": 1.9129262356991528,
259
+ "learning_rate": 1.977472602334609e-05,
260
+ "loss": 1.033,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.16605166051660517,
265
+ "grad_norm": 1.944891593472809,
266
+ "learning_rate": 1.973945529181572e-05,
267
+ "loss": 1.0346,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.1706642066420664,
272
+ "grad_norm": 1.939828197754357,
273
+ "learning_rate": 1.9701656684106764e-05,
274
+ "loss": 1.0404,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.1752767527675277,
279
+ "grad_norm": 1.8683452140679382,
280
+ "learning_rate": 1.9661340010850025e-05,
281
+ "loss": 1.0296,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.17988929889298894,
286
+ "grad_norm": 1.7548439453154032,
287
+ "learning_rate": 1.9618515736240353e-05,
288
+ "loss": 1.0511,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.18450184501845018,
293
+ "grad_norm": 1.8566237273323596,
294
+ "learning_rate": 1.9573194975320672e-05,
295
+ "loss": 1.0381,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.18911439114391143,
300
+ "grad_norm": 1.7903039830605652,
301
+ "learning_rate": 1.952538949109708e-05,
302
+ "loss": 1.0451,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.1937269372693727,
307
+ "grad_norm": 1.731821818321784,
308
+ "learning_rate": 1.9475111691485737e-05,
309
+ "loss": 1.0224,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.19833948339483395,
314
+ "grad_norm": 1.6546119685109435,
315
+ "learning_rate": 1.9422374626092414e-05,
316
+ "loss": 1.0312,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.2029520295202952,
321
+ "grad_norm": 1.6376019106503288,
322
+ "learning_rate": 1.936719198282545e-05,
323
+ "loss": 1.0307,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.20756457564575645,
328
+ "grad_norm": 1.5223498231652732,
329
+ "learning_rate": 1.930957808434307e-05,
330
+ "loss": 1.0276,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.21217712177121772,
335
+ "grad_norm": 1.5819696965290848,
336
+ "learning_rate": 1.9249547884335917e-05,
337
+ "loss": 1.0303,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.21678966789667897,
342
+ "grad_norm": 1.7581116028501704,
343
+ "learning_rate": 1.9187116963645845e-05,
344
+ "loss": 1.0357,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.22140221402214022,
349
+ "grad_norm": 1.7533150380634188,
350
+ "learning_rate": 1.912230152622189e-05,
351
+ "loss": 1.0329,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.22601476014760147,
356
+ "grad_norm": 1.838564355654299,
357
+ "learning_rate": 1.9055118394914545e-05,
358
+ "loss": 1.0361,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.23062730627306274,
363
+ "grad_norm": 1.584225431462834,
364
+ "learning_rate": 1.898558500710939e-05,
365
+ "loss": 1.0328,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.235239852398524,
370
+ "grad_norm": 1.665130350791042,
371
+ "learning_rate": 1.891371941020121e-05,
372
+ "loss": 1.0532,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.23985239852398524,
377
+ "grad_norm": 1.5868432052387649,
378
+ "learning_rate": 1.88395402569098e-05,
379
+ "loss": 1.0162,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.2444649446494465,
384
+ "grad_norm": 1.612399555276925,
385
+ "learning_rate": 1.8763066800438638e-05,
386
+ "loss": 1.0252,
387
+ "step": 265
388
+ },
389
+ {
390
+ "epoch": 0.24907749077490776,
391
+ "grad_norm": 1.743031080914046,
392
+ "learning_rate": 1.868431888947773e-05,
393
+ "loss": 1.0367,
394
+ "step": 270
395
+ },
396
+ {
397
+ "epoch": 0.253690036900369,
398
+ "grad_norm": 1.5784408195964303,
399
+ "learning_rate": 1.860331696305188e-05,
400
+ "loss": 1.0271,
401
+ "step": 275
402
+ },
403
+ {
404
+ "epoch": 0.25830258302583026,
405
+ "grad_norm": 1.654736307660285,
406
+ "learning_rate": 1.852008204521572e-05,
407
+ "loss": 1.0197,
408
+ "step": 280
409
+ },
410
+ {
411
+ "epoch": 0.2629151291512915,
412
+ "grad_norm": 1.5194226783030391,
413
+ "learning_rate": 1.8434635739596945e-05,
414
+ "loss": 1.0364,
415
+ "step": 285
416
+ },
417
+ {
418
+ "epoch": 0.26752767527675275,
419
+ "grad_norm": 1.5778882827801204,
420
+ "learning_rate": 1.834700022378907e-05,
421
+ "loss": 1.0543,
422
+ "step": 290
423
+ },
424
+ {
425
+ "epoch": 0.272140221402214,
426
+ "grad_norm": 1.6004388094053688,
427
+ "learning_rate": 1.825719824359524e-05,
428
+ "loss": 1.0403,
429
+ "step": 295
430
+ },
431
+ {
432
+ "epoch": 0.2767527675276753,
433
+ "grad_norm": 1.5745761589908656,
434
+ "learning_rate": 1.816525310712456e-05,
435
+ "loss": 1.0098,
436
+ "step": 300
437
+ },
438
+ {
439
+ "epoch": 0.28136531365313655,
440
+ "grad_norm": 1.6069217146531487,
441
+ "learning_rate": 1.8071188678742457e-05,
442
+ "loss": 1.0168,
443
+ "step": 305
444
+ },
445
+ {
446
+ "epoch": 0.2859778597785978,
447
+ "grad_norm": 1.5646672793401115,
448
+ "learning_rate": 1.7975029372876706e-05,
449
+ "loss": 1.0215,
450
+ "step": 310
451
+ },
452
+ {
453
+ "epoch": 0.29059040590405905,
454
+ "grad_norm": 1.543486819243714,
455
+ "learning_rate": 1.787680014768065e-05,
456
+ "loss": 1.0439,
457
+ "step": 315
458
+ },
459
+ {
460
+ "epoch": 0.2952029520295203,
461
+ "grad_norm": 1.7337050309315247,
462
+ "learning_rate": 1.777652649855531e-05,
463
+ "loss": 1.0183,
464
+ "step": 320
465
+ },
466
+ {
467
+ "epoch": 0.29981549815498154,
468
+ "grad_norm": 1.680429731562412,
469
+ "learning_rate": 1.7674234451532065e-05,
470
+ "loss": 1.031,
471
+ "step": 325
472
+ },
473
+ {
474
+ "epoch": 0.3044280442804428,
475
+ "grad_norm": 1.6116817757891433,
476
+ "learning_rate": 1.7569950556517566e-05,
477
+ "loss": 1.0513,
478
+ "step": 330
479
+ },
480
+ {
481
+ "epoch": 0.30904059040590404,
482
+ "grad_norm": 1.5206192596129697,
483
+ "learning_rate": 1.7463701880402738e-05,
484
+ "loss": 1.0204,
485
+ "step": 335
486
+ },
487
+ {
488
+ "epoch": 0.31365313653136534,
489
+ "grad_norm": 1.5945571759720971,
490
+ "learning_rate": 1.7355516000037555e-05,
491
+ "loss": 1.0278,
492
+ "step": 340
493
+ },
494
+ {
495
+ "epoch": 0.3182656826568266,
496
+ "grad_norm": 1.5938814901923302,
497
+ "learning_rate": 1.7245420995073453e-05,
498
+ "loss": 1.0401,
499
+ "step": 345
500
+ },
501
+ {
502
+ "epoch": 0.32287822878228783,
503
+ "grad_norm": 1.6084910280435076,
504
+ "learning_rate": 1.7133445440675268e-05,
505
+ "loss": 1.0269,
506
+ "step": 350
507
+ },
508
+ {
509
+ "epoch": 0.3274907749077491,
510
+ "grad_norm": 1.6256132765843339,
511
+ "learning_rate": 1.7019618400104572e-05,
512
+ "loss": 1.0259,
513
+ "step": 355
514
+ },
515
+ {
516
+ "epoch": 0.33210332103321033,
517
+ "grad_norm": 1.5411807350221058,
518
+ "learning_rate": 1.6903969417176244e-05,
519
+ "loss": 1.0186,
520
+ "step": 360
521
+ },
522
+ {
523
+ "epoch": 0.3367158671586716,
524
+ "grad_norm": 1.52016944094053,
525
+ "learning_rate": 1.6786528508590436e-05,
526
+ "loss": 1.033,
527
+ "step": 365
528
+ },
529
+ {
530
+ "epoch": 0.3413284132841328,
531
+ "grad_norm": 1.8803691215532436,
532
+ "learning_rate": 1.666732615614169e-05,
533
+ "loss": 1.02,
534
+ "step": 370
535
+ },
536
+ {
537
+ "epoch": 0.3459409594095941,
538
+ "grad_norm": 1.499271413842361,
539
+ "learning_rate": 1.6546393298807405e-05,
540
+ "loss": 1.0178,
541
+ "step": 375
542
+ },
543
+ {
544
+ "epoch": 0.3505535055350554,
545
+ "grad_norm": 1.6023328682377094,
546
+ "learning_rate": 1.6423761324717636e-05,
547
+ "loss": 1.023,
548
+ "step": 380
549
+ },
550
+ {
551
+ "epoch": 0.3551660516605166,
552
+ "grad_norm": 1.5304826653478296,
553
+ "learning_rate": 1.6299462063008272e-05,
554
+ "loss": 1.0155,
555
+ "step": 385
556
+ },
557
+ {
558
+ "epoch": 0.35977859778597787,
559
+ "grad_norm": 1.5394673610887177,
560
+ "learning_rate": 1.61735277755598e-05,
561
+ "loss": 1.023,
562
+ "step": 390
563
+ },
564
+ {
565
+ "epoch": 0.3643911439114391,
566
+ "grad_norm": 1.813007395614376,
567
+ "learning_rate": 1.6045991148623752e-05,
568
+ "loss": 1.0168,
569
+ "step": 395
570
+ },
571
+ {
572
+ "epoch": 0.36900369003690037,
573
+ "grad_norm": 1.5526018219731565,
574
+ "learning_rate": 1.5916885284338937e-05,
575
+ "loss": 1.0045,
576
+ "step": 400
577
+ },
578
+ {
579
+ "epoch": 0.3736162361623616,
580
+ "grad_norm": 1.4318690915846115,
581
+ "learning_rate": 1.5786243692139826e-05,
582
+ "loss": 1.0265,
583
+ "step": 405
584
+ },
585
+ {
586
+ "epoch": 0.37822878228782286,
587
+ "grad_norm": 1.65542416947911,
588
+ "learning_rate": 1.5654100280059155e-05,
589
+ "loss": 1.0038,
590
+ "step": 410
591
+ },
592
+ {
593
+ "epoch": 0.3828413284132841,
594
+ "grad_norm": 1.5036498809187224,
595
+ "learning_rate": 1.5520489345927095e-05,
596
+ "loss": 1.003,
597
+ "step": 415
598
+ },
599
+ {
600
+ "epoch": 0.3874538745387454,
601
+ "grad_norm": 1.571209743645404,
602
+ "learning_rate": 1.538544556846925e-05,
603
+ "loss": 1.0251,
604
+ "step": 420
605
+ },
606
+ {
607
+ "epoch": 0.39206642066420666,
608
+ "grad_norm": 1.4496764368758055,
609
+ "learning_rate": 1.5249003998305787e-05,
610
+ "loss": 1.0136,
611
+ "step": 425
612
+ },
613
+ {
614
+ "epoch": 0.3966789667896679,
615
+ "grad_norm": 1.5466821761987914,
616
+ "learning_rate": 1.5111200048854055e-05,
617
+ "loss": 1.0194,
618
+ "step": 430
619
+ },
620
+ {
621
+ "epoch": 0.40129151291512916,
622
+ "grad_norm": 1.4987449489819178,
623
+ "learning_rate": 1.4972069487137024e-05,
624
+ "loss": 1.009,
625
+ "step": 435
626
+ },
627
+ {
628
+ "epoch": 0.4059040590405904,
629
+ "grad_norm": 1.4694129512198164,
630
+ "learning_rate": 1.4831648424499953e-05,
631
+ "loss": 1.0149,
632
+ "step": 440
633
+ },
634
+ {
635
+ "epoch": 0.41051660516605165,
636
+ "grad_norm": 1.434796826955623,
637
+ "learning_rate": 1.4689973307237687e-05,
638
+ "loss": 1.0285,
639
+ "step": 445
640
+ },
641
+ {
642
+ "epoch": 0.4151291512915129,
643
+ "grad_norm": 1.510041122963638,
644
+ "learning_rate": 1.4547080907135024e-05,
645
+ "loss": 1.013,
646
+ "step": 450
647
+ },
648
+ {
649
+ "epoch": 0.41974169741697415,
650
+ "grad_norm": 1.4273777036410285,
651
+ "learning_rate": 1.4403008311922593e-05,
652
+ "loss": 1.0106,
653
+ "step": 455
654
+ },
655
+ {
656
+ "epoch": 0.42435424354243545,
657
+ "grad_norm": 1.434029200696217,
658
+ "learning_rate": 1.4257792915650728e-05,
659
+ "loss": 1.0126,
660
+ "step": 460
661
+ },
662
+ {
663
+ "epoch": 0.4289667896678967,
664
+ "grad_norm": 1.3526682049403003,
665
+ "learning_rate": 1.4111472408983843e-05,
666
+ "loss": 1.0056,
667
+ "step": 465
668
+ },
669
+ {
670
+ "epoch": 0.43357933579335795,
671
+ "grad_norm": 1.560807386777768,
672
+ "learning_rate": 1.3964084769417823e-05,
673
+ "loss": 1.0293,
674
+ "step": 470
675
+ },
676
+ {
677
+ "epoch": 0.4381918819188192,
678
+ "grad_norm": 1.6384296659793862,
679
+ "learning_rate": 1.3815668251422953e-05,
680
+ "loss": 0.9952,
681
+ "step": 475
682
+ },
683
+ {
684
+ "epoch": 0.44280442804428044,
685
+ "grad_norm": 1.4643281809184674,
686
+ "learning_rate": 1.3666261376514978e-05,
687
+ "loss": 0.9927,
688
+ "step": 480
689
+ },
690
+ {
691
+ "epoch": 0.4474169741697417,
692
+ "grad_norm": 1.4762693155800783,
693
+ "learning_rate": 1.3515902923256832e-05,
694
+ "loss": 1.0082,
695
+ "step": 485
696
+ },
697
+ {
698
+ "epoch": 0.45202952029520294,
699
+ "grad_norm": 1.575453671421441,
700
+ "learning_rate": 1.3364631917193671e-05,
701
+ "loss": 1.0018,
702
+ "step": 490
703
+ },
704
+ {
705
+ "epoch": 0.4566420664206642,
706
+ "grad_norm": 1.7509884043580597,
707
+ "learning_rate": 1.321248762072377e-05,
708
+ "loss": 1.0031,
709
+ "step": 495
710
+ },
711
+ {
712
+ "epoch": 0.4612546125461255,
713
+ "grad_norm": 1.4633205407763117,
714
+ "learning_rate": 1.3059509522907998e-05,
715
+ "loss": 0.9977,
716
+ "step": 500
717
+ },
718
+ {
719
+ "epoch": 0.46586715867158673,
720
+ "grad_norm": 1.4464913424686388,
721
+ "learning_rate": 1.2905737329220394e-05,
722
+ "loss": 0.9962,
723
+ "step": 505
724
+ },
725
+ {
726
+ "epoch": 0.470479704797048,
727
+ "grad_norm": 1.5019368087709983,
728
+ "learning_rate": 1.2751210951242636e-05,
729
+ "loss": 1.0047,
730
+ "step": 510
731
+ },
732
+ {
733
+ "epoch": 0.47509225092250923,
734
+ "grad_norm": 1.424993301093826,
735
+ "learning_rate": 1.2595970496304975e-05,
736
+ "loss": 0.9935,
737
+ "step": 515
738
+ },
739
+ {
740
+ "epoch": 0.4797047970479705,
741
+ "grad_norm": 1.397860945200415,
742
+ "learning_rate": 1.2440056257076376e-05,
743
+ "loss": 1.0072,
744
+ "step": 520
745
+ },
746
+ {
747
+ "epoch": 0.4843173431734317,
748
+ "grad_norm": 1.4591151077748796,
749
+ "learning_rate": 1.2283508701106559e-05,
750
+ "loss": 0.9785,
751
+ "step": 525
752
+ },
753
+ {
754
+ "epoch": 0.488929889298893,
755
+ "grad_norm": 1.3985403114867758,
756
+ "learning_rate": 1.2126368460322637e-05,
757
+ "loss": 1.0073,
758
+ "step": 530
759
+ },
760
+ {
761
+ "epoch": 0.4935424354243542,
762
+ "grad_norm": 1.3622030450903666,
763
+ "learning_rate": 1.1968676320483103e-05,
764
+ "loss": 0.9679,
765
+ "step": 535
766
+ },
767
+ {
768
+ "epoch": 0.4981549815498155,
769
+ "grad_norm": 1.4585144558991643,
770
+ "learning_rate": 1.1810473210591882e-05,
771
+ "loss": 0.9846,
772
+ "step": 540
773
+ },
774
+ {
775
+ "epoch": 0.5027675276752768,
776
+ "grad_norm": 1.4502741925244078,
777
+ "learning_rate": 1.1651800192275197e-05,
778
+ "loss": 0.9792,
779
+ "step": 545
780
+ },
781
+ {
782
+ "epoch": 0.507380073800738,
783
+ "grad_norm": 1.4112975394158198,
784
+ "learning_rate": 1.1492698449124042e-05,
785
+ "loss": 0.9882,
786
+ "step": 550
787
+ },
788
+ {
789
+ "epoch": 0.5119926199261993,
790
+ "grad_norm": 1.3902737511333096,
791
+ "learning_rate": 1.1333209276004959e-05,
792
+ "loss": 0.996,
793
+ "step": 555
794
+ },
795
+ {
796
+ "epoch": 0.5166051660516605,
797
+ "grad_norm": 1.4114090100980268,
798
+ "learning_rate": 1.1173374068341962e-05,
799
+ "loss": 1.0053,
800
+ "step": 560
801
+ },
802
+ {
803
+ "epoch": 0.5212177121771218,
804
+ "grad_norm": 1.4743164096907508,
805
+ "learning_rate": 1.1013234311372353e-05,
806
+ "loss": 0.9835,
807
+ "step": 565
808
+ },
809
+ {
810
+ "epoch": 0.525830258302583,
811
+ "grad_norm": 1.519429603617901,
812
+ "learning_rate": 1.0852831569379217e-05,
813
+ "loss": 0.9908,
814
+ "step": 570
815
+ },
816
+ {
817
+ "epoch": 0.5304428044280443,
818
+ "grad_norm": 1.453522178819256,
819
+ "learning_rate": 1.0692207474903421e-05,
820
+ "loss": 0.991,
821
+ "step": 575
822
+ },
823
+ {
824
+ "epoch": 0.5350553505535055,
825
+ "grad_norm": 1.37300728807636,
826
+ "learning_rate": 1.0531403717937888e-05,
827
+ "loss": 0.9887,
828
+ "step": 580
829
+ },
830
+ {
831
+ "epoch": 0.5396678966789668,
832
+ "grad_norm": 1.4658227361639529,
833
+ "learning_rate": 1.037046203510694e-05,
834
+ "loss": 0.975,
835
+ "step": 585
836
+ },
837
+ {
838
+ "epoch": 0.544280442804428,
839
+ "grad_norm": 1.3920165357631773,
840
+ "learning_rate": 1.0209424198833571e-05,
841
+ "loss": 0.987,
842
+ "step": 590
843
+ },
844
+ {
845
+ "epoch": 0.5488929889298892,
846
+ "grad_norm": 1.5056794833077605,
847
+ "learning_rate": 1.0048332006497406e-05,
848
+ "loss": 0.9936,
849
+ "step": 595
850
+ },
851
+ {
852
+ "epoch": 0.5535055350553506,
853
+ "grad_norm": 1.3882676355943753,
854
+ "learning_rate": 9.887227269586184e-06,
855
+ "loss": 0.9807,
856
+ "step": 600
857
+ },
858
+ {
859
+ "epoch": 0.5581180811808119,
860
+ "grad_norm": 1.466557983615673,
861
+ "learning_rate": 9.7261518028436e-06,
862
+ "loss": 0.99,
863
+ "step": 605
864
+ },
865
+ {
866
+ "epoch": 0.5627306273062731,
867
+ "grad_norm": 1.4193893381047582,
868
+ "learning_rate": 9.565147413416266e-06,
869
+ "loss": 0.9782,
870
+ "step": 610
871
+ },
872
+ {
873
+ "epoch": 0.5673431734317343,
874
+ "grad_norm": 1.3465535691560198,
875
+ "learning_rate": 9.404255890002677e-06,
876
+ "loss": 0.9963,
877
+ "step": 615
878
+ },
879
+ {
880
+ "epoch": 0.5719557195571956,
881
+ "grad_norm": 1.2838850891021258,
882
+ "learning_rate": 9.243518992006944e-06,
883
+ "loss": 0.9735,
884
+ "step": 620
885
+ },
886
+ {
887
+ "epoch": 0.5765682656826568,
888
+ "grad_norm": 1.3173792395018515,
889
+ "learning_rate": 9.082978438700138e-06,
890
+ "loss": 0.968,
891
+ "step": 625
892
+ },
893
+ {
894
+ "epoch": 0.5811808118081181,
895
+ "grad_norm": 1.4339058777471723,
896
+ "learning_rate": 8.922675898392072e-06,
897
+ "loss": 0.9663,
898
+ "step": 630
899
+ },
900
+ {
901
+ "epoch": 0.5857933579335793,
902
+ "grad_norm": 1.3593794949774558,
903
+ "learning_rate": 8.762652977616258e-06,
904
+ "loss": 0.9861,
905
+ "step": 635
906
+ },
907
+ {
908
+ "epoch": 0.5904059040590406,
909
+ "grad_norm": 1.3750915173575393,
910
+ "learning_rate": 8.602951210330942e-06,
911
+ "loss": 0.9658,
912
+ "step": 640
913
+ },
914
+ {
915
+ "epoch": 0.5950184501845018,
916
+ "grad_norm": 1.3517116878942004,
917
+ "learning_rate": 8.443612047138965e-06,
918
+ "loss": 0.982,
919
+ "step": 645
920
+ },
921
+ {
922
+ "epoch": 0.5996309963099631,
923
+ "grad_norm": 1.4091397156737253,
924
+ "learning_rate": 8.284676844529258e-06,
925
+ "loss": 0.945,
926
+ "step": 650
927
+ },
928
+ {
929
+ "epoch": 0.6042435424354243,
930
+ "grad_norm": 1.2934337488666847,
931
+ "learning_rate": 8.126186854142752e-06,
932
+ "loss": 0.9526,
933
+ "step": 655
934
+ },
935
+ {
936
+ "epoch": 0.6088560885608856,
937
+ "grad_norm": 1.336680447919326,
938
+ "learning_rate": 7.968183212065537e-06,
939
+ "loss": 0.9648,
940
+ "step": 660
941
+ },
942
+ {
943
+ "epoch": 0.6134686346863468,
944
+ "grad_norm": 1.3335063906222555,
945
+ "learning_rate": 7.81070692815195e-06,
946
+ "loss": 0.9616,
947
+ "step": 665
948
+ },
949
+ {
950
+ "epoch": 0.6180811808118081,
951
+ "grad_norm": 1.4028504727483062,
952
+ "learning_rate": 7.6537988753805e-06,
953
+ "loss": 0.9559,
954
+ "step": 670
955
+ },
956
+ {
957
+ "epoch": 0.6226937269372693,
958
+ "grad_norm": 1.3423760518443784,
959
+ "learning_rate": 7.497499779245268e-06,
960
+ "loss": 0.9649,
961
+ "step": 675
962
+ },
963
+ {
964
+ "epoch": 0.6273062730627307,
965
+ "grad_norm": 1.3763883937348929,
966
+ "learning_rate": 7.3418502071856004e-06,
967
+ "loss": 0.9727,
968
+ "step": 680
969
+ },
970
+ {
971
+ "epoch": 0.6319188191881919,
972
+ "grad_norm": 1.3538188475466573,
973
+ "learning_rate": 7.186890558056836e-06,
974
+ "loss": 0.9525,
975
+ "step": 685
976
+ },
977
+ {
978
+ "epoch": 0.6365313653136532,
979
+ "grad_norm": 1.4363152298279036,
980
+ "learning_rate": 7.0326610516447825e-06,
981
+ "loss": 0.9519,
982
+ "step": 690
983
+ },
984
+ {
985
+ "epoch": 0.6411439114391144,
986
+ "grad_norm": 1.4065045912643077,
987
+ "learning_rate": 6.879201718226658e-06,
988
+ "loss": 0.9433,
989
+ "step": 695
990
+ },
991
+ {
992
+ "epoch": 0.6457564575645757,
993
+ "grad_norm": 1.2689078033695567,
994
+ "learning_rate": 6.7265523881812335e-06,
995
+ "loss": 0.9567,
996
+ "step": 700
997
+ },
998
+ {
999
+ "epoch": 0.6503690036900369,
1000
+ "grad_norm": 1.2774626635023132,
1001
+ "learning_rate": 6.574752681650864e-06,
1002
+ "loss": 0.9423,
1003
+ "step": 705
1004
+ },
1005
+ {
1006
+ "epoch": 0.6549815498154982,
1007
+ "grad_norm": 1.2857657790671198,
1008
+ "learning_rate": 6.423841998258069e-06,
1009
+ "loss": 0.9488,
1010
+ "step": 710
1011
+ },
1012
+ {
1013
+ "epoch": 0.6595940959409594,
1014
+ "grad_norm": 1.3218825385433717,
1015
+ "learning_rate": 6.273859506879365e-06,
1016
+ "loss": 0.9675,
1017
+ "step": 715
1018
+ },
1019
+ {
1020
+ "epoch": 0.6642066420664207,
1021
+ "grad_norm": 1.3140499267955912,
1022
+ "learning_rate": 6.124844135478971e-06,
1023
+ "loss": 0.9645,
1024
+ "step": 720
1025
+ },
1026
+ {
1027
+ "epoch": 0.6688191881918819,
1028
+ "grad_norm": 1.2553616859013494,
1029
+ "learning_rate": 5.976834561005069e-06,
1030
+ "loss": 0.9593,
1031
+ "step": 725
1032
+ },
1033
+ {
1034
+ "epoch": 0.6734317343173432,
1035
+ "grad_norm": 1.3433064911689856,
1036
+ "learning_rate": 5.829869199351188e-06,
1037
+ "loss": 0.9491,
1038
+ "step": 730
1039
+ },
1040
+ {
1041
+ "epoch": 0.6780442804428044,
1042
+ "grad_norm": 1.3274312519554137,
1043
+ "learning_rate": 5.68398619538536e-06,
1044
+ "loss": 0.9425,
1045
+ "step": 735
1046
+ },
1047
+ {
1048
+ "epoch": 0.6826568265682657,
1049
+ "grad_norm": 1.3743238375183062,
1050
+ "learning_rate": 5.53922341304961e-06,
1051
+ "loss": 0.9378,
1052
+ "step": 740
1053
+ },
1054
+ {
1055
+ "epoch": 0.6872693726937269,
1056
+ "grad_norm": 1.3880330372119891,
1057
+ "learning_rate": 5.39561842553239e-06,
1058
+ "loss": 0.9759,
1059
+ "step": 745
1060
+ },
1061
+ {
1062
+ "epoch": 0.6918819188191881,
1063
+ "grad_norm": 1.3337273623917134,
1064
+ "learning_rate": 5.2532085055164205e-06,
1065
+ "loss": 0.9493,
1066
+ "step": 750
1067
+ },
1068
+ {
1069
+ "epoch": 0.6964944649446494,
1070
+ "grad_norm": 1.3077178571915684,
1071
+ "learning_rate": 5.112030615504601e-06,
1072
+ "loss": 0.9521,
1073
+ "step": 755
1074
+ },
1075
+ {
1076
+ "epoch": 0.7011070110701108,
1077
+ "grad_norm": 1.2742302447784248,
1078
+ "learning_rate": 4.972121398226371e-06,
1079
+ "loss": 0.9382,
1080
+ "step": 760
1081
+ },
1082
+ {
1083
+ "epoch": 0.705719557195572,
1084
+ "grad_norm": 1.3201506076517637,
1085
+ "learning_rate": 4.833517167127077e-06,
1086
+ "loss": 0.9525,
1087
+ "step": 765
1088
+ },
1089
+ {
1090
+ "epoch": 0.7103321033210332,
1091
+ "grad_norm": 1.3060162706383742,
1092
+ "learning_rate": 4.6962538969428416e-06,
1093
+ "loss": 0.9518,
1094
+ "step": 770
1095
+ },
1096
+ {
1097
+ "epoch": 0.7149446494464945,
1098
+ "grad_norm": 1.2666833397246693,
1099
+ "learning_rate": 4.560367214363295e-06,
1100
+ "loss": 0.9459,
1101
+ "step": 775
1102
+ },
1103
+ {
1104
+ "epoch": 0.7195571955719557,
1105
+ "grad_norm": 1.3752391118024025,
1106
+ "learning_rate": 4.425892388784681e-06,
1107
+ "loss": 0.9481,
1108
+ "step": 780
1109
+ },
1110
+ {
1111
+ "epoch": 0.724169741697417,
1112
+ "grad_norm": 1.381288016089436,
1113
+ "learning_rate": 4.292864323155684e-06,
1114
+ "loss": 0.9351,
1115
+ "step": 785
1116
+ },
1117
+ {
1118
+ "epoch": 0.7287822878228782,
1119
+ "grad_norm": 1.3075298809556533,
1120
+ "learning_rate": 4.161317544918345e-06,
1121
+ "loss": 0.9382,
1122
+ "step": 790
1123
+ },
1124
+ {
1125
+ "epoch": 0.7333948339483395,
1126
+ "grad_norm": 1.2816897642857799,
1127
+ "learning_rate": 4.031286197046493e-06,
1128
+ "loss": 0.9423,
1129
+ "step": 795
1130
+ },
1131
+ {
1132
+ "epoch": 0.7380073800738007,
1133
+ "grad_norm": 1.2968384739032883,
1134
+ "learning_rate": 3.902804029183907e-06,
1135
+ "loss": 0.9446,
1136
+ "step": 800
1137
+ },
1138
+ {
1139
+ "epoch": 0.742619926199262,
1140
+ "grad_norm": 1.329056519068419,
1141
+ "learning_rate": 3.775904388884618e-06,
1142
+ "loss": 0.9422,
1143
+ "step": 805
1144
+ },
1145
+ {
1146
+ "epoch": 0.7472324723247232,
1147
+ "grad_norm": 1.2876813358003367,
1148
+ "learning_rate": 3.650620212957524e-06,
1149
+ "loss": 0.9404,
1150
+ "step": 810
1151
+ },
1152
+ {
1153
+ "epoch": 0.7518450184501845,
1154
+ "grad_norm": 1.4667315877372276,
1155
+ "learning_rate": 3.5269840189176616e-06,
1156
+ "loss": 0.9282,
1157
+ "step": 815
1158
+ },
1159
+ {
1160
+ "epoch": 0.7564575645756457,
1161
+ "grad_norm": 1.301616249350314,
1162
+ "learning_rate": 3.405027896546277e-06,
1163
+ "loss": 0.9259,
1164
+ "step": 820
1165
+ },
1166
+ {
1167
+ "epoch": 0.761070110701107,
1168
+ "grad_norm": 1.3223905832430802,
1169
+ "learning_rate": 3.2847834995619067e-06,
1170
+ "loss": 0.9447,
1171
+ "step": 825
1172
+ },
1173
+ {
1174
+ "epoch": 0.7656826568265682,
1175
+ "grad_norm": 1.262448560449103,
1176
+ "learning_rate": 3.1662820374046776e-06,
1177
+ "loss": 0.9353,
1178
+ "step": 830
1179
+ },
1180
+ {
1181
+ "epoch": 0.7702952029520295,
1182
+ "grad_norm": 1.2951473418615072,
1183
+ "learning_rate": 3.0495542671358745e-06,
1184
+ "loss": 0.9291,
1185
+ "step": 835
1186
+ },
1187
+ {
1188
+ "epoch": 0.7749077490774908,
1189
+ "grad_norm": 1.2820260584603826,
1190
+ "learning_rate": 2.934630485454948e-06,
1191
+ "loss": 0.9356,
1192
+ "step": 840
1193
+ },
1194
+ {
1195
+ "epoch": 0.7795202952029521,
1196
+ "grad_norm": 1.2899359050310029,
1197
+ "learning_rate": 2.8215405208360237e-06,
1198
+ "loss": 0.9394,
1199
+ "step": 845
1200
+ },
1201
+ {
1202
+ "epoch": 0.7841328413284133,
1203
+ "grad_norm": 1.3428587589811845,
1204
+ "learning_rate": 2.7103137257858867e-06,
1205
+ "loss": 0.9355,
1206
+ "step": 850
1207
+ },
1208
+ {
1209
+ "epoch": 0.7887453874538746,
1210
+ "grad_norm": 1.3675117810995234,
1211
+ "learning_rate": 2.600978969225558e-06,
1212
+ "loss": 0.9406,
1213
+ "step": 855
1214
+ },
1215
+ {
1216
+ "epoch": 0.7933579335793358,
1217
+ "grad_norm": 1.285843469838144,
1218
+ "learning_rate": 2.493564628997369e-06,
1219
+ "loss": 0.9472,
1220
+ "step": 860
1221
+ },
1222
+ {
1223
+ "epoch": 0.7979704797047971,
1224
+ "grad_norm": 1.308062981287854,
1225
+ "learning_rate": 2.3880985844994674e-06,
1226
+ "loss": 0.9284,
1227
+ "step": 865
1228
+ },
1229
+ {
1230
+ "epoch": 0.8025830258302583,
1231
+ "grad_norm": 1.311817269933356,
1232
+ "learning_rate": 2.284608209449746e-06,
1233
+ "loss": 0.9357,
1234
+ "step": 870
1235
+ },
1236
+ {
1237
+ "epoch": 0.8071955719557196,
1238
+ "grad_norm": 1.3458370437390148,
1239
+ "learning_rate": 2.183120364780975e-06,
1240
+ "loss": 0.9387,
1241
+ "step": 875
1242
+ },
1243
+ {
1244
+ "epoch": 0.8118081180811808,
1245
+ "grad_norm": 1.2723012259493913,
1246
+ "learning_rate": 2.083661391669043e-06,
1247
+ "loss": 0.9252,
1248
+ "step": 880
1249
+ },
1250
+ {
1251
+ "epoch": 0.816420664206642,
1252
+ "grad_norm": 1.3187842301744654,
1253
+ "learning_rate": 1.986257104696121e-06,
1254
+ "loss": 0.9368,
1255
+ "step": 885
1256
+ },
1257
+ {
1258
+ "epoch": 0.8210332103321033,
1259
+ "grad_norm": 1.3542997263964067,
1260
+ "learning_rate": 1.8909327851504633e-06,
1261
+ "loss": 0.9526,
1262
+ "step": 890
1263
+ },
1264
+ {
1265
+ "epoch": 0.8256457564575646,
1266
+ "grad_norm": 1.3403454711197158,
1267
+ "learning_rate": 1.7977131744646724e-06,
1268
+ "loss": 0.9327,
1269
+ "step": 895
1270
+ },
1271
+ {
1272
+ "epoch": 0.8302583025830258,
1273
+ "grad_norm": 1.3077988340377713,
1274
+ "learning_rate": 1.7066224677940313e-06,
1275
+ "loss": 0.9431,
1276
+ "step": 900
1277
+ },
1278
+ {
1279
+ "epoch": 0.834870848708487,
1280
+ "grad_norm": 1.2548232997350641,
1281
+ "learning_rate": 1.6176843077366755e-06,
1282
+ "loss": 0.9136,
1283
+ "step": 905
1284
+ },
1285
+ {
1286
+ "epoch": 0.8394833948339483,
1287
+ "grad_norm": 1.2897227556284836,
1288
+ "learning_rate": 1.5309217781971419e-06,
1289
+ "loss": 0.9134,
1290
+ "step": 910
1291
+ },
1292
+ {
1293
+ "epoch": 0.8440959409594095,
1294
+ "grad_norm": 1.2995199761283964,
1295
+ "learning_rate": 1.446357398394934e-06,
1296
+ "loss": 0.9231,
1297
+ "step": 915
1298
+ },
1299
+ {
1300
+ "epoch": 0.8487084870848709,
1301
+ "grad_norm": 1.3112786455057477,
1302
+ "learning_rate": 1.3640131170196758e-06,
1303
+ "loss": 0.9157,
1304
+ "step": 920
1305
+ },
1306
+ {
1307
+ "epoch": 0.8533210332103321,
1308
+ "grad_norm": 1.3030169209685578,
1309
+ "learning_rate": 1.2839103065343084e-06,
1310
+ "loss": 0.9278,
1311
+ "step": 925
1312
+ },
1313
+ {
1314
+ "epoch": 0.8579335793357934,
1315
+ "grad_norm": 1.3094669969598054,
1316
+ "learning_rate": 1.2060697576278812e-06,
1317
+ "loss": 0.9329,
1318
+ "step": 930
1319
+ },
1320
+ {
1321
+ "epoch": 0.8625461254612546,
1322
+ "grad_norm": 1.276775688304098,
1323
+ "learning_rate": 1.1305116738193211e-06,
1324
+ "loss": 0.9346,
1325
+ "step": 935
1326
+ },
1327
+ {
1328
+ "epoch": 0.8671586715867159,
1329
+ "grad_norm": 1.2892434172054825,
1330
+ "learning_rate": 1.0572556662136036e-06,
1331
+ "loss": 0.9301,
1332
+ "step": 940
1333
+ },
1334
+ {
1335
+ "epoch": 0.8717712177121771,
1336
+ "grad_norm": 1.3021701449964453,
1337
+ "learning_rate": 9.863207484116987e-07,
1338
+ "loss": 0.9177,
1339
+ "step": 945
1340
+ },
1341
+ {
1342
+ "epoch": 0.8763837638376384,
1343
+ "grad_norm": 1.2852161940710303,
1344
+ "learning_rate": 9.177253315755796e-07,
1345
+ "loss": 0.9378,
1346
+ "step": 950
1347
+ },
1348
+ {
1349
+ "epoch": 0.8809963099630996,
1350
+ "grad_norm": 1.2451978022882328,
1351
+ "learning_rate": 8.514872196496182e-07,
1352
+ "loss": 0.9348,
1353
+ "step": 955
1354
+ },
1355
+ {
1356
+ "epoch": 0.8856088560885609,
1357
+ "grad_norm": 1.2141597430813968,
1358
+ "learning_rate": 7.876236047395525e-07,
1359
+ "loss": 0.9265,
1360
+ "step": 960
1361
+ },
1362
+ {
1363
+ "epoch": 0.8902214022140221,
1364
+ "grad_norm": 1.296008639609934,
1365
+ "learning_rate": 7.26151062650291e-07,
1366
+ "loss": 0.9338,
1367
+ "step": 965
1368
+ },
1369
+ {
1370
+ "epoch": 0.8948339483394834,
1371
+ "grad_norm": 1.3124661816253649,
1372
+ "learning_rate": 6.670855485836525e-07,
1373
+ "loss": 0.9281,
1374
+ "step": 970
1375
+ },
1376
+ {
1377
+ "epoch": 0.8994464944649446,
1378
+ "grad_norm": 1.2666059464350587,
1379
+ "learning_rate": 6.104423929971948e-07,
1380
+ "loss": 0.9327,
1381
+ "step": 975
1382
+ },
1383
+ {
1384
+ "epoch": 0.9040590405904059,
1385
+ "grad_norm": 1.2784756107848798,
1386
+ "learning_rate": 5.562362976251901e-07,
1387
+ "loss": 0.938,
1388
+ "step": 980
1389
+ },
1390
+ {
1391
+ "epoch": 0.9086715867158671,
1392
+ "grad_norm": 1.2246240779069228,
1393
+ "learning_rate": 5.044813316627994e-07,
1394
+ "loss": 0.9268,
1395
+ "step": 985
1396
+ },
1397
+ {
1398
+ "epoch": 0.9132841328413284,
1399
+ "grad_norm": 1.2911170040738167,
1400
+ "learning_rate": 4.5519092811439627e-07,
1401
+ "loss": 0.9248,
1402
+ "step": 990
1403
+ },
1404
+ {
1405
+ "epoch": 0.9178966789667896,
1406
+ "grad_norm": 1.320133517164497,
1407
+ "learning_rate": 4.083778803070504e-07,
1408
+ "loss": 0.9437,
1409
+ "step": 995
1410
+ },
1411
+ {
1412
+ "epoch": 0.922509225092251,
1413
+ "grad_norm": 1.2549143119794002,
1414
+ "learning_rate": 3.6405433856999684e-07,
1415
+ "loss": 0.9565,
1416
+ "step": 1000
1417
+ },
1418
+ {
1419
+ "epoch": 0.9271217712177122,
1420
+ "grad_norm": 1.324909180737576,
1421
+ "learning_rate": 3.2223180708102933e-07,
1422
+ "loss": 0.9322,
1423
+ "step": 1005
1424
+ },
1425
+ {
1426
+ "epoch": 0.9317343173431735,
1427
+ "grad_norm": 1.2882570015971073,
1428
+ "learning_rate": 2.829211408805932e-07,
1429
+ "loss": 0.9173,
1430
+ "step": 1010
1431
+ },
1432
+ {
1433
+ "epoch": 0.9363468634686347,
1434
+ "grad_norm": 1.67906584184866,
1435
+ "learning_rate": 2.461325430543482e-07,
1436
+ "loss": 0.9319,
1437
+ "step": 1015
1438
+ },
1439
+ {
1440
+ "epoch": 0.940959409594096,
1441
+ "grad_norm": 1.2543153354274186,
1442
+ "learning_rate": 2.1187556208496885e-07,
1443
+ "loss": 0.9212,
1444
+ "step": 1020
1445
+ },
1446
+ {
1447
+ "epoch": 0.9455719557195572,
1448
+ "grad_norm": 1.230292220248059,
1449
+ "learning_rate": 1.8015908937382587e-07,
1450
+ "loss": 0.9329,
1451
+ "step": 1025
1452
+ },
1453
+ {
1454
+ "epoch": 0.9501845018450185,
1455
+ "grad_norm": 1.2685375441760296,
1456
+ "learning_rate": 1.5099135693322776e-07,
1457
+ "loss": 0.9335,
1458
+ "step": 1030
1459
+ },
1460
+ {
1461
+ "epoch": 0.9547970479704797,
1462
+ "grad_norm": 1.2252210378690607,
1463
+ "learning_rate": 1.2437993524979984e-07,
1464
+ "loss": 0.9029,
1465
+ "step": 1035
1466
+ },
1467
+ {
1468
+ "epoch": 0.959409594095941,
1469
+ "grad_norm": 1.2888608840481186,
1470
+ "learning_rate": 1.0033173131956175e-07,
1471
+ "loss": 0.9211,
1472
+ "step": 1040
1473
+ },
1474
+ {
1475
+ "epoch": 0.9640221402214022,
1476
+ "grad_norm": 1.2495839916442792,
1477
+ "learning_rate": 7.885298685522235e-08,
1478
+ "loss": 0.9152,
1479
+ "step": 1045
1480
+ },
1481
+ {
1482
+ "epoch": 0.9686346863468634,
1483
+ "grad_norm": 1.279639118993962,
1484
+ "learning_rate": 5.99492766661347e-08,
1485
+ "loss": 0.9212,
1486
+ "step": 1050
1487
+ },
1488
+ {
1489
+ "epoch": 0.9732472324723247,
1490
+ "grad_norm": 1.2775107158396932,
1491
+ "learning_rate": 4.362550721136338e-08,
1492
+ "loss": 0.9252,
1493
+ "step": 1055
1494
+ },
1495
+ {
1496
+ "epoch": 0.977859778597786,
1497
+ "grad_norm": 1.2306806429728543,
1498
+ "learning_rate": 2.988591532620322e-08,
1499
+ "loss": 0.9151,
1500
+ "step": 1060
1501
+ },
1502
+ {
1503
+ "epoch": 0.9824723247232472,
1504
+ "grad_norm": 1.2648118187129338,
1505
+ "learning_rate": 1.8734067122514464e-08,
1506
+ "loss": 0.9222,
1507
+ "step": 1065
1508
+ },
1509
+ {
1510
+ "epoch": 0.9870848708487084,
1511
+ "grad_norm": 1.3467629616768353,
1512
+ "learning_rate": 1.0172857063137643e-08,
1513
+ "loss": 0.9251,
1514
+ "step": 1070
1515
+ },
1516
+ {
1517
+ "epoch": 0.9916974169741697,
1518
+ "grad_norm": 1.281124121994121,
1519
+ "learning_rate": 4.204507210633368e-09,
1520
+ "loss": 0.9465,
1521
+ "step": 1075
1522
+ },
1523
+ {
1524
+ "epoch": 0.996309963099631,
1525
+ "grad_norm": 1.2725002065016535,
1526
+ "learning_rate": 8.30566650548148e-10,
1527
+ "loss": 0.9402,
1528
+ "step": 1080
1529
+ },
1530
+ {
1531
+ "epoch": 1.0,
1532
+ "eval_loss": 0.9401968121528625,
1533
+ "eval_runtime": 156.1033,
1534
+ "eval_samples_per_second": 98.332,
1535
+ "eval_steps_per_second": 1.537,
1536
+ "step": 1084
1537
+ },
1538
+ {
1539
+ "epoch": 1.0,
1540
+ "step": 1084,
1541
+ "total_flos": 453935093514240.0,
1542
+ "train_loss": 0.982664321841349,
1543
+ "train_runtime": 6266.1955,
1544
+ "train_samples_per_second": 22.133,
1545
+ "train_steps_per_second": 0.173
1546
+ }
1547
+ ],
1548
+ "logging_steps": 5,
1549
+ "max_steps": 1084,
1550
+ "num_input_tokens_seen": 0,
1551
+ "num_train_epochs": 1,
1552
+ "save_steps": 100,
1553
+ "stateful_callbacks": {
1554
+ "TrainerControl": {
1555
+ "args": {
1556
+ "should_epoch_stop": false,
1557
+ "should_evaluate": false,
1558
+ "should_log": false,
1559
+ "should_save": true,
1560
+ "should_training_stop": true
1561
+ },
1562
+ "attributes": {}
1563
+ }
1564
+ },
1565
+ "total_flos": 453935093514240.0,
1566
+ "train_batch_size": 16,
1567
+ "trial_name": null,
1568
+ "trial_params": null
1569
+ }