Hugging Face's logo

WiroAI
/
OpenR1-Qwen-7B-Italian

Text Generation
Transformers
Safetensors
Italian
qwen2
Generated from Trainer
trl
sft
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
bezir commited on
Commit
e2aaf21
·
verified ·
1 Parent(s): aec177d

Model save

Browse files
Files changed (5) hide show
  1. README.md +58 -0
  2. all_results.json +8 -0
  3. generation_config.json +14 -0
  4. train_results.json +8 -0
  5. trainer_state.json +1683 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: Qwen/Qwen2.5-7B-Instruct
3
+ library_name: transformers
4
+ model_name: Qwen2.5-7B-Open-R1-Distill-Italian
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for Qwen2.5-7B-Open-R1-Distill-Italian
13
+
14
+ This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct).
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="bezir/Qwen2.5-7B-Open-R1-Distill-Italian", 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
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/bezirglasgow/huggingface/runs/a4k6nluk)
31
+
32
+
33
+ This model was trained with SFT.
34
+
35
+ ### Framework versions
36
+
37
+ - TRL: 0.16.0.dev0
38
+ - Transformers: 4.49.0
39
+ - Pytorch: 2.5.1
40
+ - Datasets: 3.3.2
41
+ - Tokenizers: 0.21.0
42
+
43
+ ## Citations
44
+
45
+
46
+
47
+ Cite TRL as:
48
+
49
+ ```bibtex
50
+ @misc{vonwerra2022trl,
51
+ title = {{TRL: Transformer Reinforcement Learning}},
52
+ 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},
53
+ year = 2020,
54
+ journal = {GitHub repository},
55
+ publisher = {GitHub},
56
+ howpublished = {\url{https://github.com/huggingface/trl}}
57
+ }
58
+ ```
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 1074983740637184.0,
3
+ "train_loss": 0.49954891321022377,
4
+ "train_runtime": 168401.8035,
5
+ "train_samples": 104697,
6
+ "train_samples_per_second": 0.78,
7
+ "train_steps_per_second": 0.006
8
+ }
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.05,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "4.49.0"
14
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 1074983740637184.0,
3
+ "train_loss": 0.49954891321022377,
4
+ "train_runtime": 168401.8035,
5
+ "train_samples": 104697,
6
+ "train_samples_per_second": 0.78,
7
+ "train_steps_per_second": 0.006
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1683 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.9973216459702945,
5
+ "eval_steps": 100,
6
+ "global_step": 1026,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.009739469198928659,
13
+ "grad_norm": 5.16649192443276,
14
+ "learning_rate": 4.854368932038835e-07,
15
+ "loss": 0.956,
16
+ "mean_token_accuracy": 0.7770209729671478,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.019478938397857318,
21
+ "grad_norm": 4.487616874759018,
22
+ "learning_rate": 9.70873786407767e-07,
23
+ "loss": 0.9595,
24
+ "mean_token_accuracy": 0.774706457555294,
25
+ "step": 10
26
+ },
27
+ {
28
+ "epoch": 0.029218407596785977,
29
+ "grad_norm": 1.9918521253065147,
30
+ "learning_rate": 1.4563106796116506e-06,
31
+ "loss": 0.9087,
32
+ "mean_token_accuracy": 0.7794785097241401,
33
+ "step": 15
34
+ },
35
+ {
36
+ "epoch": 0.038957876795714635,
37
+ "grad_norm": 2.140741610293528,
38
+ "learning_rate": 1.941747572815534e-06,
39
+ "loss": 0.8516,
40
+ "mean_token_accuracy": 0.7845764443278312,
41
+ "step": 20
42
+ },
43
+ {
44
+ "epoch": 0.048697345994643294,
45
+ "grad_norm": 1.577104941303547,
46
+ "learning_rate": 2.427184466019418e-06,
47
+ "loss": 0.7835,
48
+ "mean_token_accuracy": 0.795551997423172,
49
+ "step": 25
50
+ },
51
+ {
52
+ "epoch": 0.05843681519357195,
53
+ "grad_norm": 0.9851513931833208,
54
+ "learning_rate": 2.912621359223301e-06,
55
+ "loss": 0.7536,
56
+ "mean_token_accuracy": 0.801626966893673,
57
+ "step": 30
58
+ },
59
+ {
60
+ "epoch": 0.0681762843925006,
61
+ "grad_norm": 0.7665830684245213,
62
+ "learning_rate": 3.398058252427185e-06,
63
+ "loss": 0.7207,
64
+ "mean_token_accuracy": 0.8073863789439202,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.07791575359142927,
69
+ "grad_norm": 0.6623786102965908,
70
+ "learning_rate": 3.883495145631068e-06,
71
+ "loss": 0.6976,
72
+ "mean_token_accuracy": 0.8124146014451981,
73
+ "step": 40
74
+ },
75
+ {
76
+ "epoch": 0.08765522279035792,
77
+ "grad_norm": 0.5409465013517523,
78
+ "learning_rate": 4.368932038834952e-06,
79
+ "loss": 0.6791,
80
+ "mean_token_accuracy": 0.8153241157531739,
81
+ "step": 45
82
+ },
83
+ {
84
+ "epoch": 0.09739469198928659,
85
+ "grad_norm": 0.533911697336328,
86
+ "learning_rate": 4.854368932038836e-06,
87
+ "loss": 0.6494,
88
+ "mean_token_accuracy": 0.8212776482105255,
89
+ "step": 50
90
+ },
91
+ {
92
+ "epoch": 0.10713416118821524,
93
+ "grad_norm": 0.46335668910229383,
94
+ "learning_rate": 5.3398058252427185e-06,
95
+ "loss": 0.6374,
96
+ "mean_token_accuracy": 0.8242872759699822,
97
+ "step": 55
98
+ },
99
+ {
100
+ "epoch": 0.1168736303871439,
101
+ "grad_norm": 0.48547854119926936,
102
+ "learning_rate": 5.825242718446602e-06,
103
+ "loss": 0.6321,
104
+ "mean_token_accuracy": 0.8252887204289436,
105
+ "step": 60
106
+ },
107
+ {
108
+ "epoch": 0.12661309958607256,
109
+ "grad_norm": 0.4864652293922326,
110
+ "learning_rate": 6.310679611650487e-06,
111
+ "loss": 0.6246,
112
+ "mean_token_accuracy": 0.8261281028389931,
113
+ "step": 65
114
+ },
115
+ {
116
+ "epoch": 0.1363525687850012,
117
+ "grad_norm": 0.48690138801181443,
118
+ "learning_rate": 6.79611650485437e-06,
119
+ "loss": 0.6168,
120
+ "mean_token_accuracy": 0.8278339207172394,
121
+ "step": 70
122
+ },
123
+ {
124
+ "epoch": 0.14609203798392986,
125
+ "grad_norm": 0.4337713878880355,
126
+ "learning_rate": 7.2815533980582534e-06,
127
+ "loss": 0.5925,
128
+ "mean_token_accuracy": 0.8333460614085197,
129
+ "step": 75
130
+ },
131
+ {
132
+ "epoch": 0.15583150718285854,
133
+ "grad_norm": 0.45660620623082415,
134
+ "learning_rate": 7.766990291262136e-06,
135
+ "loss": 0.5973,
136
+ "mean_token_accuracy": 0.8320671111345291,
137
+ "step": 80
138
+ },
139
+ {
140
+ "epoch": 0.1655709763817872,
141
+ "grad_norm": 0.4424374164027964,
142
+ "learning_rate": 8.25242718446602e-06,
143
+ "loss": 0.5873,
144
+ "mean_token_accuracy": 0.8338468298316002,
145
+ "step": 85
146
+ },
147
+ {
148
+ "epoch": 0.17531044558071585,
149
+ "grad_norm": 0.4949541791856808,
150
+ "learning_rate": 8.737864077669904e-06,
151
+ "loss": 0.5846,
152
+ "mean_token_accuracy": 0.8346328064799309,
153
+ "step": 90
154
+ },
155
+ {
156
+ "epoch": 0.1850499147796445,
157
+ "grad_norm": 0.44716595203588544,
158
+ "learning_rate": 9.223300970873788e-06,
159
+ "loss": 0.574,
160
+ "mean_token_accuracy": 0.8365551233291626,
161
+ "step": 95
162
+ },
163
+ {
164
+ "epoch": 0.19478938397857318,
165
+ "grad_norm": 0.40922461158607365,
166
+ "learning_rate": 9.708737864077671e-06,
167
+ "loss": 0.5745,
168
+ "mean_token_accuracy": 0.8366570115089417,
169
+ "step": 100
170
+ },
171
+ {
172
+ "epoch": 0.20452885317750183,
173
+ "grad_norm": 0.5016773528604923,
174
+ "learning_rate": 9.99988415036596e-06,
175
+ "loss": 0.5649,
176
+ "mean_token_accuracy": 0.8382768034934998,
177
+ "step": 105
178
+ },
179
+ {
180
+ "epoch": 0.21426832237643048,
181
+ "grad_norm": 0.5010899629730204,
182
+ "learning_rate": 9.99858090363555e-06,
183
+ "loss": 0.569,
184
+ "mean_token_accuracy": 0.8377967774868011,
185
+ "step": 110
186
+ },
187
+ {
188
+ "epoch": 0.22400779157535913,
189
+ "grad_norm": 0.5010895934649102,
190
+ "learning_rate": 9.995829976834402e-06,
191
+ "loss": 0.5654,
192
+ "mean_token_accuracy": 0.8383171066641808,
193
+ "step": 115
194
+ },
195
+ {
196
+ "epoch": 0.2337472607742878,
197
+ "grad_norm": 0.5864787846174829,
198
+ "learning_rate": 9.99163216668102e-06,
199
+ "loss": 0.5703,
200
+ "mean_token_accuracy": 0.8369988009333611,
201
+ "step": 120
202
+ },
203
+ {
204
+ "epoch": 0.24348672997321646,
205
+ "grad_norm": 0.5502383423638442,
206
+ "learning_rate": 9.985988688937684e-06,
207
+ "loss": 0.5632,
208
+ "mean_token_accuracy": 0.8387595832347869,
209
+ "step": 125
210
+ },
211
+ {
212
+ "epoch": 0.2532261991721451,
213
+ "grad_norm": 0.4838385873501471,
214
+ "learning_rate": 9.978901178058333e-06,
215
+ "loss": 0.5472,
216
+ "mean_token_accuracy": 0.8424718379974365,
217
+ "step": 130
218
+ },
219
+ {
220
+ "epoch": 0.26296566837107377,
221
+ "grad_norm": 0.5371306966614764,
222
+ "learning_rate": 9.970371686715205e-06,
223
+ "loss": 0.5431,
224
+ "mean_token_accuracy": 0.8432716697454452,
225
+ "step": 135
226
+ },
227
+ {
228
+ "epoch": 0.2727051375700024,
229
+ "grad_norm": 0.5584506793424302,
230
+ "learning_rate": 9.960402685204347e-06,
231
+ "loss": 0.5516,
232
+ "mean_token_accuracy": 0.8408507108688354,
233
+ "step": 140
234
+ },
235
+ {
236
+ "epoch": 0.2824446067689311,
237
+ "grad_norm": 0.5467853450677332,
238
+ "learning_rate": 9.948997060730161e-06,
239
+ "loss": 0.5464,
240
+ "mean_token_accuracy": 0.8424971371889114,
241
+ "step": 145
242
+ },
243
+ {
244
+ "epoch": 0.2921840759678597,
245
+ "grad_norm": 0.48818575924287444,
246
+ "learning_rate": 9.936158116569231e-06,
247
+ "loss": 0.5489,
248
+ "mean_token_accuracy": 0.8420207649469376,
249
+ "step": 150
250
+ },
251
+ {
252
+ "epoch": 0.30192354516678843,
253
+ "grad_norm": 0.46723378629583207,
254
+ "learning_rate": 9.921889571113629e-06,
255
+ "loss": 0.5462,
256
+ "mean_token_accuracy": 0.8419363871216774,
257
+ "step": 155
258
+ },
259
+ {
260
+ "epoch": 0.3116630143657171,
261
+ "grad_norm": 0.561719955013033,
262
+ "learning_rate": 9.906195556793996e-06,
263
+ "loss": 0.546,
264
+ "mean_token_accuracy": 0.8423524782061577,
265
+ "step": 160
266
+ },
267
+ {
268
+ "epoch": 0.32140248356464574,
269
+ "grad_norm": 0.625160714021079,
270
+ "learning_rate": 9.889080618882719e-06,
271
+ "loss": 0.5335,
272
+ "mean_token_accuracy": 0.8451830595731735,
273
+ "step": 165
274
+ },
275
+ {
276
+ "epoch": 0.3311419527635744,
277
+ "grad_norm": 0.5878540451317104,
278
+ "learning_rate": 9.870549714177538e-06,
279
+ "loss": 0.5463,
280
+ "mean_token_accuracy": 0.8417419150471688,
281
+ "step": 170
282
+ },
283
+ {
284
+ "epoch": 0.34088142196250304,
285
+ "grad_norm": 0.6154889939045929,
286
+ "learning_rate": 9.850608209565967e-06,
287
+ "loss": 0.5327,
288
+ "mean_token_accuracy": 0.8450453072786331,
289
+ "step": 175
290
+ },
291
+ {
292
+ "epoch": 0.3506208911614317,
293
+ "grad_norm": 0.4329465610848603,
294
+ "learning_rate": 9.829261880470941e-06,
295
+ "loss": 0.5392,
296
+ "mean_token_accuracy": 0.8434989348053932,
297
+ "step": 180
298
+ },
299
+ {
300
+ "epoch": 0.36036036036036034,
301
+ "grad_norm": 0.5971707984256834,
302
+ "learning_rate": 9.806516909178161e-06,
303
+ "loss": 0.5324,
304
+ "mean_token_accuracy": 0.84499292075634,
305
+ "step": 185
306
+ },
307
+ {
308
+ "epoch": 0.370099829559289,
309
+ "grad_norm": 0.44190322710346147,
310
+ "learning_rate": 9.78237988304557e-06,
311
+ "loss": 0.5332,
312
+ "mean_token_accuracy": 0.845010556280613,
313
+ "step": 190
314
+ },
315
+ {
316
+ "epoch": 0.3798392987582177,
317
+ "grad_norm": 0.4258724885095506,
318
+ "learning_rate": 9.756857792595555e-06,
319
+ "loss": 0.5319,
320
+ "mean_token_accuracy": 0.845000034570694,
321
+ "step": 195
322
+ },
323
+ {
324
+ "epoch": 0.38957876795714635,
325
+ "grad_norm": 0.4729261486427116,
326
+ "learning_rate": 9.729958029490353e-06,
327
+ "loss": 0.5336,
328
+ "mean_token_accuracy": 0.8447421163320541,
329
+ "step": 200
330
+ },
331
+ {
332
+ "epoch": 0.399318237156075,
333
+ "grad_norm": 0.4528083565277609,
334
+ "learning_rate": 9.701688384391296e-06,
335
+ "loss": 0.5347,
336
+ "mean_token_accuracy": 0.8443498685956001,
337
+ "step": 205
338
+ },
339
+ {
340
+ "epoch": 0.40905770635500366,
341
+ "grad_norm": 0.5259149339810829,
342
+ "learning_rate": 9.672057044702492e-06,
343
+ "loss": 0.5199,
344
+ "mean_token_accuracy": 0.848120279610157,
345
+ "step": 210
346
+ },
347
+ {
348
+ "epoch": 0.4187971755539323,
349
+ "grad_norm": 0.46836949308672154,
350
+ "learning_rate": 9.641072592199599e-06,
351
+ "loss": 0.5219,
352
+ "mean_token_accuracy": 0.8473641723394394,
353
+ "step": 215
354
+ },
355
+ {
356
+ "epoch": 0.42853664475286096,
357
+ "grad_norm": 0.5302502163183013,
358
+ "learning_rate": 9.608744000544392e-06,
359
+ "loss": 0.5174,
360
+ "mean_token_accuracy": 0.8485160410404206,
361
+ "step": 220
362
+ },
363
+ {
364
+ "epoch": 0.4382761139517896,
365
+ "grad_norm": 0.51343914785562,
366
+ "learning_rate": 9.575080632685832e-06,
367
+ "loss": 0.5239,
368
+ "mean_token_accuracy": 0.846913392841816,
369
+ "step": 225
370
+ },
371
+ {
372
+ "epoch": 0.44801558315071827,
373
+ "grad_norm": 0.5925559680424145,
374
+ "learning_rate": 9.54009223814837e-06,
375
+ "loss": 0.5277,
376
+ "mean_token_accuracy": 0.8454070091247559,
377
+ "step": 230
378
+ },
379
+ {
380
+ "epoch": 0.4577550523496469,
381
+ "grad_norm": 0.426110283810703,
382
+ "learning_rate": 9.503788950208324e-06,
383
+ "loss": 0.5215,
384
+ "mean_token_accuracy": 0.8473676040768623,
385
+ "step": 235
386
+ },
387
+ {
388
+ "epoch": 0.4674945215485756,
389
+ "grad_norm": 0.5303378941560718,
390
+ "learning_rate": 9.466181282959083e-06,
391
+ "loss": 0.5282,
392
+ "mean_token_accuracy": 0.8465007901191711,
393
+ "step": 240
394
+ },
395
+ {
396
+ "epoch": 0.4772339907475043,
397
+ "grad_norm": 0.4965315113021109,
398
+ "learning_rate": 9.427280128266049e-06,
399
+ "loss": 0.5179,
400
+ "mean_token_accuracy": 0.8484022691845894,
401
+ "step": 245
402
+ },
403
+ {
404
+ "epoch": 0.48697345994643293,
405
+ "grad_norm": 0.44982624201366644,
406
+ "learning_rate": 9.387096752612144e-06,
407
+ "loss": 0.5224,
408
+ "mean_token_accuracy": 0.8463509559631348,
409
+ "step": 250
410
+ },
411
+ {
412
+ "epoch": 0.4967129291453616,
413
+ "grad_norm": 0.4623459721033935,
414
+ "learning_rate": 9.345642793834825e-06,
415
+ "loss": 0.5271,
416
+ "mean_token_accuracy": 0.8463737353682518,
417
+ "step": 255
418
+ },
419
+ {
420
+ "epoch": 0.5064523983442902,
421
+ "grad_norm": 0.5825923961149249,
422
+ "learning_rate": 9.302930257755579e-06,
423
+ "loss": 0.53,
424
+ "mean_token_accuracy": 0.8450088694691658,
425
+ "step": 260
426
+ },
427
+ {
428
+ "epoch": 0.5161918675432189,
429
+ "grad_norm": 0.45940800329264964,
430
+ "learning_rate": 9.258971514702789e-06,
431
+ "loss": 0.507,
432
+ "mean_token_accuracy": 0.8508246764540672,
433
+ "step": 265
434
+ },
435
+ {
436
+ "epoch": 0.5259313367421475,
437
+ "grad_norm": 0.4465364801256697,
438
+ "learning_rate": 9.213779295929082e-06,
439
+ "loss": 0.5087,
440
+ "mean_token_accuracy": 0.8500014141201973,
441
+ "step": 270
442
+ },
443
+ {
444
+ "epoch": 0.5356708059410762,
445
+ "grad_norm": 0.4686695606038613,
446
+ "learning_rate": 9.167366689924116e-06,
447
+ "loss": 0.5163,
448
+ "mean_token_accuracy": 0.8484609499573708,
449
+ "step": 275
450
+ },
451
+ {
452
+ "epoch": 0.5454102751400048,
453
+ "grad_norm": 0.5572145736502692,
454
+ "learning_rate": 9.119747138623925e-06,
455
+ "loss": 0.5221,
456
+ "mean_token_accuracy": 0.8470426678657532,
457
+ "step": 280
458
+ },
459
+ {
460
+ "epoch": 0.5551497443389335,
461
+ "grad_norm": 0.4765827682250088,
462
+ "learning_rate": 9.070934433517872e-06,
463
+ "loss": 0.5068,
464
+ "mean_token_accuracy": 0.8509115263819694,
465
+ "step": 285
466
+ },
467
+ {
468
+ "epoch": 0.5648892135378621,
469
+ "grad_norm": 0.49934327207360063,
470
+ "learning_rate": 9.020942711654404e-06,
471
+ "loss": 0.5106,
472
+ "mean_token_accuracy": 0.8498208403587342,
473
+ "step": 290
474
+ },
475
+ {
476
+ "epoch": 0.5746286827367908,
477
+ "grad_norm": 0.5101324493724546,
478
+ "learning_rate": 8.969786451546691e-06,
479
+ "loss": 0.5123,
480
+ "mean_token_accuracy": 0.8496938437223435,
481
+ "step": 295
482
+ },
483
+ {
484
+ "epoch": 0.5843681519357194,
485
+ "grad_norm": 0.5089021103596151,
486
+ "learning_rate": 8.917480468979387e-06,
487
+ "loss": 0.5128,
488
+ "mean_token_accuracy": 0.8487787261605263,
489
+ "step": 300
490
+ },
491
+ {
492
+ "epoch": 0.5941076211346482,
493
+ "grad_norm": 0.5501775930667125,
494
+ "learning_rate": 8.864039912717713e-06,
495
+ "loss": 0.5123,
496
+ "mean_token_accuracy": 0.849444879591465,
497
+ "step": 305
498
+ },
499
+ {
500
+ "epoch": 0.6038470903335769,
501
+ "grad_norm": 0.41563520519489466,
502
+ "learning_rate": 8.809480260120096e-06,
503
+ "loss": 0.5048,
504
+ "mean_token_accuracy": 0.8513683333992959,
505
+ "step": 310
506
+ },
507
+ {
508
+ "epoch": 0.6135865595325055,
509
+ "grad_norm": 0.4999791105897695,
510
+ "learning_rate": 8.753817312655642e-06,
511
+ "loss": 0.514,
512
+ "mean_token_accuracy": 0.8484693005681038,
513
+ "step": 315
514
+ },
515
+ {
516
+ "epoch": 0.6233260287314342,
517
+ "grad_norm": 0.48088239973601826,
518
+ "learning_rate": 8.697067191327748e-06,
519
+ "loss": 0.5114,
520
+ "mean_token_accuracy": 0.8495015501976013,
521
+ "step": 320
522
+ },
523
+ {
524
+ "epoch": 0.6330654979303628,
525
+ "grad_norm": 0.4440141428026334,
526
+ "learning_rate": 8.639246332005163e-06,
527
+ "loss": 0.5064,
528
+ "mean_token_accuracy": 0.8507678374648094,
529
+ "step": 325
530
+ },
531
+ {
532
+ "epoch": 0.6428049671292915,
533
+ "grad_norm": 0.5288324873519791,
534
+ "learning_rate": 8.580371480661857e-06,
535
+ "loss": 0.5024,
536
+ "mean_token_accuracy": 0.8514497712254524,
537
+ "step": 330
538
+ },
539
+ {
540
+ "epoch": 0.6525444363282201,
541
+ "grad_norm": 0.4563934748637001,
542
+ "learning_rate": 8.520459688527091e-06,
543
+ "loss": 0.5108,
544
+ "mean_token_accuracy": 0.849525935947895,
545
+ "step": 335
546
+ },
547
+ {
548
+ "epoch": 0.6622839055271488,
549
+ "grad_norm": 0.4756063591717307,
550
+ "learning_rate": 8.459528307147066e-06,
551
+ "loss": 0.51,
552
+ "mean_token_accuracy": 0.8501726359128952,
553
+ "step": 340
554
+ },
555
+ {
556
+ "epoch": 0.6720233747260774,
557
+ "grad_norm": 0.5732019741491505,
558
+ "learning_rate": 8.397594983359591e-06,
559
+ "loss": 0.5062,
560
+ "mean_token_accuracy": 0.850397090613842,
561
+ "step": 345
562
+ },
563
+ {
564
+ "epoch": 0.6817628439250061,
565
+ "grad_norm": 0.5267653966986828,
566
+ "learning_rate": 8.334677654183254e-06,
567
+ "loss": 0.5065,
568
+ "mean_token_accuracy": 0.8505988359451294,
569
+ "step": 350
570
+ },
571
+ {
572
+ "epoch": 0.6915023131239347,
573
+ "grad_norm": 0.4796149747636043,
574
+ "learning_rate": 8.27079454162252e-06,
575
+ "loss": 0.5028,
576
+ "mean_token_accuracy": 0.851087860763073,
577
+ "step": 355
578
+ },
579
+ {
580
+ "epoch": 0.7012417823228634,
581
+ "grad_norm": 0.4809967856664383,
582
+ "learning_rate": 8.205964147390313e-06,
583
+ "loss": 0.5084,
584
+ "mean_token_accuracy": 0.8496400877833367,
585
+ "step": 360
586
+ },
587
+ {
588
+ "epoch": 0.710981251521792,
589
+ "grad_norm": 0.4348534092140185,
590
+ "learning_rate": 8.140205247549583e-06,
591
+ "loss": 0.4983,
592
+ "mean_token_accuracy": 0.8522587567567825,
593
+ "step": 365
594
+ },
595
+ {
596
+ "epoch": 0.7207207207207207,
597
+ "grad_norm": 0.4799630529257426,
598
+ "learning_rate": 8.073536887075417e-06,
599
+ "loss": 0.5119,
600
+ "mean_token_accuracy": 0.8493492469191551,
601
+ "step": 370
602
+ },
603
+ {
604
+ "epoch": 0.7304601899196493,
605
+ "grad_norm": 0.5305052950944196,
606
+ "learning_rate": 8.005978374339264e-06,
607
+ "loss": 0.4946,
608
+ "mean_token_accuracy": 0.8531364649534225,
609
+ "step": 375
610
+ },
611
+ {
612
+ "epoch": 0.740199659118578,
613
+ "grad_norm": 0.46401531374429855,
614
+ "learning_rate": 7.937549275516882e-06,
615
+ "loss": 0.493,
616
+ "mean_token_accuracy": 0.8535278528928757,
617
+ "step": 380
618
+ },
619
+ {
620
+ "epoch": 0.7499391283175066,
621
+ "grad_norm": 0.46753323308197925,
622
+ "learning_rate": 7.868269408921614e-06,
623
+ "loss": 0.504,
624
+ "mean_token_accuracy": 0.8511819407343865,
625
+ "step": 385
626
+ },
627
+ {
628
+ "epoch": 0.7596785975164354,
629
+ "grad_norm": 0.43897436674183826,
630
+ "learning_rate": 7.798158839264645e-06,
631
+ "loss": 0.4983,
632
+ "mean_token_accuracy": 0.8521765768527985,
633
+ "step": 390
634
+ },
635
+ {
636
+ "epoch": 0.7694180667153641,
637
+ "grad_norm": 0.4030095081507055,
638
+ "learning_rate": 7.7272378718439e-06,
639
+ "loss": 0.5092,
640
+ "mean_token_accuracy": 0.8500049978494644,
641
+ "step": 395
642
+ },
643
+ {
644
+ "epoch": 0.7791575359142927,
645
+ "grad_norm": 0.393465790048579,
646
+ "learning_rate": 7.655527046663254e-06,
647
+ "loss": 0.5022,
648
+ "mean_token_accuracy": 0.8510702222585678,
649
+ "step": 400
650
+ },
651
+ {
652
+ "epoch": 0.7888970051132214,
653
+ "grad_norm": 0.49269595855244275,
654
+ "learning_rate": 7.5830471324837765e-06,
655
+ "loss": 0.4945,
656
+ "mean_token_accuracy": 0.8529314771294594,
657
+ "step": 405
658
+ },
659
+ {
660
+ "epoch": 0.79863647431215,
661
+ "grad_norm": 0.4729001525106615,
662
+ "learning_rate": 7.5098191208087144e-06,
663
+ "loss": 0.5,
664
+ "mean_token_accuracy": 0.8525989070534706,
665
+ "step": 410
666
+ },
667
+ {
668
+ "epoch": 0.8083759435110787,
669
+ "grad_norm": 0.44720074409020516,
670
+ "learning_rate": 7.4358642198039835e-06,
671
+ "loss": 0.4946,
672
+ "mean_token_accuracy": 0.8530281245708465,
673
+ "step": 415
674
+ },
675
+ {
676
+ "epoch": 0.8181154127100073,
677
+ "grad_norm": 0.4192773521625985,
678
+ "learning_rate": 7.36120384815588e-06,
679
+ "loss": 0.4927,
680
+ "mean_token_accuracy": 0.8539465010166168,
681
+ "step": 420
682
+ },
683
+ {
684
+ "epoch": 0.827854881908936,
685
+ "grad_norm": 0.41037683317185314,
686
+ "learning_rate": 7.285859628867851e-06,
687
+ "loss": 0.4952,
688
+ "mean_token_accuracy": 0.8532393842935562,
689
+ "step": 425
690
+ },
691
+ {
692
+ "epoch": 0.8375943511078646,
693
+ "grad_norm": 0.4536782151813187,
694
+ "learning_rate": 7.209853382998077e-06,
695
+ "loss": 0.4983,
696
+ "mean_token_accuracy": 0.8528945103287697,
697
+ "step": 430
698
+ },
699
+ {
700
+ "epoch": 0.8473338203067933,
701
+ "grad_norm": 0.4895833290473578,
702
+ "learning_rate": 7.133207123339689e-06,
703
+ "loss": 0.4939,
704
+ "mean_token_accuracy": 0.853129243850708,
705
+ "step": 435
706
+ },
707
+ {
708
+ "epoch": 0.8570732895057219,
709
+ "grad_norm": 0.4771689532642086,
710
+ "learning_rate": 7.055943048045476e-06,
711
+ "loss": 0.5002,
712
+ "mean_token_accuracy": 0.8518661975860595,
713
+ "step": 440
714
+ },
715
+ {
716
+ "epoch": 0.8668127587046506,
717
+ "grad_norm": 0.4439177979495324,
718
+ "learning_rate": 6.978083534198878e-06,
719
+ "loss": 0.4977,
720
+ "mean_token_accuracy": 0.8526063248515129,
721
+ "step": 445
722
+ },
723
+ {
724
+ "epoch": 0.8765522279035792,
725
+ "grad_norm": 0.45795986197220584,
726
+ "learning_rate": 6.899651131333194e-06,
727
+ "loss": 0.4876,
728
+ "mean_token_accuracy": 0.8546572834253311,
729
+ "step": 450
730
+ },
731
+ {
732
+ "epoch": 0.8862916971025079,
733
+ "grad_norm": 0.38084088770564956,
734
+ "learning_rate": 6.82066855490081e-06,
735
+ "loss": 0.5009,
736
+ "mean_token_accuracy": 0.851148933172226,
737
+ "step": 455
738
+ },
739
+ {
740
+ "epoch": 0.8960311663014365,
741
+ "grad_norm": 0.4552737931829716,
742
+ "learning_rate": 6.741158679694403e-06,
743
+ "loss": 0.4968,
744
+ "mean_token_accuracy": 0.8524380102753639,
745
+ "step": 460
746
+ },
747
+ {
748
+ "epoch": 0.9057706355003652,
749
+ "grad_norm": 0.38831398930391275,
750
+ "learning_rate": 6.661144533221974e-06,
751
+ "loss": 0.4897,
752
+ "mean_token_accuracy": 0.8537176489830017,
753
+ "step": 465
754
+ },
755
+ {
756
+ "epoch": 0.9155101046992938,
757
+ "grad_norm": 0.42794585960634596,
758
+ "learning_rate": 6.58064928903767e-06,
759
+ "loss": 0.4942,
760
+ "mean_token_accuracy": 0.8529506504535675,
761
+ "step": 470
762
+ },
763
+ {
764
+ "epoch": 0.9252495738982226,
765
+ "grad_norm": 0.42890975667323755,
766
+ "learning_rate": 6.499696260030297e-06,
767
+ "loss": 0.5064,
768
+ "mean_token_accuracy": 0.8502799227833748,
769
+ "step": 475
770
+ },
771
+ {
772
+ "epoch": 0.9349890430971513,
773
+ "grad_norm": 0.4478332747744812,
774
+ "learning_rate": 6.418308891671484e-06,
775
+ "loss": 0.4855,
776
+ "mean_token_accuracy": 0.8555838361382484,
777
+ "step": 480
778
+ },
779
+ {
780
+ "epoch": 0.9447285122960799,
781
+ "grad_norm": 0.435211802978247,
782
+ "learning_rate": 6.336510755225447e-06,
783
+ "loss": 0.4835,
784
+ "mean_token_accuracy": 0.8558964654803276,
785
+ "step": 485
786
+ },
787
+ {
788
+ "epoch": 0.9544679814950086,
789
+ "grad_norm": 0.4462726381437146,
790
+ "learning_rate": 6.25432554092232e-06,
791
+ "loss": 0.4898,
792
+ "mean_token_accuracy": 0.8544702440500259,
793
+ "step": 490
794
+ },
795
+ {
796
+ "epoch": 0.9642074506939372,
797
+ "grad_norm": 0.44503915866443544,
798
+ "learning_rate": 6.171777051097037e-06,
799
+ "loss": 0.4858,
800
+ "mean_token_accuracy": 0.8554443955421448,
801
+ "step": 495
802
+ },
803
+ {
804
+ "epoch": 0.9739469198928659,
805
+ "grad_norm": 0.47966098493432496,
806
+ "learning_rate": 6.088889193295738e-06,
807
+ "loss": 0.4929,
808
+ "mean_token_accuracy": 0.8535514727234841,
809
+ "step": 500
810
+ },
811
+ {
812
+ "epoch": 0.9836863890917945,
813
+ "grad_norm": 0.5437585793823457,
814
+ "learning_rate": 6.005685973351708e-06,
815
+ "loss": 0.4931,
816
+ "mean_token_accuracy": 0.8531181156635285,
817
+ "step": 505
818
+ },
819
+ {
820
+ "epoch": 0.9934258582907232,
821
+ "grad_norm": 0.4124694681433172,
822
+ "learning_rate": 5.922191488432857e-06,
823
+ "loss": 0.4917,
824
+ "mean_token_accuracy": 0.8535135626792908,
825
+ "step": 510
826
+ },
827
+ {
828
+ "epoch": 1.0019478938397857,
829
+ "grad_norm": 0.803029283965446,
830
+ "learning_rate": 5.838429920062734e-06,
831
+ "loss": 0.4727,
832
+ "mean_token_accuracy": 0.8567549926894051,
833
+ "step": 515
834
+ },
835
+ {
836
+ "epoch": 1.0116873630387144,
837
+ "grad_norm": 0.3927635667112885,
838
+ "learning_rate": 5.754425527117118e-06,
839
+ "loss": 0.4479,
840
+ "mean_token_accuracy": 0.864441742002964,
841
+ "step": 520
842
+ },
843
+ {
844
+ "epoch": 1.021426832237643,
845
+ "grad_norm": 0.45928387456979786,
846
+ "learning_rate": 5.670202638798213e-06,
847
+ "loss": 0.4598,
848
+ "mean_token_accuracy": 0.8615871027112008,
849
+ "step": 525
850
+ },
851
+ {
852
+ "epoch": 1.0311663014365717,
853
+ "grad_norm": 0.517389580126703,
854
+ "learning_rate": 5.585785647588458e-06,
855
+ "loss": 0.4572,
856
+ "mean_token_accuracy": 0.8620010375976562,
857
+ "step": 530
858
+ },
859
+ {
860
+ "epoch": 1.0409057706355003,
861
+ "grad_norm": 0.37886924917283415,
862
+ "learning_rate": 5.501199002186024e-06,
863
+ "loss": 0.455,
864
+ "mean_token_accuracy": 0.8625968441367149,
865
+ "step": 535
866
+ },
867
+ {
868
+ "epoch": 1.050645239834429,
869
+ "grad_norm": 0.39773153996958316,
870
+ "learning_rate": 5.416467200424032e-06,
871
+ "loss": 0.45,
872
+ "mean_token_accuracy": 0.8637859463691712,
873
+ "step": 540
874
+ },
875
+ {
876
+ "epoch": 1.0603847090333576,
877
+ "grad_norm": 0.40508395528042257,
878
+ "learning_rate": 5.33161478217552e-06,
879
+ "loss": 0.4516,
880
+ "mean_token_accuracy": 0.8634377360343933,
881
+ "step": 545
882
+ },
883
+ {
884
+ "epoch": 1.0701241782322863,
885
+ "grad_norm": 0.37429603209812456,
886
+ "learning_rate": 5.246666322246267e-06,
887
+ "loss": 0.4445,
888
+ "mean_token_accuracy": 0.8651037693023682,
889
+ "step": 550
890
+ },
891
+ {
892
+ "epoch": 1.079863647431215,
893
+ "grad_norm": 0.47125189912053295,
894
+ "learning_rate": 5.1616464232574635e-06,
895
+ "loss": 0.4626,
896
+ "mean_token_accuracy": 0.8602706581354141,
897
+ "step": 555
898
+ },
899
+ {
900
+ "epoch": 1.0896031166301436,
901
+ "grad_norm": 0.435383825468146,
902
+ "learning_rate": 5.076579708520355e-06,
903
+ "loss": 0.4497,
904
+ "mean_token_accuracy": 0.8635805040597916,
905
+ "step": 560
906
+ },
907
+ {
908
+ "epoch": 1.0993425858290724,
909
+ "grad_norm": 0.412421290706311,
910
+ "learning_rate": 4.991490814904888e-06,
911
+ "loss": 0.4378,
912
+ "mean_token_accuracy": 0.8668369174003601,
913
+ "step": 565
914
+ },
915
+ {
916
+ "epoch": 1.109082055028001,
917
+ "grad_norm": 0.41209358309681204,
918
+ "learning_rate": 4.906404385704402e-06,
919
+ "loss": 0.4525,
920
+ "mean_token_accuracy": 0.8637306377291679,
921
+ "step": 570
922
+ },
923
+ {
924
+ "epoch": 1.1188215242269297,
925
+ "grad_norm": 0.3821281290056491,
926
+ "learning_rate": 4.82134506349851e-06,
927
+ "loss": 0.4564,
928
+ "mean_token_accuracy": 0.8625956058502198,
929
+ "step": 575
930
+ },
931
+ {
932
+ "epoch": 1.1285609934258582,
933
+ "grad_norm": 0.34980242368851916,
934
+ "learning_rate": 4.736337483016138e-06,
935
+ "loss": 0.4513,
936
+ "mean_token_accuracy": 0.8634121060371399,
937
+ "step": 580
938
+ },
939
+ {
940
+ "epoch": 1.138300462624787,
941
+ "grad_norm": 0.39297932611221925,
942
+ "learning_rate": 4.651406264000871e-06,
943
+ "loss": 0.4512,
944
+ "mean_token_accuracy": 0.8632175624370575,
945
+ "step": 585
946
+ },
947
+ {
948
+ "epoch": 1.1480399318237156,
949
+ "grad_norm": 0.4317512663581136,
950
+ "learning_rate": 4.5665760040806174e-06,
951
+ "loss": 0.4558,
952
+ "mean_token_accuracy": 0.8621700823307037,
953
+ "step": 590
954
+ },
955
+ {
956
+ "epoch": 1.1577794010226443,
957
+ "grad_norm": 0.4166519413744608,
958
+ "learning_rate": 4.481871271643698e-06,
959
+ "loss": 0.4543,
960
+ "mean_token_accuracy": 0.8628953084349632,
961
+ "step": 595
962
+ },
963
+ {
964
+ "epoch": 1.1675188702215729,
965
+ "grad_norm": 0.3817664896689048,
966
+ "learning_rate": 4.397316598723385e-06,
967
+ "loss": 0.4599,
968
+ "mean_token_accuracy": 0.8614401906728745,
969
+ "step": 600
970
+ },
971
+ {
972
+ "epoch": 1.1772583394205016,
973
+ "grad_norm": 0.39223028227054235,
974
+ "learning_rate": 4.312936473892984e-06,
975
+ "loss": 0.4559,
976
+ "mean_token_accuracy": 0.8621881052851676,
977
+ "step": 605
978
+ },
979
+ {
980
+ "epoch": 1.1869978086194302,
981
+ "grad_norm": 0.39870192645434044,
982
+ "learning_rate": 4.228755335173488e-06,
983
+ "loss": 0.4554,
984
+ "mean_token_accuracy": 0.8622724115848541,
985
+ "step": 610
986
+ },
987
+ {
988
+ "epoch": 1.196737277818359,
989
+ "grad_norm": 0.3968827579618214,
990
+ "learning_rate": 4.1447975629559e-06,
991
+ "loss": 0.4496,
992
+ "mean_token_accuracy": 0.8639340966939926,
993
+ "step": 615
994
+ },
995
+ {
996
+ "epoch": 1.2064767470172875,
997
+ "grad_norm": 0.3525974840125754,
998
+ "learning_rate": 4.061087472940204e-06,
999
+ "loss": 0.4468,
1000
+ "mean_token_accuracy": 0.8643500834703446,
1001
+ "step": 620
1002
+ },
1003
+ {
1004
+ "epoch": 1.2162162162162162,
1005
+ "grad_norm": 0.34784808537613615,
1006
+ "learning_rate": 3.977649309093113e-06,
1007
+ "loss": 0.4463,
1008
+ "mean_token_accuracy": 0.8645710095763206,
1009
+ "step": 625
1010
+ },
1011
+ {
1012
+ "epoch": 1.225955685415145,
1013
+ "grad_norm": 0.3588527258157974,
1014
+ "learning_rate": 3.89450723662657e-06,
1015
+ "loss": 0.4517,
1016
+ "mean_token_accuracy": 0.863483439385891,
1017
+ "step": 630
1018
+ },
1019
+ {
1020
+ "epoch": 1.2356951546140735,
1021
+ "grad_norm": 0.495886348965792,
1022
+ "learning_rate": 3.8116853349990574e-06,
1023
+ "loss": 0.4507,
1024
+ "mean_token_accuracy": 0.8636451244354248,
1025
+ "step": 635
1026
+ },
1027
+ {
1028
+ "epoch": 1.245434623813002,
1029
+ "grad_norm": 0.3842625748309975,
1030
+ "learning_rate": 3.729207590941753e-06,
1031
+ "loss": 0.4377,
1032
+ "mean_token_accuracy": 0.8670691177248955,
1033
+ "step": 640
1034
+ },
1035
+ {
1036
+ "epoch": 1.2551740930119308,
1037
+ "grad_norm": 0.4426936471853252,
1038
+ "learning_rate": 3.647097891511536e-06,
1039
+ "loss": 0.4495,
1040
+ "mean_token_accuracy": 0.8640359625220299,
1041
+ "step": 645
1042
+ },
1043
+ {
1044
+ "epoch": 1.2649135622108596,
1045
+ "grad_norm": 0.3707914127374137,
1046
+ "learning_rate": 3.565380017172854e-06,
1047
+ "loss": 0.4397,
1048
+ "mean_token_accuracy": 0.8666136890649796,
1049
+ "step": 650
1050
+ },
1051
+ {
1052
+ "epoch": 1.2746530314097881,
1053
+ "grad_norm": 0.35267532874755536,
1054
+ "learning_rate": 3.4840776349104755e-06,
1055
+ "loss": 0.4539,
1056
+ "mean_token_accuracy": 0.8626947477459908,
1057
+ "step": 655
1058
+ },
1059
+ {
1060
+ "epoch": 1.284392500608717,
1061
+ "grad_norm": 0.3749743227529099,
1062
+ "learning_rate": 3.4032142913750956e-06,
1063
+ "loss": 0.4497,
1064
+ "mean_token_accuracy": 0.8637418314814568,
1065
+ "step": 660
1066
+ },
1067
+ {
1068
+ "epoch": 1.2941319698076454,
1069
+ "grad_norm": 0.37600006002277636,
1070
+ "learning_rate": 3.322813406063794e-06,
1071
+ "loss": 0.4559,
1072
+ "mean_token_accuracy": 0.8622224271297455,
1073
+ "step": 665
1074
+ },
1075
+ {
1076
+ "epoch": 1.3038714390065742,
1077
+ "grad_norm": 0.4005804398308692,
1078
+ "learning_rate": 3.242898264537331e-06,
1079
+ "loss": 0.4521,
1080
+ "mean_token_accuracy": 0.8632362619042396,
1081
+ "step": 670
1082
+ },
1083
+ {
1084
+ "epoch": 1.3136109082055027,
1085
+ "grad_norm": 0.38630202996272256,
1086
+ "learning_rate": 3.1634920116762175e-06,
1087
+ "loss": 0.4499,
1088
+ "mean_token_accuracy": 0.8635998621582985,
1089
+ "step": 675
1090
+ },
1091
+ {
1092
+ "epoch": 1.3233503774044315,
1093
+ "grad_norm": 0.36180685239531096,
1094
+ "learning_rate": 3.0846176449775363e-06,
1095
+ "loss": 0.4508,
1096
+ "mean_token_accuracy": 0.8636634424328804,
1097
+ "step": 680
1098
+ },
1099
+ {
1100
+ "epoch": 1.33308984660336,
1101
+ "grad_norm": 0.3357931590954893,
1102
+ "learning_rate": 3.0062980078944515e-06,
1103
+ "loss": 0.4379,
1104
+ "mean_token_accuracy": 0.8665053129196167,
1105
+ "step": 685
1106
+ },
1107
+ {
1108
+ "epoch": 1.3428293158022888,
1109
+ "grad_norm": 0.35662293115272325,
1110
+ "learning_rate": 2.9285557832203328e-06,
1111
+ "loss": 0.4458,
1112
+ "mean_token_accuracy": 0.8648849859833717,
1113
+ "step": 690
1114
+ },
1115
+ {
1116
+ "epoch": 1.3525687850012174,
1117
+ "grad_norm": 0.35254913645932934,
1118
+ "learning_rate": 2.851413486519388e-06,
1119
+ "loss": 0.4413,
1120
+ "mean_token_accuracy": 0.8654858738183975,
1121
+ "step": 695
1122
+ },
1123
+ {
1124
+ "epoch": 1.3623082542001461,
1125
+ "grad_norm": 0.3448432687225989,
1126
+ "learning_rate": 2.774893459605766e-06,
1127
+ "loss": 0.4431,
1128
+ "mean_token_accuracy": 0.86555365473032,
1129
+ "step": 700
1130
+ },
1131
+ {
1132
+ "epoch": 1.3720477233990747,
1133
+ "grad_norm": 0.35405617501926967,
1134
+ "learning_rate": 2.69901786407295e-06,
1135
+ "loss": 0.444,
1136
+ "mean_token_accuracy": 0.8651208564639091,
1137
+ "step": 705
1138
+ },
1139
+ {
1140
+ "epoch": 1.3817871925980034,
1141
+ "grad_norm": 0.36859590132072323,
1142
+ "learning_rate": 2.6238086748753587e-06,
1143
+ "loss": 0.456,
1144
+ "mean_token_accuracy": 0.8627100110054016,
1145
+ "step": 710
1146
+ },
1147
+ {
1148
+ "epoch": 1.3915266617969322,
1149
+ "grad_norm": 0.35246470987862855,
1150
+ "learning_rate": 2.5492876739639912e-06,
1151
+ "loss": 0.4533,
1152
+ "mean_token_accuracy": 0.8628792524337768,
1153
+ "step": 715
1154
+ },
1155
+ {
1156
+ "epoch": 1.4012661309958607,
1157
+ "grad_norm": 0.36351033252725057,
1158
+ "learning_rate": 2.475476443977996e-06,
1159
+ "loss": 0.4469,
1160
+ "mean_token_accuracy": 0.8642540082335473,
1161
+ "step": 720
1162
+ },
1163
+ {
1164
+ "epoch": 1.4110056001947893,
1165
+ "grad_norm": 0.35938233389923585,
1166
+ "learning_rate": 2.40239636199393e-06,
1167
+ "loss": 0.4451,
1168
+ "mean_token_accuracy": 0.8650069192051888,
1169
+ "step": 725
1170
+ },
1171
+ {
1172
+ "epoch": 1.420745069393718,
1173
+ "grad_norm": 0.3673326064574297,
1174
+ "learning_rate": 2.3300685933345656e-06,
1175
+ "loss": 0.4485,
1176
+ "mean_token_accuracy": 0.8642319470643998,
1177
+ "step": 730
1178
+ },
1179
+ {
1180
+ "epoch": 1.4304845385926468,
1181
+ "grad_norm": 0.37166399663475924,
1182
+ "learning_rate": 2.2585140854390432e-06,
1183
+ "loss": 0.4496,
1184
+ "mean_token_accuracy": 0.8641144469380379,
1185
+ "step": 735
1186
+ },
1187
+ {
1188
+ "epoch": 1.4402240077915753,
1189
+ "grad_norm": 0.34969080725493995,
1190
+ "learning_rate": 2.187753561796097e-06,
1191
+ "loss": 0.449,
1192
+ "mean_token_accuracy": 0.8638374775648117,
1193
+ "step": 740
1194
+ },
1195
+ {
1196
+ "epoch": 1.449963476990504,
1197
+ "grad_norm": 0.33346605175419364,
1198
+ "learning_rate": 2.117807515942163e-06,
1199
+ "loss": 0.4487,
1200
+ "mean_token_accuracy": 0.8639461770653725,
1201
+ "step": 745
1202
+ },
1203
+ {
1204
+ "epoch": 1.4597029461894326,
1205
+ "grad_norm": 0.3402282068550279,
1206
+ "learning_rate": 2.0486962055260744e-06,
1207
+ "loss": 0.4381,
1208
+ "mean_token_accuracy": 0.8668898791074753,
1209
+ "step": 750
1210
+ },
1211
+ {
1212
+ "epoch": 1.4694424153883614,
1213
+ "grad_norm": 0.3454615445905308,
1214
+ "learning_rate": 1.9804396464420798e-06,
1215
+ "loss": 0.4407,
1216
+ "mean_token_accuracy": 0.8662415385246277,
1217
+ "step": 755
1218
+ },
1219
+ {
1220
+ "epoch": 1.47918188458729,
1221
+ "grad_norm": 0.3815154168732918,
1222
+ "learning_rate": 1.9130576070328695e-06,
1223
+ "loss": 0.4502,
1224
+ "mean_token_accuracy": 0.863922019302845,
1225
+ "step": 760
1226
+ },
1227
+ {
1228
+ "epoch": 1.4889213537862187,
1229
+ "grad_norm": 0.35782414062256107,
1230
+ "learning_rate": 1.8465696023643115e-06,
1231
+ "loss": 0.4484,
1232
+ "mean_token_accuracy": 0.8640252217650414,
1233
+ "step": 765
1234
+ },
1235
+ {
1236
+ "epoch": 1.4986608229851472,
1237
+ "grad_norm": 0.34966179386580715,
1238
+ "learning_rate": 1.7809948885735295e-06,
1239
+ "loss": 0.4476,
1240
+ "mean_token_accuracy": 0.8640663206577301,
1241
+ "step": 770
1242
+ },
1243
+ {
1244
+ "epoch": 1.508400292184076,
1245
+ "grad_norm": 0.3283826370466805,
1246
+ "learning_rate": 1.7163524572919748e-06,
1247
+ "loss": 0.4535,
1248
+ "mean_token_accuracy": 0.8629004299640656,
1249
+ "step": 775
1250
+ },
1251
+ {
1252
+ "epoch": 1.5181397613830048,
1253
+ "grad_norm": 0.3309412101835561,
1254
+ "learning_rate": 1.6526610301451028e-06,
1255
+ "loss": 0.4374,
1256
+ "mean_token_accuracy": 0.8666589662432671,
1257
+ "step": 780
1258
+ },
1259
+ {
1260
+ "epoch": 1.5278792305819333,
1261
+ "grad_norm": 0.35820157894670385,
1262
+ "learning_rate": 1.5899390533302538e-06,
1263
+ "loss": 0.4387,
1264
+ "mean_token_accuracy": 0.8667290091514588,
1265
+ "step": 785
1266
+ },
1267
+ {
1268
+ "epoch": 1.5376186997808619,
1269
+ "grad_norm": 0.34472450626785944,
1270
+ "learning_rate": 1.5282046922742876e-06,
1271
+ "loss": 0.4502,
1272
+ "mean_token_accuracy": 0.8635128363966942,
1273
+ "step": 790
1274
+ },
1275
+ {
1276
+ "epoch": 1.5473581689797906,
1277
+ "grad_norm": 0.31690650065908227,
1278
+ "learning_rate": 1.4674758263725614e-06,
1279
+ "loss": 0.4461,
1280
+ "mean_token_accuracy": 0.8644041374325753,
1281
+ "step": 795
1282
+ },
1283
+ {
1284
+ "epoch": 1.5570976381787194,
1285
+ "grad_norm": 0.3261992242400567,
1286
+ "learning_rate": 1.4077700438107183e-06,
1287
+ "loss": 0.445,
1288
+ "mean_token_accuracy": 0.8651425749063492,
1289
+ "step": 800
1290
+ },
1291
+ {
1292
+ "epoch": 1.566837107377648,
1293
+ "grad_norm": 0.37132210271715693,
1294
+ "learning_rate": 1.3491046364708294e-06,
1295
+ "loss": 0.445,
1296
+ "mean_token_accuracy": 0.8648254871368408,
1297
+ "step": 805
1298
+ },
1299
+ {
1300
+ "epoch": 1.5765765765765765,
1301
+ "grad_norm": 0.34739884194127096,
1302
+ "learning_rate": 1.2914965949233572e-06,
1303
+ "loss": 0.4474,
1304
+ "mean_token_accuracy": 0.8643729150295257,
1305
+ "step": 810
1306
+ },
1307
+ {
1308
+ "epoch": 1.5863160457755052,
1309
+ "grad_norm": 0.3092847696386134,
1310
+ "learning_rate": 1.2349626035063705e-06,
1311
+ "loss": 0.4389,
1312
+ "mean_token_accuracy": 0.8666358023881913,
1313
+ "step": 815
1314
+ },
1315
+ {
1316
+ "epoch": 1.596055514974434,
1317
+ "grad_norm": 0.33368514787450476,
1318
+ "learning_rate": 1.1795190354934587e-06,
1319
+ "loss": 0.4635,
1320
+ "mean_token_accuracy": 0.8606240957975387,
1321
+ "step": 820
1322
+ },
1323
+ {
1324
+ "epoch": 1.6057949841733625,
1325
+ "grad_norm": 0.3355701607651583,
1326
+ "learning_rate": 1.1251819483517334e-06,
1327
+ "loss": 0.4469,
1328
+ "mean_token_accuracy": 0.8647612199187279,
1329
+ "step": 825
1330
+ },
1331
+ {
1332
+ "epoch": 1.615534453372291,
1333
+ "grad_norm": 0.3215469441235877,
1334
+ "learning_rate": 1.0719670790912928e-06,
1335
+ "loss": 0.4479,
1336
+ "mean_token_accuracy": 0.8641064539551735,
1337
+ "step": 830
1338
+ },
1339
+ {
1340
+ "epoch": 1.6252739225712198,
1341
+ "grad_norm": 0.3331920310980409,
1342
+ "learning_rate": 1.019889839707498e-06,
1343
+ "loss": 0.447,
1344
+ "mean_token_accuracy": 0.8645351231098175,
1345
+ "step": 835
1346
+ },
1347
+ {
1348
+ "epoch": 1.6350133917701486,
1349
+ "grad_norm": 0.3311460501397384,
1350
+ "learning_rate": 9.689653127173743e-07,
1351
+ "loss": 0.4548,
1352
+ "mean_token_accuracy": 0.8624306350946427,
1353
+ "step": 840
1354
+ },
1355
+ {
1356
+ "epoch": 1.6447528609690771,
1357
+ "grad_norm": 0.3387701231769685,
1358
+ "learning_rate": 9.192082467914465e-07,
1359
+ "loss": 0.444,
1360
+ "mean_token_accuracy": 0.8649628892540931,
1361
+ "step": 845
1362
+ },
1363
+ {
1364
+ "epoch": 1.654492330168006,
1365
+ "grad_norm": 0.33969198086970775,
1366
+ "learning_rate": 8.706330524822548e-07,
1367
+ "loss": 0.4413,
1368
+ "mean_token_accuracy": 0.865912164747715,
1369
+ "step": 850
1370
+ },
1371
+ {
1372
+ "epoch": 1.6642317993669344,
1373
+ "grad_norm": 0.32626435288941413,
1374
+ "learning_rate": 8.232537980507848e-07,
1375
+ "loss": 0.4454,
1376
+ "mean_token_accuracy": 0.8650734156370163,
1377
+ "step": 855
1378
+ },
1379
+ {
1380
+ "epoch": 1.6739712685658632,
1381
+ "grad_norm": 0.32534723430048623,
1382
+ "learning_rate": 7.770842053920585e-07,
1383
+ "loss": 0.4424,
1384
+ "mean_token_accuracy": 0.8653772249817848,
1385
+ "step": 860
1386
+ },
1387
+ {
1388
+ "epoch": 1.683710737764792,
1389
+ "grad_norm": 0.33973250353406415,
1390
+ "learning_rate": 7.321376460610136e-07,
1391
+ "loss": 0.4398,
1392
+ "mean_token_accuracy": 0.865781269967556,
1393
+ "step": 865
1394
+ },
1395
+ {
1396
+ "epoch": 1.6934502069637205,
1397
+ "grad_norm": 0.31477206750092435,
1398
+ "learning_rate": 6.884271373998608e-07,
1399
+ "loss": 0.4402,
1400
+ "mean_token_accuracy": 0.8660067468881607,
1401
+ "step": 870
1402
+ },
1403
+ {
1404
+ "epoch": 1.703189676162649,
1405
+ "grad_norm": 0.3537664329765484,
1406
+ "learning_rate": 6.459653387680248e-07,
1407
+ "loss": 0.4426,
1408
+ "mean_token_accuracy": 0.8656087100505829,
1409
+ "step": 875
1410
+ },
1411
+ {
1412
+ "epoch": 1.7129291453615778,
1413
+ "grad_norm": 0.32858889216418335,
1414
+ "learning_rate": 6.047645478757635e-07,
1415
+ "loss": 0.4367,
1416
+ "mean_token_accuracy": 0.8670719146728516,
1417
+ "step": 880
1418
+ },
1419
+ {
1420
+ "epoch": 1.7226686145605066,
1421
+ "grad_norm": 0.32864732265316704,
1422
+ "learning_rate": 5.648366972225222e-07,
1423
+ "loss": 0.4527,
1424
+ "mean_token_accuracy": 0.8630405649542808,
1425
+ "step": 885
1426
+ },
1427
+ {
1428
+ "epoch": 1.7324080837594351,
1429
+ "grad_norm": 0.3173505795175411,
1430
+ "learning_rate": 5.261933506410722e-07,
1431
+ "loss": 0.4401,
1432
+ "mean_token_accuracy": 0.8665132194757461,
1433
+ "step": 890
1434
+ },
1435
+ {
1436
+ "epoch": 1.7421475529583637,
1437
+ "grad_norm": 0.36040952773775503,
1438
+ "learning_rate": 4.888456999484098e-07,
1439
+ "loss": 0.4465,
1440
+ "mean_token_accuracy": 0.8646954327821732,
1441
+ "step": 895
1442
+ },
1443
+ {
1444
+ "epoch": 1.7518870221572924,
1445
+ "grad_norm": 0.31604383070212816,
1446
+ "learning_rate": 4.528045617044019e-07,
1447
+ "loss": 0.443,
1448
+ "mean_token_accuracy": 0.8652036920189857,
1449
+ "step": 900
1450
+ },
1451
+ {
1452
+ "epoch": 1.7616264913562212,
1453
+ "grad_norm": 0.31389323834734445,
1454
+ "learning_rate": 4.180803740791156e-07,
1455
+ "loss": 0.4426,
1456
+ "mean_token_accuracy": 0.8656812936067582,
1457
+ "step": 905
1458
+ },
1459
+ {
1460
+ "epoch": 1.7713659605551497,
1461
+ "grad_norm": 0.3097110381509705,
1462
+ "learning_rate": 3.846831938297324e-07,
1463
+ "loss": 0.4468,
1464
+ "mean_token_accuracy": 0.8643257409334183,
1465
+ "step": 910
1466
+ },
1467
+ {
1468
+ "epoch": 1.7811054297540783,
1469
+ "grad_norm": 0.31179540512244613,
1470
+ "learning_rate": 3.5262269338792623e-07,
1471
+ "loss": 0.4447,
1472
+ "mean_token_accuracy": 0.8651932507753373,
1473
+ "step": 915
1474
+ },
1475
+ {
1476
+ "epoch": 1.790844898953007,
1477
+ "grad_norm": 0.33502042463366694,
1478
+ "learning_rate": 3.219081580585548e-07,
1479
+ "loss": 0.4508,
1480
+ "mean_token_accuracy": 0.863429008424282,
1481
+ "step": 920
1482
+ },
1483
+ {
1484
+ "epoch": 1.8005843681519358,
1485
+ "grad_norm": 0.3164739894198197,
1486
+ "learning_rate": 2.9254848333046817e-07,
1487
+ "loss": 0.4528,
1488
+ "mean_token_accuracy": 0.8630515649914742,
1489
+ "step": 925
1490
+ },
1491
+ {
1492
+ "epoch": 1.8103238373508643,
1493
+ "grad_norm": 0.3030075688047627,
1494
+ "learning_rate": 2.645521723002037e-07,
1495
+ "loss": 0.4507,
1496
+ "mean_token_accuracy": 0.8635053560137749,
1497
+ "step": 930
1498
+ },
1499
+ {
1500
+ "epoch": 1.8200633065497929,
1501
+ "grad_norm": 0.3371646202292577,
1502
+ "learning_rate": 2.3792733320934348e-07,
1503
+ "loss": 0.4441,
1504
+ "mean_token_accuracy": 0.8654543533921242,
1505
+ "step": 935
1506
+ },
1507
+ {
1508
+ "epoch": 1.8298027757487216,
1509
+ "grad_norm": 0.3074575906934633,
1510
+ "learning_rate": 2.12681677096217e-07,
1511
+ "loss": 0.4351,
1512
+ "mean_token_accuracy": 0.8675669968128205,
1513
+ "step": 940
1514
+ },
1515
+ {
1516
+ "epoch": 1.8395422449476504,
1517
+ "grad_norm": 0.30631898093344184,
1518
+ "learning_rate": 1.888225155626433e-07,
1519
+ "loss": 0.444,
1520
+ "mean_token_accuracy": 0.8653656959533691,
1521
+ "step": 945
1522
+ },
1523
+ {
1524
+ "epoch": 1.8492817141465792,
1525
+ "grad_norm": 0.2958594503607962,
1526
+ "learning_rate": 1.6635675865635859e-07,
1527
+ "loss": 0.4568,
1528
+ "mean_token_accuracy": 0.8619549512863159,
1529
+ "step": 950
1530
+ },
1531
+ {
1532
+ "epoch": 1.8590211833455077,
1533
+ "grad_norm": 0.31258753396965294,
1534
+ "learning_rate": 1.4529091286973994e-07,
1535
+ "loss": 0.444,
1536
+ "mean_token_accuracy": 0.8656645834445953,
1537
+ "step": 955
1538
+ },
1539
+ {
1540
+ "epoch": 1.8687606525444362,
1541
+ "grad_norm": 0.30999125345039114,
1542
+ "learning_rate": 1.2563107925540774e-07,
1543
+ "loss": 0.4444,
1544
+ "mean_token_accuracy": 0.8654705569148063,
1545
+ "step": 960
1546
+ },
1547
+ {
1548
+ "epoch": 1.878500121743365,
1549
+ "grad_norm": 0.3160340098992061,
1550
+ "learning_rate": 1.0738295165924783e-07,
1551
+ "loss": 0.4459,
1552
+ "mean_token_accuracy": 0.8648449763655662,
1553
+ "step": 965
1554
+ },
1555
+ {
1556
+ "epoch": 1.8882395909422938,
1557
+ "grad_norm": 0.3187758679651324,
1558
+ "learning_rate": 9.055181507137245e-08,
1559
+ "loss": 0.444,
1560
+ "mean_token_accuracy": 0.8654528453946113,
1561
+ "step": 970
1562
+ },
1563
+ {
1564
+ "epoch": 1.8979790601412223,
1565
+ "grad_norm": 0.33617665839091787,
1566
+ "learning_rate": 7.514254409549005e-08,
1567
+ "loss": 0.4481,
1568
+ "mean_token_accuracy": 0.8641701668500901,
1569
+ "step": 975
1570
+ },
1571
+ {
1572
+ "epoch": 1.9077185293401508,
1573
+ "grad_norm": 0.31232672260705024,
1574
+ "learning_rate": 6.115960153712963e-08,
1575
+ "loss": 0.4414,
1576
+ "mean_token_accuracy": 0.8660656422376632,
1577
+ "step": 980
1578
+ },
1579
+ {
1580
+ "epoch": 1.9174579985390796,
1581
+ "grad_norm": 0.303119007372895,
1582
+ "learning_rate": 4.860703711113246e-08,
1583
+ "loss": 0.4362,
1584
+ "mean_token_accuracy": 0.8676238685846329,
1585
+ "step": 985
1586
+ },
1587
+ {
1588
+ "epoch": 1.9271974677380084,
1589
+ "grad_norm": 0.30341116951092534,
1590
+ "learning_rate": 3.748848626878132e-08,
1591
+ "loss": 0.4508,
1592
+ "mean_token_accuracy": 0.8636912703514099,
1593
+ "step": 990
1594
+ },
1595
+ {
1596
+ "epoch": 1.936936936936937,
1597
+ "grad_norm": 0.31515978754450974,
1598
+ "learning_rate": 2.7807169144906108e-08,
1599
+ "loss": 0.4528,
1600
+ "mean_token_accuracy": 0.8630853027105332,
1601
+ "step": 995
1602
+ },
1603
+ {
1604
+ "epoch": 1.9466764061358655,
1605
+ "grad_norm": 0.30748596132213357,
1606
+ "learning_rate": 1.9565889625275945e-08,
1607
+ "loss": 0.4558,
1608
+ "mean_token_accuracy": 0.8628044292330742,
1609
+ "step": 1000
1610
+ },
1611
+ {
1612
+ "epoch": 1.9564158753347942,
1613
+ "grad_norm": 0.30248992557073057,
1614
+ "learning_rate": 1.2767034534540978e-08,
1615
+ "loss": 0.4467,
1616
+ "mean_token_accuracy": 0.86439578384161,
1617
+ "step": 1005
1618
+ },
1619
+ {
1620
+ "epoch": 1.966155344533723,
1621
+ "grad_norm": 0.3256862744869616,
1622
+ "learning_rate": 7.412572944965335e-09,
1623
+ "loss": 0.4548,
1624
+ "mean_token_accuracy": 0.8625592529773712,
1625
+ "step": 1010
1626
+ },
1627
+ {
1628
+ "epoch": 1.9758948137326515,
1629
+ "grad_norm": 0.3147340514506444,
1630
+ "learning_rate": 3.5040556061483043e-09,
1631
+ "loss": 0.4422,
1632
+ "mean_token_accuracy": 0.8657747611403466,
1633
+ "step": 1015
1634
+ },
1635
+ {
1636
+ "epoch": 1.98563428293158,
1637
+ "grad_norm": 0.32764257560082577,
1638
+ "learning_rate": 1.0426144958985974e-09,
1639
+ "loss": 0.4419,
1640
+ "mean_token_accuracy": 0.8660721600055694,
1641
+ "step": 1020
1642
+ },
1643
+ {
1644
+ "epoch": 1.9953737521305088,
1645
+ "grad_norm": 0.3098108021812636,
1646
+ "learning_rate": 2.8962492393258546e-11,
1647
+ "loss": 0.4452,
1648
+ "mean_token_accuracy": 0.8650736406445503,
1649
+ "step": 1025
1650
+ },
1651
+ {
1652
+ "epoch": 1.9973216459702945,
1653
+ "mean_token_accuracy": 0.8681519404053688,
1654
+ "step": 1026,
1655
+ "total_flos": 1074983740637184.0,
1656
+ "train_loss": 0.49954891321022377,
1657
+ "train_runtime": 168401.8035,
1658
+ "train_samples_per_second": 0.78,
1659
+ "train_steps_per_second": 0.006
1660
+ }
1661
+ ],
1662
+ "logging_steps": 5,
1663
+ "max_steps": 1026,
1664
+ "num_input_tokens_seen": 0,
1665
+ "num_train_epochs": 2,
1666
+ "save_steps": 50,
1667
+ "stateful_callbacks": {
1668
+ "TrainerControl": {
1669
+ "args": {
1670
+ "should_epoch_stop": false,
1671
+ "should_evaluate": false,
1672
+ "should_log": false,
1673
+ "should_save": true,
1674
+ "should_training_stop": true
1675
+ },
1676
+ "attributes": {}
1677
+ }
1678
+ },
1679
+ "total_flos": 1074983740637184.0,
1680
+ "train_batch_size": 2,
1681
+ "trial_name": null,
1682
+ "trial_params": null
1683
+ }