Model save
Browse files- README.md +2 -4
- all_results.json +2 -7
- train_results.json +2 -2
- trainer_state.json +565 -565
README.md
CHANGED
|
@@ -1,11 +1,9 @@
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen2.5-3B
|
| 3 |
-
datasets: ebony59/MATH-lighteval-gen-correct
|
| 4 |
library_name: transformers
|
| 5 |
model_name: Qwen2.5-3B-MATH-lighteval-gen-SFT-15epoch
|
| 6 |
tags:
|
| 7 |
- generated_from_trainer
|
| 8 |
-
- open-r1
|
| 9 |
- trl
|
| 10 |
- sft
|
| 11 |
licence: license
|
|
@@ -13,7 +11,7 @@ licence: license
|
|
| 13 |
|
| 14 |
# Model Card for Qwen2.5-3B-MATH-lighteval-gen-SFT-15epoch
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [Qwen/Qwen2.5-3B](https://huggingface.co/Qwen/Qwen2.5-3B)
|
| 17 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 18 |
|
| 19 |
## Quick start
|
|
@@ -29,7 +27,7 @@ print(output["generated_text"])
|
|
| 29 |
|
| 30 |
## Training procedure
|
| 31 |
|
| 32 |
-
[<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/pl03818948-ludwig-maximilianuniversity-of-munich/qwen-math-sft/runs/
|
| 33 |
|
| 34 |
|
| 35 |
This model was trained with SFT.
|
|
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen2.5-3B
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
model_name: Qwen2.5-3B-MATH-lighteval-gen-SFT-15epoch
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
|
|
| 7 |
- trl
|
| 8 |
- sft
|
| 9 |
licence: license
|
|
|
|
| 11 |
|
| 12 |
# Model Card for Qwen2.5-3B-MATH-lighteval-gen-SFT-15epoch
|
| 13 |
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-3B](https://huggingface.co/Qwen/Qwen2.5-3B).
|
| 15 |
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
|
| 17 |
## Quick start
|
|
|
|
| 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/pl03818948-ludwig-maximilianuniversity-of-munich/qwen-math-sft/runs/uujrvm5v)
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with SFT.
|
all_results.json
CHANGED
|
@@ -1,12 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"eval_loss": 0.46884241700172424,
|
| 3 |
-
"eval_runtime": 69.6122,
|
| 4 |
-
"eval_samples": 3537,
|
| 5 |
-
"eval_samples_per_second": 12.282,
|
| 6 |
-
"eval_steps_per_second": 1.537,
|
| 7 |
"total_flos": 24409842647040.0,
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
"train_samples": 6726,
|
| 11 |
"train_samples_per_second": 2.194,
|
| 12 |
"train_steps_per_second": 0.067
|
|
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"total_flos": 24409842647040.0,
|
| 3 |
+
"train_loss": 0.12274208276295194,
|
| 4 |
+
"train_runtime": 3782.9235,
|
| 5 |
"train_samples": 6726,
|
| 6 |
"train_samples_per_second": 2.194,
|
| 7 |
"train_steps_per_second": 0.067
|
train_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"total_flos": 24409842647040.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 6726,
|
| 6 |
"train_samples_per_second": 2.194,
|
| 7 |
"train_steps_per_second": 0.067
|
|
|
|
| 1 |
{
|
| 2 |
"total_flos": 24409842647040.0,
|
| 3 |
+
"train_loss": 0.12274208276295194,
|
| 4 |
+
"train_runtime": 3782.9235,
|
| 5 |
"train_samples": 6726,
|
| 6 |
"train_samples_per_second": 2.194,
|
| 7 |
"train_steps_per_second": 0.067
|
trainer_state.json
CHANGED
|
@@ -11,8 +11,8 @@
|
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 0.03855421686746988,
|
| 14 |
-
"grad_norm": 2.
|
| 15 |
-
"learning_rate":
|
| 16 |
"loss": 0.2354,
|
| 17 |
"mean_token_accuracy": 0.930065356194973,
|
| 18 |
"num_tokens": 131072.0,
|
|
@@ -20,1266 +20,1266 @@
|
|
| 20 |
},
|
| 21 |
{
|
| 22 |
"epoch": 0.07710843373493977,
|
| 23 |
-
"grad_norm": 2.
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss": 0.
|
| 26 |
-
"mean_token_accuracy": 0.
|
| 27 |
"num_tokens": 262144.0,
|
| 28 |
"step": 4
|
| 29 |
},
|
| 30 |
{
|
| 31 |
"epoch": 0.11566265060240964,
|
| 32 |
-
"grad_norm": 1.
|
| 33 |
-
"learning_rate":
|
| 34 |
-
"loss": 0.
|
| 35 |
-
"mean_token_accuracy": 0.
|
| 36 |
"num_tokens": 393216.0,
|
| 37 |
"step": 6
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.15421686746987953,
|
| 41 |
-
"grad_norm":
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 0.
|
| 44 |
-
"mean_token_accuracy": 0.
|
| 45 |
"num_tokens": 524288.0,
|
| 46 |
"step": 8
|
| 47 |
},
|
| 48 |
{
|
| 49 |
"epoch": 0.1927710843373494,
|
| 50 |
-
"grad_norm":
|
| 51 |
-
"learning_rate":
|
| 52 |
-
"loss": 0.
|
| 53 |
-
"mean_token_accuracy": 0.
|
| 54 |
"num_tokens": 654484.0,
|
| 55 |
"step": 10
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"epoch": 0.23132530120481928,
|
| 59 |
-
"grad_norm":
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss": 0.
|
| 62 |
-
"mean_token_accuracy": 0.
|
| 63 |
"num_tokens": 785556.0,
|
| 64 |
"step": 12
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"epoch": 0.26987951807228916,
|
| 68 |
-
"grad_norm": 0.
|
| 69 |
-
"learning_rate":
|
| 70 |
-
"loss": 0.
|
| 71 |
-
"mean_token_accuracy": 0.
|
| 72 |
"num_tokens": 915519.0,
|
| 73 |
"step": 14
|
| 74 |
},
|
| 75 |
{
|
| 76 |
"epoch": 0.30843373493975906,
|
| 77 |
-
"grad_norm": 0.
|
| 78 |
-
"learning_rate":
|
| 79 |
-
"loss": 0.
|
| 80 |
-
"mean_token_accuracy": 0.
|
| 81 |
"num_tokens": 1046591.0,
|
| 82 |
"step": 16
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"epoch": 0.3469879518072289,
|
| 86 |
-
"grad_norm": 0.
|
| 87 |
-
"learning_rate":
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"mean_token_accuracy": 0.
|
| 90 |
"num_tokens": 1177663.0,
|
| 91 |
"step": 18
|
| 92 |
},
|
| 93 |
{
|
| 94 |
"epoch": 0.3855421686746988,
|
| 95 |
-
"grad_norm": 0.
|
| 96 |
-
"learning_rate":
|
| 97 |
-
"loss": 0.
|
| 98 |
-
"mean_token_accuracy": 0.
|
| 99 |
"num_tokens": 1308735.0,
|
| 100 |
"step": 20
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.3855421686746988,
|
| 104 |
-
"eval_loss": 0.
|
| 105 |
-
"eval_mean_token_accuracy": 0.
|
| 106 |
"eval_num_tokens": 1308735.0,
|
| 107 |
-
"eval_runtime": 70.
|
| 108 |
-
"eval_samples_per_second": 12.
|
| 109 |
-
"eval_steps_per_second": 1.
|
| 110 |
"step": 20
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.42409638554216866,
|
| 114 |
-
"grad_norm": 0.
|
| 115 |
-
"learning_rate":
|
| 116 |
-
"loss": 0.
|
| 117 |
-
"mean_token_accuracy": 0.
|
| 118 |
"num_tokens": 1439807.0,
|
| 119 |
"step": 22
|
| 120 |
},
|
| 121 |
{
|
| 122 |
"epoch": 0.46265060240963857,
|
| 123 |
-
"grad_norm": 0.
|
| 124 |
-
"learning_rate":
|
| 125 |
-
"loss": 0.
|
| 126 |
-
"mean_token_accuracy": 0.
|
| 127 |
"num_tokens": 1570062.0,
|
| 128 |
"step": 24
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.5012048192771085,
|
| 132 |
-
"grad_norm": 0.
|
| 133 |
-
"learning_rate": 9.
|
| 134 |
-
"loss": 0.
|
| 135 |
-
"mean_token_accuracy": 0.
|
| 136 |
"num_tokens": 1701134.0,
|
| 137 |
"step": 26
|
| 138 |
},
|
| 139 |
{
|
| 140 |
"epoch": 0.5397590361445783,
|
| 141 |
-
"grad_norm": 0.
|
| 142 |
-
"learning_rate": 9.
|
| 143 |
-
"loss": 0.
|
| 144 |
-
"mean_token_accuracy": 0.
|
| 145 |
"num_tokens": 1832133.0,
|
| 146 |
"step": 28
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.5783132530120482,
|
| 150 |
-
"grad_norm": 0.
|
| 151 |
-
"learning_rate": 9.
|
| 152 |
-
"loss": 0.
|
| 153 |
-
"mean_token_accuracy": 0.
|
| 154 |
"num_tokens": 1963205.0,
|
| 155 |
"step": 30
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 0.6168674698795181,
|
| 159 |
-
"grad_norm": 0.
|
| 160 |
-
"learning_rate": 9.
|
| 161 |
-
"loss": 0.
|
| 162 |
-
"mean_token_accuracy": 0.
|
| 163 |
"num_tokens": 2094277.0,
|
| 164 |
"step": 32
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.655421686746988,
|
| 168 |
-
"grad_norm": 0.
|
| 169 |
-
"learning_rate": 9.
|
| 170 |
-
"loss": 0.
|
| 171 |
-
"mean_token_accuracy": 0.
|
| 172 |
"num_tokens": 2225349.0,
|
| 173 |
"step": 34
|
| 174 |
},
|
| 175 |
{
|
| 176 |
"epoch": 0.6939759036144578,
|
| 177 |
-
"grad_norm": 0.
|
| 178 |
-
"learning_rate": 9.
|
| 179 |
-
"loss": 0.
|
| 180 |
-
"mean_token_accuracy": 0.
|
| 181 |
"num_tokens": 2355263.0,
|
| 182 |
"step": 36
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.7325301204819277,
|
| 186 |
-
"grad_norm":
|
| 187 |
-
"learning_rate": 9.
|
| 188 |
-
"loss": 0.
|
| 189 |
-
"mean_token_accuracy": 0.
|
| 190 |
"num_tokens": 2486335.0,
|
| 191 |
"step": 38
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.7710843373493976,
|
| 195 |
-
"grad_norm": 0.
|
| 196 |
-
"learning_rate": 9.
|
| 197 |
-
"loss": 0.
|
| 198 |
-
"mean_token_accuracy": 0.
|
| 199 |
"num_tokens": 2617407.0,
|
| 200 |
"step": 40
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.7710843373493976,
|
| 204 |
-
"eval_loss": 0.
|
| 205 |
-
"eval_mean_token_accuracy": 0.
|
| 206 |
"eval_num_tokens": 2617407.0,
|
| 207 |
-
"eval_runtime": 69.
|
| 208 |
-
"eval_samples_per_second": 12.
|
| 209 |
"eval_steps_per_second": 1.536,
|
| 210 |
"step": 40
|
| 211 |
},
|
| 212 |
{
|
| 213 |
"epoch": 0.8096385542168675,
|
| 214 |
-
"grad_norm": 0.
|
| 215 |
-
"learning_rate": 9.
|
| 216 |
-
"loss": 0.
|
| 217 |
-
"mean_token_accuracy": 0.
|
| 218 |
"num_tokens": 2748479.0,
|
| 219 |
"step": 42
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.8481927710843373,
|
| 223 |
-
"grad_norm": 0.
|
| 224 |
-
"learning_rate": 9.
|
| 225 |
-
"loss": 0.
|
| 226 |
-
"mean_token_accuracy": 0.
|
| 227 |
"num_tokens": 2879551.0,
|
| 228 |
"step": 44
|
| 229 |
},
|
| 230 |
{
|
| 231 |
"epoch": 0.8867469879518072,
|
| 232 |
-
"grad_norm": 0.
|
| 233 |
-
"learning_rate": 9.
|
| 234 |
-
"loss": 0.
|
| 235 |
-
"mean_token_accuracy": 0.
|
| 236 |
"num_tokens": 3010185.0,
|
| 237 |
"step": 46
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.9253012048192771,
|
| 241 |
-
"grad_norm": 0.
|
| 242 |
-
"learning_rate": 9.
|
| 243 |
-
"loss": 0.
|
| 244 |
-
"mean_token_accuracy": 0.
|
| 245 |
"num_tokens": 3141257.0,
|
| 246 |
"step": 48
|
| 247 |
},
|
| 248 |
{
|
| 249 |
"epoch": 0.963855421686747,
|
| 250 |
-
"grad_norm": 0.
|
| 251 |
-
"learning_rate": 9.
|
| 252 |
-
"loss": 0.
|
| 253 |
-
"mean_token_accuracy": 0.
|
| 254 |
"num_tokens": 3271788.0,
|
| 255 |
"step": 50
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 1.0192771084337349,
|
| 259 |
-
"grad_norm": 0.
|
| 260 |
-
"learning_rate": 9.
|
| 261 |
-
"loss": 0.
|
| 262 |
-
"mean_token_accuracy": 0.
|
| 263 |
"num_tokens": 3435628.0,
|
| 264 |
"step": 52
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 1.0578313253012048,
|
| 268 |
-
"grad_norm": 0.
|
| 269 |
-
"learning_rate": 9.
|
| 270 |
-
"loss": 0.
|
| 271 |
-
"mean_token_accuracy": 0.
|
| 272 |
"num_tokens": 3566700.0,
|
| 273 |
"step": 54
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 1.0963855421686748,
|
| 277 |
-
"grad_norm": 0.
|
| 278 |
-
"learning_rate": 9.
|
| 279 |
-
"loss": 0.
|
| 280 |
-
"mean_token_accuracy": 0.
|
| 281 |
"num_tokens": 3697772.0,
|
| 282 |
"step": 56
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 1.1349397590361445,
|
| 286 |
-
"grad_norm": 0.
|
| 287 |
-
"learning_rate": 9.
|
| 288 |
-
"loss": 0.
|
| 289 |
-
"mean_token_accuracy": 0.
|
| 290 |
"num_tokens": 3828844.0,
|
| 291 |
"step": 58
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 1.1734939759036145,
|
| 295 |
-
"grad_norm": 0.
|
| 296 |
-
"learning_rate": 9.
|
| 297 |
-
"loss": 0.
|
| 298 |
-
"mean_token_accuracy": 0.
|
| 299 |
"num_tokens": 3959916.0,
|
| 300 |
"step": 60
|
| 301 |
},
|
| 302 |
{
|
| 303 |
"epoch": 1.1734939759036145,
|
| 304 |
-
"eval_loss": 0.
|
| 305 |
-
"eval_mean_token_accuracy": 0.
|
| 306 |
"eval_num_tokens": 3959916.0,
|
| 307 |
-
"eval_runtime": 69.
|
| 308 |
-
"eval_samples_per_second": 12.
|
| 309 |
"eval_steps_per_second": 1.532,
|
| 310 |
"step": 60
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 1.2120481927710842,
|
| 314 |
-
"grad_norm": 0.
|
| 315 |
-
"learning_rate": 9.
|
| 316 |
-
"loss": 0.
|
| 317 |
-
"mean_token_accuracy": 0.
|
| 318 |
"num_tokens": 4089879.0,
|
| 319 |
"step": 62
|
| 320 |
},
|
| 321 |
{
|
| 322 |
"epoch": 1.2506024096385542,
|
| 323 |
-
"grad_norm": 0.
|
| 324 |
-
"learning_rate":
|
| 325 |
-
"loss": 0.
|
| 326 |
-
"mean_token_accuracy": 0.
|
| 327 |
"num_tokens": 4220951.0,
|
| 328 |
"step": 64
|
| 329 |
},
|
| 330 |
{
|
| 331 |
"epoch": 1.2891566265060241,
|
| 332 |
-
"grad_norm": 0.
|
| 333 |
-
"learning_rate":
|
| 334 |
-
"loss": 0.
|
| 335 |
-
"mean_token_accuracy": 0.
|
| 336 |
"num_tokens": 4352023.0,
|
| 337 |
"step": 66
|
| 338 |
},
|
| 339 |
{
|
| 340 |
"epoch": 1.3277108433734939,
|
| 341 |
-
"grad_norm": 0.
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss": 0.
|
| 344 |
-
"mean_token_accuracy": 0.
|
| 345 |
"num_tokens": 4483095.0,
|
| 346 |
"step": 68
|
| 347 |
},
|
| 348 |
{
|
| 349 |
"epoch": 1.3662650602409638,
|
| 350 |
-
"grad_norm": 0.
|
| 351 |
-
"learning_rate":
|
| 352 |
-
"loss": 0.
|
| 353 |
-
"mean_token_accuracy": 0.
|
| 354 |
"num_tokens": 4614167.0,
|
| 355 |
"step": 70
|
| 356 |
},
|
| 357 |
{
|
| 358 |
"epoch": 1.4048192771084338,
|
| 359 |
-
"grad_norm": 0.
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss": 0.
|
| 362 |
-
"mean_token_accuracy": 0.
|
| 363 |
"num_tokens": 4745239.0,
|
| 364 |
"step": 72
|
| 365 |
},
|
| 366 |
{
|
| 367 |
"epoch": 1.4433734939759035,
|
| 368 |
-
"grad_norm": 0.
|
| 369 |
-
"learning_rate": 8.
|
| 370 |
-
"loss": 0.
|
| 371 |
-
"mean_token_accuracy": 0.
|
| 372 |
"num_tokens": 4876311.0,
|
| 373 |
"step": 74
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 1.4819277108433735,
|
| 377 |
-
"grad_norm": 0.
|
| 378 |
-
"learning_rate": 8.
|
| 379 |
-
"loss": 0.
|
| 380 |
-
"mean_token_accuracy": 0.
|
| 381 |
"num_tokens": 5007383.0,
|
| 382 |
"step": 76
|
| 383 |
},
|
| 384 |
{
|
| 385 |
"epoch": 1.5204819277108435,
|
| 386 |
-
"grad_norm": 0.
|
| 387 |
-
"learning_rate": 8.
|
| 388 |
-
"loss": 0.
|
| 389 |
-
"mean_token_accuracy": 0.
|
| 390 |
"num_tokens": 5138455.0,
|
| 391 |
"step": 78
|
| 392 |
},
|
| 393 |
{
|
| 394 |
"epoch": 1.5590361445783132,
|
| 395 |
-
"grad_norm": 0.
|
| 396 |
-
"learning_rate": 8.
|
| 397 |
-
"loss": 0.
|
| 398 |
-
"mean_token_accuracy": 0.
|
| 399 |
"num_tokens": 5269527.0,
|
| 400 |
"step": 80
|
| 401 |
},
|
| 402 |
{
|
| 403 |
"epoch": 1.5590361445783132,
|
| 404 |
-
"eval_loss": 0.
|
| 405 |
-
"eval_mean_token_accuracy": 0.
|
| 406 |
"eval_num_tokens": 5269527.0,
|
| 407 |
-
"eval_runtime": 69.
|
| 408 |
-
"eval_samples_per_second": 12.
|
| 409 |
"eval_steps_per_second": 1.536,
|
| 410 |
"step": 80
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 1.5975903614457831,
|
| 414 |
-
"grad_norm": 0.
|
| 415 |
-
"learning_rate": 8.
|
| 416 |
-
"loss": 0.
|
| 417 |
-
"mean_token_accuracy": 0.
|
| 418 |
"num_tokens": 5400161.0,
|
| 419 |
"step": 82
|
| 420 |
},
|
| 421 |
{
|
| 422 |
"epoch": 1.636144578313253,
|
| 423 |
-
"grad_norm": 0.
|
| 424 |
-
"learning_rate": 8.
|
| 425 |
-
"loss": 0.
|
| 426 |
-
"mean_token_accuracy": 0.
|
| 427 |
"num_tokens": 5531233.0,
|
| 428 |
"step": 84
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 1.6746987951807228,
|
| 432 |
-
"grad_norm": 0.
|
| 433 |
-
"learning_rate":
|
| 434 |
-
"loss": 0.
|
| 435 |
-
"mean_token_accuracy": 0.
|
| 436 |
"num_tokens": 5662305.0,
|
| 437 |
"step": 86
|
| 438 |
},
|
| 439 |
{
|
| 440 |
"epoch": 1.7132530120481928,
|
| 441 |
-
"grad_norm": 0.
|
| 442 |
-
"learning_rate":
|
| 443 |
-
"loss": 0.
|
| 444 |
-
"mean_token_accuracy": 0.
|
| 445 |
"num_tokens": 5792219.0,
|
| 446 |
"step": 88
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 1.7518072289156628,
|
| 450 |
-
"grad_norm": 0.
|
| 451 |
-
"learning_rate":
|
| 452 |
-
"loss": 0.
|
| 453 |
-
"mean_token_accuracy": 0.
|
| 454 |
"num_tokens": 5923291.0,
|
| 455 |
"step": 90
|
| 456 |
},
|
| 457 |
{
|
| 458 |
"epoch": 1.7903614457831325,
|
| 459 |
-
"grad_norm": 0.
|
| 460 |
-
"learning_rate":
|
| 461 |
-
"loss": 0.
|
| 462 |
-
"mean_token_accuracy": 0.
|
| 463 |
"num_tokens": 6053822.0,
|
| 464 |
"step": 92
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 1.8289156626506025,
|
| 468 |
-
"grad_norm": 0.
|
| 469 |
-
"learning_rate":
|
| 470 |
-
"loss": 0.
|
| 471 |
-
"mean_token_accuracy": 0.
|
| 472 |
"num_tokens": 6184894.0,
|
| 473 |
"step": 94
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 1.8674698795180724,
|
| 477 |
-
"grad_norm": 0.
|
| 478 |
-
"learning_rate": 7.
|
| 479 |
-
"loss": 0.
|
| 480 |
-
"mean_token_accuracy": 0.
|
| 481 |
"num_tokens": 6315149.0,
|
| 482 |
"step": 96
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 1.9060240963855422,
|
| 486 |
-
"grad_norm": 0.
|
| 487 |
-
"learning_rate": 7.
|
| 488 |
-
"loss": 0.
|
| 489 |
-
"mean_token_accuracy": 0.
|
| 490 |
"num_tokens": 6445345.0,
|
| 491 |
"step": 98
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 1.944578313253012,
|
| 495 |
-
"grad_norm": 0.
|
| 496 |
-
"learning_rate": 7.
|
| 497 |
-
"loss": 0.
|
| 498 |
-
"mean_token_accuracy": 0.
|
| 499 |
"num_tokens": 6576344.0,
|
| 500 |
"step": 100
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 1.944578313253012,
|
| 504 |
-
"eval_loss": 0.
|
| 505 |
-
"eval_mean_token_accuracy": 0.
|
| 506 |
"eval_num_tokens": 6576344.0,
|
| 507 |
-
"eval_runtime": 69.
|
| 508 |
-
"eval_samples_per_second": 12.
|
| 509 |
-
"eval_steps_per_second": 1.
|
| 510 |
"step": 100
|
| 511 |
},
|
| 512 |
{
|
| 513 |
"epoch": 1.983132530120482,
|
| 514 |
-
"grad_norm": 0.
|
| 515 |
-
"learning_rate": 7.
|
| 516 |
-
"loss": 0.
|
| 517 |
-
"mean_token_accuracy": 0.
|
| 518 |
"num_tokens": 6707416.0,
|
| 519 |
"step": 102
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 2.0385542168674697,
|
| 523 |
-
"grad_norm": 0.
|
| 524 |
-
"learning_rate":
|
| 525 |
-
"loss": 0.
|
| 526 |
-
"mean_token_accuracy": 0.
|
| 527 |
"num_tokens": 6871256.0,
|
| 528 |
"step": 104
|
| 529 |
},
|
| 530 |
{
|
| 531 |
"epoch": 2.07710843373494,
|
| 532 |
-
"grad_norm": 0.
|
| 533 |
-
"learning_rate":
|
| 534 |
-
"loss": 0.
|
| 535 |
-
"mean_token_accuracy": 0.
|
| 536 |
"num_tokens": 7002255.0,
|
| 537 |
"step": 106
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 2.1156626506024097,
|
| 541 |
-
"grad_norm": 0.
|
| 542 |
-
"learning_rate":
|
| 543 |
-
"loss": 0.
|
| 544 |
-
"mean_token_accuracy": 0.
|
| 545 |
"num_tokens": 7133327.0,
|
| 546 |
"step": 108
|
| 547 |
},
|
| 548 |
{
|
| 549 |
"epoch": 2.1542168674698794,
|
| 550 |
-
"grad_norm": 0.
|
| 551 |
-
"learning_rate":
|
| 552 |
-
"loss": 0.
|
| 553 |
-
"mean_token_accuracy": 0.
|
| 554 |
"num_tokens": 7263523.0,
|
| 555 |
"step": 110
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 2.1927710843373496,
|
| 559 |
-
"grad_norm": 0.
|
| 560 |
-
"learning_rate": 6.
|
| 561 |
-
"loss": 0.
|
| 562 |
-
"mean_token_accuracy": 0.
|
| 563 |
"num_tokens": 7394595.0,
|
| 564 |
"step": 112
|
| 565 |
},
|
| 566 |
{
|
| 567 |
"epoch": 2.2313253012048193,
|
| 568 |
-
"grad_norm": 0.
|
| 569 |
-
"learning_rate": 6.
|
| 570 |
-
"loss": 0.
|
| 571 |
-
"mean_token_accuracy": 0.
|
| 572 |
"num_tokens": 7525667.0,
|
| 573 |
"step": 114
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 2.269879518072289,
|
| 577 |
-
"grad_norm": 0.
|
| 578 |
-
"learning_rate": 6.
|
| 579 |
-
"loss": 0.
|
| 580 |
-
"mean_token_accuracy": 0.
|
| 581 |
"num_tokens": 7656739.0,
|
| 582 |
"step": 116
|
| 583 |
},
|
| 584 |
{
|
| 585 |
"epoch": 2.3084337349397592,
|
| 586 |
-
"grad_norm": 0.
|
| 587 |
-
"learning_rate": 6.
|
| 588 |
-
"loss": 0.
|
| 589 |
-
"mean_token_accuracy": 0.
|
| 590 |
"num_tokens": 7787811.0,
|
| 591 |
"step": 118
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 2.346987951807229,
|
| 595 |
-
"grad_norm": 0.
|
| 596 |
-
"learning_rate":
|
| 597 |
-
"loss": 0.
|
| 598 |
-
"mean_token_accuracy": 0.
|
| 599 |
"num_tokens": 7916616.0,
|
| 600 |
"step": 120
|
| 601 |
},
|
| 602 |
{
|
| 603 |
"epoch": 2.346987951807229,
|
| 604 |
-
"eval_loss": 0.
|
| 605 |
-
"eval_mean_token_accuracy": 0.
|
| 606 |
"eval_num_tokens": 7916616.0,
|
| 607 |
-
"eval_runtime": 69.
|
| 608 |
-
"eval_samples_per_second": 12.
|
| 609 |
-
"eval_steps_per_second": 1.
|
| 610 |
"step": 120
|
| 611 |
},
|
| 612 |
{
|
| 613 |
"epoch": 2.3855421686746987,
|
| 614 |
-
"grad_norm": 0.
|
| 615 |
-
"learning_rate":
|
| 616 |
-
"loss": 0.
|
| 617 |
-
"mean_token_accuracy": 0.
|
| 618 |
"num_tokens": 8047688.0,
|
| 619 |
"step": 122
|
| 620 |
},
|
| 621 |
{
|
| 622 |
"epoch": 2.4240963855421684,
|
| 623 |
-
"grad_norm": 0.
|
| 624 |
-
"learning_rate":
|
| 625 |
-
"loss": 0.
|
| 626 |
-
"mean_token_accuracy": 0.
|
| 627 |
"num_tokens": 8178760.0,
|
| 628 |
"step": 124
|
| 629 |
},
|
| 630 |
{
|
| 631 |
"epoch": 2.4626506024096386,
|
| 632 |
-
"grad_norm": 0.
|
| 633 |
-
"learning_rate":
|
| 634 |
-
"loss": 0.
|
| 635 |
-
"mean_token_accuracy": 0.
|
| 636 |
"num_tokens": 8309832.0,
|
| 637 |
"step": 126
|
| 638 |
},
|
| 639 |
{
|
| 640 |
"epoch": 2.5012048192771084,
|
| 641 |
-
"grad_norm": 0.
|
| 642 |
-
"learning_rate": 5.
|
| 643 |
-
"loss": 0.
|
| 644 |
-
"mean_token_accuracy": 0.
|
| 645 |
"num_tokens": 8440904.0,
|
| 646 |
"step": 128
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 2.539759036144578,
|
| 650 |
-
"grad_norm": 0.
|
| 651 |
-
"learning_rate": 5.
|
| 652 |
-
"loss": 0.
|
| 653 |
-
"mean_token_accuracy": 0.
|
| 654 |
"num_tokens": 8571976.0,
|
| 655 |
"step": 130
|
| 656 |
},
|
| 657 |
{
|
| 658 |
"epoch": 2.5783132530120483,
|
| 659 |
-
"grad_norm": 0.
|
| 660 |
-
"learning_rate": 5.
|
| 661 |
-
"loss": 0.
|
| 662 |
-
"mean_token_accuracy": 0.
|
| 663 |
"num_tokens": 8703048.0,
|
| 664 |
"step": 132
|
| 665 |
},
|
| 666 |
{
|
| 667 |
"epoch": 2.616867469879518,
|
| 668 |
-
"grad_norm": 0.
|
| 669 |
-
"learning_rate": 5.
|
| 670 |
-
"loss": 0.
|
| 671 |
-
"mean_token_accuracy": 0.
|
| 672 |
"num_tokens": 8834120.0,
|
| 673 |
"step": 134
|
| 674 |
},
|
| 675 |
{
|
| 676 |
"epoch": 2.6554216867469878,
|
| 677 |
-
"grad_norm": 0.
|
| 678 |
-
"learning_rate":
|
| 679 |
-
"loss": 0.
|
| 680 |
-
"mean_token_accuracy": 0.
|
| 681 |
"num_tokens": 8964375.0,
|
| 682 |
"step": 136
|
| 683 |
},
|
| 684 |
{
|
| 685 |
"epoch": 2.693975903614458,
|
| 686 |
-
"grad_norm": 0.
|
| 687 |
-
"learning_rate":
|
| 688 |
-
"loss": 0.
|
| 689 |
-
"mean_token_accuracy": 0.
|
| 690 |
"num_tokens": 9094906.0,
|
| 691 |
"step": 138
|
| 692 |
},
|
| 693 |
{
|
| 694 |
"epoch": 2.7325301204819277,
|
| 695 |
-
"grad_norm": 0.
|
| 696 |
-
"learning_rate":
|
| 697 |
-
"loss": 0.
|
| 698 |
-
"mean_token_accuracy": 0.
|
| 699 |
"num_tokens": 9225978.0,
|
| 700 |
"step": 140
|
| 701 |
},
|
| 702 |
{
|
| 703 |
"epoch": 2.7325301204819277,
|
| 704 |
-
"eval_loss": 0.
|
| 705 |
-
"eval_mean_token_accuracy": 0.
|
| 706 |
"eval_num_tokens": 9225978.0,
|
| 707 |
-
"eval_runtime": 69.
|
| 708 |
-
"eval_samples_per_second": 12.
|
| 709 |
-
"eval_steps_per_second": 1.
|
| 710 |
"step": 140
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 2.7710843373493974,
|
| 714 |
-
"grad_norm": 0.
|
| 715 |
-
"learning_rate": 4.
|
| 716 |
-
"loss": 0.
|
| 717 |
-
"mean_token_accuracy": 0.
|
| 718 |
"num_tokens": 9357050.0,
|
| 719 |
"step": 142
|
| 720 |
},
|
| 721 |
{
|
| 722 |
"epoch": 2.8096385542168676,
|
| 723 |
-
"grad_norm": 0.
|
| 724 |
-
"learning_rate": 4.
|
| 725 |
-
"loss": 0.
|
| 726 |
-
"mean_token_accuracy": 0.
|
| 727 |
"num_tokens": 9488122.0,
|
| 728 |
"step": 144
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 2.8481927710843373,
|
| 732 |
-
"grad_norm": 0.
|
| 733 |
-
"learning_rate": 4.
|
| 734 |
-
"loss": 0.
|
| 735 |
-
"mean_token_accuracy": 0.
|
| 736 |
"num_tokens": 9619194.0,
|
| 737 |
"step": 146
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"epoch": 2.886746987951807,
|
| 741 |
-
"grad_norm": 0.
|
| 742 |
-
"learning_rate": 4.
|
| 743 |
-
"loss": 0.
|
| 744 |
-
"mean_token_accuracy": 0.
|
| 745 |
"num_tokens": 9750266.0,
|
| 746 |
"step": 148
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 2.9253012048192772,
|
| 750 |
-
"grad_norm": 0.
|
| 751 |
-
"learning_rate": 4.
|
| 752 |
-
"loss": 0.
|
| 753 |
-
"mean_token_accuracy": 0.
|
| 754 |
"num_tokens": 9881338.0,
|
| 755 |
"step": 150
|
| 756 |
},
|
| 757 |
{
|
| 758 |
"epoch": 2.963855421686747,
|
| 759 |
-
"grad_norm": 0.
|
| 760 |
-
"learning_rate":
|
| 761 |
-
"loss": 0.
|
| 762 |
-
"mean_token_accuracy": 0.
|
| 763 |
"num_tokens": 10011972.0,
|
| 764 |
"step": 152
|
| 765 |
},
|
| 766 |
{
|
| 767 |
"epoch": 3.019277108433735,
|
| 768 |
-
"grad_norm":
|
| 769 |
-
"learning_rate":
|
| 770 |
-
"loss": 0.
|
| 771 |
-
"mean_token_accuracy": 0.
|
| 772 |
"num_tokens": 10175812.0,
|
| 773 |
"step": 154
|
| 774 |
},
|
| 775 |
{
|
| 776 |
"epoch": 3.057831325301205,
|
| 777 |
-
"grad_norm": 0.
|
| 778 |
-
"learning_rate":
|
| 779 |
-
"loss": 0.
|
| 780 |
-
"mean_token_accuracy": 0.
|
| 781 |
"num_tokens": 10306884.0,
|
| 782 |
"step": 156
|
| 783 |
},
|
| 784 |
{
|
| 785 |
"epoch": 3.0963855421686746,
|
| 786 |
-
"grad_norm": 0.
|
| 787 |
-
"learning_rate": 3.
|
| 788 |
-
"loss": 0.
|
| 789 |
-
"mean_token_accuracy": 0.
|
| 790 |
"num_tokens": 10437883.0,
|
| 791 |
"step": 158
|
| 792 |
},
|
| 793 |
{
|
| 794 |
"epoch": 3.1349397590361447,
|
| 795 |
-
"grad_norm": 0.
|
| 796 |
-
"learning_rate": 3.
|
| 797 |
-
"loss": 0.
|
| 798 |
-
"mean_token_accuracy": 0.
|
| 799 |
"num_tokens": 10568955.0,
|
| 800 |
"step": 160
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 3.1349397590361447,
|
| 804 |
-
"eval_loss": 0.
|
| 805 |
-
"eval_mean_token_accuracy": 0.
|
| 806 |
"eval_num_tokens": 10568955.0,
|
| 807 |
-
"eval_runtime": 69.
|
| 808 |
-
"eval_samples_per_second": 12.
|
| 809 |
-
"eval_steps_per_second": 1.
|
| 810 |
"step": 160
|
| 811 |
},
|
| 812 |
{
|
| 813 |
"epoch": 3.1734939759036145,
|
| 814 |
-
"grad_norm": 0.
|
| 815 |
-
"learning_rate": 3.
|
| 816 |
-
"loss": 0.
|
| 817 |
-
"mean_token_accuracy": 0.
|
| 818 |
"num_tokens": 10700027.0,
|
| 819 |
"step": 162
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"epoch": 3.212048192771084,
|
| 823 |
-
"grad_norm": 0.
|
| 824 |
-
"learning_rate": 3.
|
| 825 |
-
"loss": 0.
|
| 826 |
-
"mean_token_accuracy": 0.
|
| 827 |
"num_tokens": 10831099.0,
|
| 828 |
"step": 164
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"epoch": 3.2506024096385544,
|
| 832 |
-
"grad_norm": 0.
|
| 833 |
-
"learning_rate": 3.
|
| 834 |
-
"loss": 0.
|
| 835 |
-
"mean_token_accuracy": 0.
|
| 836 |
"num_tokens": 10962171.0,
|
| 837 |
"step": 166
|
| 838 |
},
|
| 839 |
{
|
| 840 |
"epoch": 3.289156626506024,
|
| 841 |
-
"grad_norm": 0.
|
| 842 |
-
"learning_rate":
|
| 843 |
-
"loss": 0.
|
| 844 |
-
"mean_token_accuracy": 0.
|
| 845 |
"num_tokens": 11092367.0,
|
| 846 |
"step": 168
|
| 847 |
},
|
| 848 |
{
|
| 849 |
"epoch": 3.327710843373494,
|
| 850 |
-
"grad_norm": 0.
|
| 851 |
-
"learning_rate":
|
| 852 |
-
"loss": 0.
|
| 853 |
-
"mean_token_accuracy": 0.
|
| 854 |
"num_tokens": 11223439.0,
|
| 855 |
"step": 170
|
| 856 |
},
|
| 857 |
{
|
| 858 |
"epoch": 3.3662650602409636,
|
| 859 |
-
"grad_norm": 0.
|
| 860 |
-
"learning_rate": 2.
|
| 861 |
-
"loss": 0.
|
| 862 |
-
"mean_token_accuracy": 0.
|
| 863 |
"num_tokens": 11354511.0,
|
| 864 |
"step": 172
|
| 865 |
},
|
| 866 |
{
|
| 867 |
"epoch": 3.404819277108434,
|
| 868 |
-
"grad_norm": 0.
|
| 869 |
-
"learning_rate": 2.
|
| 870 |
-
"loss": 0.
|
| 871 |
-
"mean_token_accuracy": 0.
|
| 872 |
"num_tokens": 11485583.0,
|
| 873 |
"step": 174
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 3.4433734939759035,
|
| 877 |
-
"grad_norm": 0.
|
| 878 |
-
"learning_rate": 2.
|
| 879 |
-
"loss": 0.
|
| 880 |
-
"mean_token_accuracy": 0.
|
| 881 |
"num_tokens": 11616217.0,
|
| 882 |
"step": 176
|
| 883 |
},
|
| 884 |
{
|
| 885 |
"epoch": 3.4819277108433733,
|
| 886 |
-
"grad_norm": 0.
|
| 887 |
-
"learning_rate": 2.
|
| 888 |
-
"loss": 0.
|
| 889 |
-
"mean_token_accuracy": 0.
|
| 890 |
"num_tokens": 11747289.0,
|
| 891 |
"step": 178
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 3.5204819277108435,
|
| 895 |
-
"grad_norm": 0.
|
| 896 |
-
"learning_rate": 2.
|
| 897 |
"loss": 0.0777,
|
| 898 |
-
"mean_token_accuracy": 0.
|
| 899 |
"num_tokens": 11878361.0,
|
| 900 |
"step": 180
|
| 901 |
},
|
| 902 |
{
|
| 903 |
"epoch": 3.5204819277108435,
|
| 904 |
-
"eval_loss": 0.
|
| 905 |
-
"eval_mean_token_accuracy": 0.
|
| 906 |
"eval_num_tokens": 11878361.0,
|
| 907 |
-
"eval_runtime": 69.
|
| 908 |
-
"eval_samples_per_second": 12.
|
| 909 |
-
"eval_steps_per_second": 1.
|
| 910 |
"step": 180
|
| 911 |
},
|
| 912 |
{
|
| 913 |
"epoch": 3.559036144578313,
|
| 914 |
-
"grad_norm": 0.
|
| 915 |
-
"learning_rate": 2.
|
| 916 |
-
"loss": 0.
|
| 917 |
-
"mean_token_accuracy": 0.
|
| 918 |
"num_tokens": 12009433.0,
|
| 919 |
"step": 182
|
| 920 |
},
|
| 921 |
{
|
| 922 |
"epoch": 3.597590361445783,
|
| 923 |
-
"grad_norm": 0.
|
| 924 |
-
"learning_rate": 2.
|
| 925 |
-
"loss": 0.
|
| 926 |
-
"mean_token_accuracy": 0.
|
| 927 |
"num_tokens": 12140505.0,
|
| 928 |
"step": 184
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 3.636144578313253,
|
| 932 |
-
"grad_norm": 0.
|
| 933 |
-
"learning_rate":
|
| 934 |
-
"loss": 0.
|
| 935 |
-
"mean_token_accuracy": 0.
|
| 936 |
"num_tokens": 12271577.0,
|
| 937 |
"step": 186
|
| 938 |
},
|
| 939 |
{
|
| 940 |
"epoch": 3.674698795180723,
|
| 941 |
-
"grad_norm": 0.
|
| 942 |
-
"learning_rate":
|
| 943 |
-
"loss": 0.
|
| 944 |
-
"mean_token_accuracy": 0.
|
| 945 |
"num_tokens": 12402649.0,
|
| 946 |
"step": 188
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 3.7132530120481926,
|
| 950 |
-
"grad_norm": 0.
|
| 951 |
-
"learning_rate": 1.
|
| 952 |
-
"loss": 0.
|
| 953 |
-
"mean_token_accuracy": 0.
|
| 954 |
"num_tokens": 12533721.0,
|
| 955 |
"step": 190
|
| 956 |
},
|
| 957 |
{
|
| 958 |
"epoch": 3.7518072289156628,
|
| 959 |
-
"grad_norm": 0.
|
| 960 |
-
"learning_rate": 1.
|
| 961 |
-
"loss": 0.
|
| 962 |
-
"mean_token_accuracy": 0.
|
| 963 |
"num_tokens": 12663435.0,
|
| 964 |
"step": 192
|
| 965 |
},
|
| 966 |
{
|
| 967 |
"epoch": 3.7903614457831325,
|
| 968 |
-
"grad_norm": 0.
|
| 969 |
-
"learning_rate": 1.
|
| 970 |
-
"loss": 0.
|
| 971 |
-
"mean_token_accuracy": 0.
|
| 972 |
"num_tokens": 12794507.0,
|
| 973 |
"step": 194
|
| 974 |
},
|
| 975 |
{
|
| 976 |
"epoch": 3.8289156626506022,
|
| 977 |
-
"grad_norm": 0.
|
| 978 |
-
"learning_rate": 1.
|
| 979 |
-
"loss": 0.
|
| 980 |
-
"mean_token_accuracy": 0.
|
| 981 |
"num_tokens": 12925579.0,
|
| 982 |
"step": 196
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 3.8674698795180724,
|
| 986 |
-
"grad_norm": 0.
|
| 987 |
-
"learning_rate": 1.
|
| 988 |
-
"loss": 0.
|
| 989 |
-
"mean_token_accuracy": 0.
|
| 990 |
"num_tokens": 13056651.0,
|
| 991 |
"step": 198
|
| 992 |
},
|
| 993 |
{
|
| 994 |
"epoch": 3.906024096385542,
|
| 995 |
-
"grad_norm": 0.
|
| 996 |
-
"learning_rate": 1.
|
| 997 |
-
"loss": 0.
|
| 998 |
-
"mean_token_accuracy": 0.
|
| 999 |
"num_tokens": 13187723.0,
|
| 1000 |
"step": 200
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 3.906024096385542,
|
| 1004 |
-
"eval_loss": 0.
|
| 1005 |
-
"eval_mean_token_accuracy": 0.
|
| 1006 |
"eval_num_tokens": 13187723.0,
|
| 1007 |
-
"eval_runtime": 69.
|
| 1008 |
-
"eval_samples_per_second": 12.
|
| 1009 |
-
"eval_steps_per_second": 1.
|
| 1010 |
"step": 200
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 3.944578313253012,
|
| 1014 |
-
"grad_norm": 0.
|
| 1015 |
-
"learning_rate": 1.
|
| 1016 |
-
"loss": 0.
|
| 1017 |
-
"mean_token_accuracy": 0.
|
| 1018 |
"num_tokens": 13317686.0,
|
| 1019 |
"step": 202
|
| 1020 |
},
|
| 1021 |
{
|
| 1022 |
"epoch": 3.983132530120482,
|
| 1023 |
-
"grad_norm": 0.
|
| 1024 |
-
"learning_rate": 1.
|
| 1025 |
-
"loss": 0.
|
| 1026 |
-
"mean_token_accuracy": 0.
|
| 1027 |
"num_tokens": 13447600.0,
|
| 1028 |
"step": 204
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
"epoch": 4.03855421686747,
|
| 1032 |
-
"grad_norm": 0.
|
| 1033 |
-
"learning_rate": 1.
|
| 1034 |
-
"loss": 0.
|
| 1035 |
-
"mean_token_accuracy": 0.
|
| 1036 |
"num_tokens": 13611440.0,
|
| 1037 |
"step": 206
|
| 1038 |
},
|
| 1039 |
{
|
| 1040 |
"epoch": 4.0771084337349395,
|
| 1041 |
-
"grad_norm": 0.
|
| 1042 |
-
"learning_rate":
|
| 1043 |
-
"loss": 0.
|
| 1044 |
-
"mean_token_accuracy": 0.
|
| 1045 |
"num_tokens": 13740537.0,
|
| 1046 |
"step": 208
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
"epoch": 4.11566265060241,
|
| 1050 |
-
"grad_norm": 0.
|
| 1051 |
-
"learning_rate":
|
| 1052 |
-
"loss": 0.
|
| 1053 |
-
"mean_token_accuracy": 0.
|
| 1054 |
"num_tokens": 13871609.0,
|
| 1055 |
"step": 210
|
| 1056 |
},
|
| 1057 |
{
|
| 1058 |
"epoch": 4.15421686746988,
|
| 1059 |
-
"grad_norm": 0.
|
| 1060 |
-
"learning_rate":
|
| 1061 |
-
"loss": 0.
|
| 1062 |
-
"mean_token_accuracy": 0.
|
| 1063 |
"num_tokens": 14002681.0,
|
| 1064 |
"step": 212
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
"epoch": 4.192771084337349,
|
| 1068 |
-
"grad_norm": 0.
|
| 1069 |
-
"learning_rate":
|
| 1070 |
-
"loss": 0.
|
| 1071 |
-
"mean_token_accuracy": 0.
|
| 1072 |
"num_tokens": 14133753.0,
|
| 1073 |
"step": 214
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
"epoch": 4.231325301204819,
|
| 1077 |
-
"grad_norm": 0.
|
| 1078 |
-
"learning_rate":
|
| 1079 |
-
"loss": 0.
|
| 1080 |
-
"mean_token_accuracy": 0.
|
| 1081 |
"num_tokens": 14264314.0,
|
| 1082 |
"step": 216
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
"epoch": 4.2698795180722895,
|
| 1086 |
-
"grad_norm": 0.
|
| 1087 |
-
"learning_rate":
|
| 1088 |
-
"loss": 0.
|
| 1089 |
-
"mean_token_accuracy": 0.
|
| 1090 |
"num_tokens": 14395386.0,
|
| 1091 |
"step": 218
|
| 1092 |
},
|
| 1093 |
{
|
| 1094 |
"epoch": 4.308433734939759,
|
| 1095 |
-
"grad_norm": 0.
|
| 1096 |
-
"learning_rate": 5.
|
| 1097 |
-
"loss": 0.
|
| 1098 |
-
"mean_token_accuracy": 0.
|
| 1099 |
"num_tokens": 14526458.0,
|
| 1100 |
"step": 220
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
"epoch": 4.308433734939759,
|
| 1104 |
-
"eval_loss": 0.
|
| 1105 |
-
"eval_mean_token_accuracy": 0.
|
| 1106 |
"eval_num_tokens": 14526458.0,
|
| 1107 |
-
"eval_runtime": 69.
|
| 1108 |
-
"eval_samples_per_second": 12.
|
| 1109 |
-
"eval_steps_per_second": 1.
|
| 1110 |
"step": 220
|
| 1111 |
},
|
| 1112 |
{
|
| 1113 |
"epoch": 4.346987951807229,
|
| 1114 |
-
"grad_norm": 0.
|
| 1115 |
-
"learning_rate":
|
| 1116 |
-
"loss": 0.
|
| 1117 |
-
"mean_token_accuracy": 0.
|
| 1118 |
"num_tokens": 14656421.0,
|
| 1119 |
"step": 222
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
"epoch": 4.385542168674699,
|
| 1123 |
-
"grad_norm": 0.
|
| 1124 |
-
"learning_rate": 4.
|
| 1125 |
-
"loss": 0.
|
| 1126 |
-
"mean_token_accuracy": 0.
|
| 1127 |
"num_tokens": 14787493.0,
|
| 1128 |
"step": 224
|
| 1129 |
},
|
| 1130 |
{
|
| 1131 |
"epoch": 4.424096385542168,
|
| 1132 |
-
"grad_norm": 0.
|
| 1133 |
-
"learning_rate":
|
| 1134 |
-
"loss": 0.
|
| 1135 |
-
"mean_token_accuracy": 0.
|
| 1136 |
"num_tokens": 14918565.0,
|
| 1137 |
"step": 226
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
"epoch": 4.462650602409639,
|
| 1141 |
-
"grad_norm": 0.
|
| 1142 |
-
"learning_rate": 3.
|
| 1143 |
-
"loss": 0.
|
| 1144 |
-
"mean_token_accuracy": 0.
|
| 1145 |
"num_tokens": 15049637.0,
|
| 1146 |
"step": 228
|
| 1147 |
},
|
| 1148 |
{
|
| 1149 |
"epoch": 4.501204819277109,
|
| 1150 |
-
"grad_norm": 0.
|
| 1151 |
-
"learning_rate":
|
| 1152 |
-
"loss": 0.
|
| 1153 |
-
"mean_token_accuracy": 0.
|
| 1154 |
"num_tokens": 15180709.0,
|
| 1155 |
"step": 230
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
"epoch": 4.539759036144578,
|
| 1159 |
-
"grad_norm": 0.
|
| 1160 |
-
"learning_rate": 2.
|
| 1161 |
-
"loss": 0.
|
| 1162 |
-
"mean_token_accuracy": 0.
|
| 1163 |
"num_tokens": 15311781.0,
|
| 1164 |
"step": 232
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 4.578313253012048,
|
| 1168 |
-
"grad_norm": 0.
|
| 1169 |
-
"learning_rate": 2.
|
| 1170 |
-
"loss": 0.
|
| 1171 |
-
"mean_token_accuracy": 0.
|
| 1172 |
"num_tokens": 15442853.0,
|
| 1173 |
"step": 234
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
"epoch": 4.6168674698795185,
|
| 1177 |
-
"grad_norm": 0.
|
| 1178 |
-
"learning_rate": 1.
|
| 1179 |
-
"loss": 0.
|
| 1180 |
-
"mean_token_accuracy": 0.
|
| 1181 |
"num_tokens": 15573925.0,
|
| 1182 |
"step": 236
|
| 1183 |
},
|
| 1184 |
{
|
| 1185 |
"epoch": 4.655421686746988,
|
| 1186 |
-
"grad_norm": 0.
|
| 1187 |
-
"learning_rate": 1.
|
| 1188 |
-
"loss": 0.
|
| 1189 |
-
"mean_token_accuracy": 0.
|
| 1190 |
"num_tokens": 15704997.0,
|
| 1191 |
"step": 238
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
"epoch": 4.693975903614458,
|
| 1195 |
-
"grad_norm": 0.
|
| 1196 |
-
"learning_rate": 1.
|
| 1197 |
-
"loss": 0.
|
| 1198 |
-
"mean_token_accuracy": 0.
|
| 1199 |
"num_tokens": 15836069.0,
|
| 1200 |
"step": 240
|
| 1201 |
},
|
| 1202 |
{
|
| 1203 |
"epoch": 4.693975903614458,
|
| 1204 |
-
"eval_loss": 0.
|
| 1205 |
-
"eval_mean_token_accuracy": 0.
|
| 1206 |
"eval_num_tokens": 15836069.0,
|
| 1207 |
-
"eval_runtime": 69.
|
| 1208 |
-
"eval_samples_per_second": 12.
|
| 1209 |
-
"eval_steps_per_second": 1.
|
| 1210 |
"step": 240
|
| 1211 |
},
|
| 1212 |
{
|
| 1213 |
"epoch": 4.732530120481927,
|
| 1214 |
-
"grad_norm": 0.
|
| 1215 |
-
"learning_rate":
|
| 1216 |
-
"loss": 0.
|
| 1217 |
-
"mean_token_accuracy": 0.
|
| 1218 |
"num_tokens": 15967141.0,
|
| 1219 |
"step": 242
|
| 1220 |
},
|
| 1221 |
{
|
| 1222 |
"epoch": 4.771084337349397,
|
| 1223 |
-
"grad_norm": 0.
|
| 1224 |
-
"learning_rate": 6.
|
| 1225 |
-
"loss": 0.
|
| 1226 |
-
"mean_token_accuracy": 0.
|
| 1227 |
"num_tokens": 16098213.0,
|
| 1228 |
"step": 244
|
| 1229 |
},
|
| 1230 |
{
|
| 1231 |
"epoch": 4.809638554216868,
|
| 1232 |
-
"grad_norm": 0.
|
| 1233 |
-
"learning_rate": 4.
|
| 1234 |
-
"loss": 0.
|
| 1235 |
-
"mean_token_accuracy": 0.
|
| 1236 |
"num_tokens": 16229285.0,
|
| 1237 |
"step": 246
|
| 1238 |
},
|
| 1239 |
{
|
| 1240 |
"epoch": 4.848192771084337,
|
| 1241 |
-
"grad_norm": 0.
|
| 1242 |
-
"learning_rate":
|
| 1243 |
-
"loss": 0.
|
| 1244 |
-
"mean_token_accuracy": 0.
|
| 1245 |
"num_tokens": 16360357.0,
|
| 1246 |
"step": 248
|
| 1247 |
},
|
| 1248 |
{
|
| 1249 |
"epoch": 4.886746987951807,
|
| 1250 |
-
"grad_norm": 0.
|
| 1251 |
-
"learning_rate": 1.
|
| 1252 |
-
"loss": 0.
|
| 1253 |
-
"mean_token_accuracy": 0.
|
| 1254 |
"num_tokens": 16491429.0,
|
| 1255 |
"step": 250
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 4.925301204819277,
|
| 1259 |
-
"grad_norm": 0.
|
| 1260 |
-
"learning_rate":
|
| 1261 |
-
"loss": 0.
|
| 1262 |
-
"mean_token_accuracy": 0.
|
| 1263 |
"num_tokens": 16621960.0,
|
| 1264 |
"step": 252
|
| 1265 |
},
|
| 1266 |
{
|
| 1267 |
"epoch": 4.9638554216867465,
|
| 1268 |
-
"grad_norm": 0.
|
| 1269 |
-
"learning_rate": 1.
|
| 1270 |
-
"loss": 0.
|
| 1271 |
-
"mean_token_accuracy": 0.
|
| 1272 |
"num_tokens": 16752156.0,
|
| 1273 |
"step": 254
|
| 1274 |
},
|
| 1275 |
{
|
| 1276 |
"epoch": 4.983132530120482,
|
| 1277 |
-
"mean_token_accuracy": 0.
|
| 1278 |
"num_tokens": 16817692.0,
|
| 1279 |
"step": 255,
|
| 1280 |
"total_flos": 24409842647040.0,
|
| 1281 |
-
"train_loss": 0.
|
| 1282 |
-
"train_runtime":
|
| 1283 |
"train_samples_per_second": 2.194,
|
| 1284 |
"train_steps_per_second": 0.067
|
| 1285 |
}
|
|
|
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 0.03855421686746988,
|
| 14 |
+
"grad_norm": 2.0457221564142256,
|
| 15 |
+
"learning_rate": 3.846153846153847e-07,
|
| 16 |
"loss": 0.2354,
|
| 17 |
"mean_token_accuracy": 0.930065356194973,
|
| 18 |
"num_tokens": 131072.0,
|
|
|
|
| 20 |
},
|
| 21 |
{
|
| 22 |
"epoch": 0.07710843373493977,
|
| 23 |
+
"grad_norm": 2.1086974646270145,
|
| 24 |
+
"learning_rate": 1.153846153846154e-06,
|
| 25 |
+
"loss": 0.2508,
|
| 26 |
+
"mean_token_accuracy": 0.9255465492606163,
|
| 27 |
"num_tokens": 262144.0,
|
| 28 |
"step": 4
|
| 29 |
},
|
| 30 |
{
|
| 31 |
"epoch": 0.11566265060240964,
|
| 32 |
+
"grad_norm": 1.698182282959437,
|
| 33 |
+
"learning_rate": 1.9230769230769234e-06,
|
| 34 |
+
"loss": 0.2473,
|
| 35 |
+
"mean_token_accuracy": 0.9256381466984749,
|
| 36 |
"num_tokens": 393216.0,
|
| 37 |
"step": 6
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.15421686746987953,
|
| 41 |
+
"grad_norm": 1.4331583326698771,
|
| 42 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 43 |
+
"loss": 0.2193,
|
| 44 |
+
"mean_token_accuracy": 0.9314393177628517,
|
| 45 |
"num_tokens": 524288.0,
|
| 46 |
"step": 8
|
| 47 |
},
|
| 48 |
{
|
| 49 |
"epoch": 0.1927710843373494,
|
| 50 |
+
"grad_norm": 1.280978852144958,
|
| 51 |
+
"learning_rate": 3.4615384615384617e-06,
|
| 52 |
+
"loss": 0.2205,
|
| 53 |
+
"mean_token_accuracy": 0.930450152605772,
|
| 54 |
"num_tokens": 654484.0,
|
| 55 |
"step": 10
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"epoch": 0.23132530120481928,
|
| 59 |
+
"grad_norm": 0.8255955634911271,
|
| 60 |
+
"learning_rate": 4.230769230769231e-06,
|
| 61 |
+
"loss": 0.2117,
|
| 62 |
+
"mean_token_accuracy": 0.9317141100764275,
|
| 63 |
"num_tokens": 785556.0,
|
| 64 |
"step": 12
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"epoch": 0.26987951807228916,
|
| 68 |
+
"grad_norm": 0.7584680371226415,
|
| 69 |
+
"learning_rate": 5e-06,
|
| 70 |
+
"loss": 0.206,
|
| 71 |
+
"mean_token_accuracy": 0.9338631108403206,
|
| 72 |
"num_tokens": 915519.0,
|
| 73 |
"step": 14
|
| 74 |
},
|
| 75 |
{
|
| 76 |
"epoch": 0.30843373493975906,
|
| 77 |
+
"grad_norm": 0.9495192852210463,
|
| 78 |
+
"learning_rate": 5.769230769230769e-06,
|
| 79 |
+
"loss": 0.1982,
|
| 80 |
+
"mean_token_accuracy": 0.9358359947800636,
|
| 81 |
"num_tokens": 1046591.0,
|
| 82 |
"step": 16
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"epoch": 0.3469879518072289,
|
| 86 |
+
"grad_norm": 0.9714974283482016,
|
| 87 |
+
"learning_rate": 6.538461538461539e-06,
|
| 88 |
+
"loss": 0.2055,
|
| 89 |
+
"mean_token_accuracy": 0.9338132180273533,
|
| 90 |
"num_tokens": 1177663.0,
|
| 91 |
"step": 18
|
| 92 |
},
|
| 93 |
{
|
| 94 |
"epoch": 0.3855421686746988,
|
| 95 |
+
"grad_norm": 0.6339236056292388,
|
| 96 |
+
"learning_rate": 7.307692307692308e-06,
|
| 97 |
+
"loss": 0.1917,
|
| 98 |
+
"mean_token_accuracy": 0.9378740377724171,
|
| 99 |
"num_tokens": 1308735.0,
|
| 100 |
"step": 20
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.3855421686746988,
|
| 104 |
+
"eval_loss": 0.3343917727470398,
|
| 105 |
+
"eval_mean_token_accuracy": 0.9013295725127247,
|
| 106 |
"eval_num_tokens": 1308735.0,
|
| 107 |
+
"eval_runtime": 70.0593,
|
| 108 |
+
"eval_samples_per_second": 12.204,
|
| 109 |
+
"eval_steps_per_second": 1.527,
|
| 110 |
"step": 20
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.42409638554216866,
|
| 114 |
+
"grad_norm": 0.7315888499202351,
|
| 115 |
+
"learning_rate": 8.076923076923077e-06,
|
| 116 |
+
"loss": 0.1809,
|
| 117 |
+
"mean_token_accuracy": 0.9400189444422722,
|
| 118 |
"num_tokens": 1439807.0,
|
| 119 |
"step": 22
|
| 120 |
},
|
| 121 |
{
|
| 122 |
"epoch": 0.46265060240963857,
|
| 123 |
+
"grad_norm": 0.7642349616310066,
|
| 124 |
+
"learning_rate": 8.846153846153847e-06,
|
| 125 |
+
"loss": 0.1928,
|
| 126 |
+
"mean_token_accuracy": 0.9367095269262791,
|
| 127 |
"num_tokens": 1570062.0,
|
| 128 |
"step": 24
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.5012048192771085,
|
| 132 |
+
"grad_norm": 0.6114978913375759,
|
| 133 |
+
"learning_rate": 9.615384615384616e-06,
|
| 134 |
+
"loss": 0.1828,
|
| 135 |
+
"mean_token_accuracy": 0.9394693598151207,
|
| 136 |
"num_tokens": 1701134.0,
|
| 137 |
"step": 26
|
| 138 |
},
|
| 139 |
{
|
| 140 |
"epoch": 0.5397590361445783,
|
| 141 |
+
"grad_norm": 0.6229653774047121,
|
| 142 |
+
"learning_rate": 9.999529497453782e-06,
|
| 143 |
+
"loss": 0.1806,
|
| 144 |
+
"mean_token_accuracy": 0.9402282536029816,
|
| 145 |
"num_tokens": 1832133.0,
|
| 146 |
"step": 28
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.5783132530120482,
|
| 150 |
+
"grad_norm": 0.6722415161460822,
|
| 151 |
+
"learning_rate": 9.99576600836172e-06,
|
| 152 |
+
"loss": 0.1896,
|
| 153 |
+
"mean_token_accuracy": 0.9363855794072151,
|
| 154 |
"num_tokens": 1963205.0,
|
| 155 |
"step": 30
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 0.6168674698795181,
|
| 159 |
+
"grad_norm": 0.5974286474799401,
|
| 160 |
+
"learning_rate": 9.988241863214212e-06,
|
| 161 |
+
"loss": 0.1814,
|
| 162 |
+
"mean_token_accuracy": 0.9404540322721004,
|
| 163 |
"num_tokens": 2094277.0,
|
| 164 |
"step": 32
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.655421686746988,
|
| 168 |
+
"grad_norm": 0.601035342701654,
|
| 169 |
+
"learning_rate": 9.976962725951878e-06,
|
| 170 |
+
"loss": 0.1801,
|
| 171 |
+
"mean_token_accuracy": 0.9400342106819153,
|
| 172 |
"num_tokens": 2225349.0,
|
| 173 |
"step": 34
|
| 174 |
},
|
| 175 |
{
|
| 176 |
"epoch": 0.6939759036144578,
|
| 177 |
+
"grad_norm": 0.5765003488310966,
|
| 178 |
+
"learning_rate": 9.961937087155697e-06,
|
| 179 |
+
"loss": 0.1828,
|
| 180 |
+
"mean_token_accuracy": 0.9392519034445286,
|
| 181 |
"num_tokens": 2355263.0,
|
| 182 |
"step": 36
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.7325301204819277,
|
| 186 |
+
"grad_norm": 34.52047518558373,
|
| 187 |
+
"learning_rate": 9.943176257655567e-06,
|
| 188 |
+
"loss": 0.2098,
|
| 189 |
+
"mean_token_accuracy": 0.9331491328775883,
|
| 190 |
"num_tokens": 2486335.0,
|
| 191 |
"step": 38
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.7710843373493976,
|
| 195 |
+
"grad_norm": 0.6276699276820382,
|
| 196 |
+
"learning_rate": 9.920694360015864e-06,
|
| 197 |
+
"loss": 0.1745,
|
| 198 |
+
"mean_token_accuracy": 0.9413929060101509,
|
| 199 |
"num_tokens": 2617407.0,
|
| 200 |
"step": 40
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.7710843373493976,
|
| 204 |
+
"eval_loss": 0.32280808687210083,
|
| 205 |
+
"eval_mean_token_accuracy": 0.9021720039510281,
|
| 206 |
"eval_num_tokens": 2617407.0,
|
| 207 |
+
"eval_runtime": 69.6577,
|
| 208 |
+
"eval_samples_per_second": 12.274,
|
| 209 |
"eval_steps_per_second": 1.536,
|
| 210 |
"step": 40
|
| 211 |
},
|
| 212 |
{
|
| 213 |
"epoch": 0.8096385542168675,
|
| 214 |
+
"grad_norm": 0.6015365123041743,
|
| 215 |
+
"learning_rate": 9.894508317904418e-06,
|
| 216 |
+
"loss": 0.1751,
|
| 217 |
+
"mean_token_accuracy": 0.9412707760930061,
|
| 218 |
"num_tokens": 2748479.0,
|
| 219 |
"step": 42
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.8481927710843373,
|
| 223 |
+
"grad_norm": 0.6316203175238668,
|
| 224 |
+
"learning_rate": 9.864637843352916e-06,
|
| 225 |
+
"loss": 0.184,
|
| 226 |
+
"mean_token_accuracy": 0.9374923817813396,
|
| 227 |
"num_tokens": 2879551.0,
|
| 228 |
"step": 44
|
| 229 |
},
|
| 230 |
{
|
| 231 |
"epoch": 0.8867469879518072,
|
| 232 |
+
"grad_norm": 0.5904610746669308,
|
| 233 |
+
"learning_rate": 9.831105421918287e-06,
|
| 234 |
+
"loss": 0.1777,
|
| 235 |
+
"mean_token_accuracy": 0.9405580870807171,
|
| 236 |
"num_tokens": 3010185.0,
|
| 237 |
"step": 46
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.9253012048192771,
|
| 241 |
+
"grad_norm": 0.5994215271575196,
|
| 242 |
+
"learning_rate": 9.793936295756292e-06,
|
| 243 |
+
"loss": 0.187,
|
| 244 |
+
"mean_token_accuracy": 0.9375152811408043,
|
| 245 |
"num_tokens": 3141257.0,
|
| 246 |
"step": 48
|
| 247 |
},
|
| 248 |
{
|
| 249 |
"epoch": 0.963855421686747,
|
| 250 |
+
"grad_norm": 0.5854742456446934,
|
| 251 |
+
"learning_rate": 9.753158444620013e-06,
|
| 252 |
+
"loss": 0.1815,
|
| 253 |
+
"mean_token_accuracy": 0.9394976831972599,
|
| 254 |
"num_tokens": 3271788.0,
|
| 255 |
"step": 50
|
| 256 |
},
|
| 257 |
{
|
| 258 |
"epoch": 1.0192771084337349,
|
| 259 |
+
"grad_norm": 0.957499837849808,
|
| 260 |
+
"learning_rate": 9.70880256479758e-06,
|
| 261 |
+
"loss": 0.2534,
|
| 262 |
+
"mean_token_accuracy": 0.9437652796506881,
|
| 263 |
"num_tokens": 3435628.0,
|
| 264 |
"step": 52
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 1.0578313253012048,
|
| 268 |
+
"grad_norm": 0.6854514205992324,
|
| 269 |
+
"learning_rate": 9.660902046004954e-06,
|
| 270 |
+
"loss": 0.151,
|
| 271 |
+
"mean_token_accuracy": 0.9503083899617195,
|
| 272 |
"num_tokens": 3566700.0,
|
| 273 |
"step": 54
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 1.0963855421686748,
|
| 277 |
+
"grad_norm": 0.6080507225701574,
|
| 278 |
+
"learning_rate": 9.60949294625121e-06,
|
| 279 |
+
"loss": 0.1415,
|
| 280 |
+
"mean_token_accuracy": 0.9535066671669483,
|
| 281 |
"num_tokens": 3697772.0,
|
| 282 |
"step": 56
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 1.1349397590361445,
|
| 286 |
+
"grad_norm": 0.6054065882233389,
|
| 287 |
+
"learning_rate": 9.554613964695189e-06,
|
| 288 |
+
"loss": 0.1493,
|
| 289 |
+
"mean_token_accuracy": 0.9502549581229687,
|
| 290 |
"num_tokens": 3828844.0,
|
| 291 |
"step": 58
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 1.1734939759036145,
|
| 295 |
+
"grad_norm": 0.7694600057204949,
|
| 296 |
+
"learning_rate": 9.496306412513989e-06,
|
| 297 |
+
"loss": 0.1462,
|
| 298 |
+
"mean_token_accuracy": 0.9519953094422817,
|
| 299 |
"num_tokens": 3959916.0,
|
| 300 |
"step": 60
|
| 301 |
},
|
| 302 |
{
|
| 303 |
"epoch": 1.1734939759036145,
|
| 304 |
+
"eval_loss": 0.359206885099411,
|
| 305 |
+
"eval_mean_token_accuracy": 0.9007850846397543,
|
| 306 |
"eval_num_tokens": 3959916.0,
|
| 307 |
+
"eval_runtime": 69.8215,
|
| 308 |
+
"eval_samples_per_second": 12.246,
|
| 309 |
"eval_steps_per_second": 1.532,
|
| 310 |
"step": 60
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 1.2120481927710842,
|
| 314 |
+
"grad_norm": 0.6845669867023433,
|
| 315 |
+
"learning_rate": 9.434614181805203e-06,
|
| 316 |
+
"loss": 0.1407,
|
| 317 |
+
"mean_token_accuracy": 0.9533876590430737,
|
| 318 |
"num_tokens": 4089879.0,
|
| 319 |
"step": 62
|
| 320 |
},
|
| 321 |
{
|
| 322 |
"epoch": 1.2506024096385542,
|
| 323 |
+
"grad_norm": 0.6197114152379135,
|
| 324 |
+
"learning_rate": 9.369583712546322e-06,
|
| 325 |
+
"loss": 0.1349,
|
| 326 |
+
"mean_token_accuracy": 0.9554836452007294,
|
| 327 |
"num_tokens": 4220951.0,
|
| 328 |
"step": 64
|
| 329 |
},
|
| 330 |
{
|
| 331 |
"epoch": 1.2891566265060241,
|
| 332 |
+
"grad_norm": 0.6172158164875755,
|
| 333 |
+
"learning_rate": 9.30126395763618e-06,
|
| 334 |
+
"loss": 0.1535,
|
| 335 |
+
"mean_token_accuracy": 0.95006413012743,
|
| 336 |
"num_tokens": 4352023.0,
|
| 337 |
"step": 66
|
| 338 |
},
|
| 339 |
{
|
| 340 |
"epoch": 1.3277108433734939,
|
| 341 |
+
"grad_norm": 0.6409060214608714,
|
| 342 |
+
"learning_rate": 9.229706346044749e-06,
|
| 343 |
+
"loss": 0.156,
|
| 344 |
+
"mean_token_accuracy": 0.9484306424856186,
|
| 345 |
"num_tokens": 4483095.0,
|
| 346 |
"step": 68
|
| 347 |
},
|
| 348 |
{
|
| 349 |
"epoch": 1.3662650602409638,
|
| 350 |
+
"grad_norm": 0.6166450609513697,
|
| 351 |
+
"learning_rate": 9.154964744099006e-06,
|
| 352 |
+
"loss": 0.1419,
|
| 353 |
+
"mean_token_accuracy": 0.9533540047705173,
|
| 354 |
"num_tokens": 4614167.0,
|
| 355 |
"step": 70
|
| 356 |
},
|
| 357 |
{
|
| 358 |
"epoch": 1.4048192771084338,
|
| 359 |
+
"grad_norm": 0.6058092262037136,
|
| 360 |
+
"learning_rate": 9.077095414934076e-06,
|
| 361 |
+
"loss": 0.1439,
|
| 362 |
+
"mean_token_accuracy": 0.9524685628712177,
|
| 363 |
"num_tokens": 4745239.0,
|
| 364 |
"step": 72
|
| 365 |
},
|
| 366 |
{
|
| 367 |
"epoch": 1.4433734939759035,
|
| 368 |
+
"grad_norm": 0.6464674278239464,
|
| 369 |
+
"learning_rate": 8.996156976140088e-06,
|
| 370 |
+
"loss": 0.1427,
|
| 371 |
+
"mean_token_accuracy": 0.9521632380783558,
|
| 372 |
"num_tokens": 4876311.0,
|
| 373 |
"step": 74
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 1.4819277108433735,
|
| 377 |
+
"grad_norm": 0.6232124362016298,
|
| 378 |
+
"learning_rate": 8.91221035563669e-06,
|
| 379 |
+
"loss": 0.1387,
|
| 380 |
+
"mean_token_accuracy": 0.9537738263607025,
|
| 381 |
"num_tokens": 5007383.0,
|
| 382 |
"step": 76
|
| 383 |
},
|
| 384 |
{
|
| 385 |
"epoch": 1.5204819277108435,
|
| 386 |
+
"grad_norm": 0.6251055517263481,
|
| 387 |
+
"learning_rate": 8.82531874580844e-06,
|
| 388 |
+
"loss": 0.1544,
|
| 389 |
+
"mean_token_accuracy": 0.9496977403759956,
|
| 390 |
"num_tokens": 5138455.0,
|
| 391 |
"step": 78
|
| 392 |
},
|
| 393 |
{
|
| 394 |
"epoch": 1.5590361445783132,
|
| 395 |
+
"grad_norm": 0.6597130966145244,
|
| 396 |
+
"learning_rate": 8.735547555935538e-06,
|
| 397 |
+
"loss": 0.1467,
|
| 398 |
+
"mean_token_accuracy": 0.951957143843174,
|
| 399 |
"num_tokens": 5269527.0,
|
| 400 |
"step": 80
|
| 401 |
},
|
| 402 |
{
|
| 403 |
"epoch": 1.5590361445783132,
|
| 404 |
+
"eval_loss": 0.34304243326187134,
|
| 405 |
+
"eval_mean_token_accuracy": 0.9011661727851796,
|
| 406 |
"eval_num_tokens": 5269527.0,
|
| 407 |
+
"eval_runtime": 69.6573,
|
| 408 |
+
"eval_samples_per_second": 12.274,
|
| 409 |
"eval_steps_per_second": 1.536,
|
| 410 |
"step": 80
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 1.5975903614457831,
|
| 414 |
+
"grad_norm": 0.6093216234766912,
|
| 415 |
+
"learning_rate": 8.642964362955781e-06,
|
| 416 |
+
"loss": 0.145,
|
| 417 |
+
"mean_token_accuracy": 0.9515700563788414,
|
| 418 |
"num_tokens": 5400161.0,
|
| 419 |
"step": 82
|
| 420 |
},
|
| 421 |
{
|
| 422 |
"epoch": 1.636144578313253,
|
| 423 |
+
"grad_norm": 0.5687703380048487,
|
| 424 |
+
"learning_rate": 8.547638860594765e-06,
|
| 425 |
+
"loss": 0.1484,
|
| 426 |
+
"mean_token_accuracy": 0.9509495720267296,
|
| 427 |
"num_tokens": 5531233.0,
|
| 428 |
"step": 84
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 1.6746987951807228,
|
| 432 |
+
"grad_norm": 0.6551898466798518,
|
| 433 |
+
"learning_rate": 8.449642806902623e-06,
|
| 434 |
+
"loss": 0.1568,
|
| 435 |
+
"mean_token_accuracy": 0.9481558501720428,
|
| 436 |
"num_tokens": 5662305.0,
|
| 437 |
"step": 86
|
| 438 |
},
|
| 439 |
{
|
| 440 |
"epoch": 1.7132530120481928,
|
| 441 |
+
"grad_norm": 0.6433780292504243,
|
| 442 |
+
"learning_rate": 8.349049970236822e-06,
|
| 443 |
+
"loss": 0.1349,
|
| 444 |
+
"mean_token_accuracy": 0.954715259373188,
|
| 445 |
"num_tokens": 5792219.0,
|
| 446 |
"step": 88
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 1.7518072289156628,
|
| 450 |
+
"grad_norm": 0.5701046312406493,
|
| 451 |
+
"learning_rate": 8.245936073731654e-06,
|
| 452 |
+
"loss": 0.147,
|
| 453 |
+
"mean_token_accuracy": 0.9507969096302986,
|
| 454 |
"num_tokens": 5923291.0,
|
| 455 |
"step": 90
|
| 456 |
},
|
| 457 |
{
|
| 458 |
"epoch": 1.7903614457831325,
|
| 459 |
+
"grad_norm": 0.6865332623152001,
|
| 460 |
+
"learning_rate": 8.140378738296233e-06,
|
| 461 |
+
"loss": 0.1529,
|
| 462 |
+
"mean_token_accuracy": 0.9498768150806427,
|
| 463 |
"num_tokens": 6053822.0,
|
| 464 |
"step": 92
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 1.8289156626506025,
|
| 468 |
+
"grad_norm": 0.6305307568855328,
|
| 469 |
+
"learning_rate": 8.032457424183909e-06,
|
| 470 |
+
"loss": 0.1476,
|
| 471 |
+
"mean_token_accuracy": 0.9505984485149384,
|
| 472 |
"num_tokens": 6184894.0,
|
| 473 |
"step": 94
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 1.8674698795180724,
|
| 477 |
+
"grad_norm": 0.5748443476790706,
|
| 478 |
+
"learning_rate": 7.922253371177081e-06,
|
| 479 |
+
"loss": 0.155,
|
| 480 |
+
"mean_token_accuracy": 0.9482144415378571,
|
| 481 |
"num_tokens": 6315149.0,
|
| 482 |
"step": 96
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 1.9060240963855422,
|
| 486 |
+
"grad_norm": 0.5993128969226361,
|
| 487 |
+
"learning_rate": 7.809849537432432e-06,
|
| 488 |
+
"loss": 0.1434,
|
| 489 |
+
"mean_token_accuracy": 0.9525645859539509,
|
| 490 |
"num_tokens": 6445345.0,
|
| 491 |
"step": 98
|
| 492 |
},
|
| 493 |
{
|
| 494 |
"epoch": 1.944578313253012,
|
| 495 |
+
"grad_norm": 0.6280456904784001,
|
| 496 |
+
"learning_rate": 7.695330537032629e-06,
|
| 497 |
+
"loss": 0.1445,
|
| 498 |
+
"mean_token_accuracy": 0.9512222707271576,
|
| 499 |
"num_tokens": 6576344.0,
|
| 500 |
"step": 100
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 1.944578313253012,
|
| 504 |
+
"eval_loss": 0.3398211598396301,
|
| 505 |
+
"eval_mean_token_accuracy": 0.901328669530209,
|
| 506 |
"eval_num_tokens": 6576344.0,
|
| 507 |
+
"eval_runtime": 69.654,
|
| 508 |
+
"eval_samples_per_second": 12.275,
|
| 509 |
+
"eval_steps_per_second": 1.536,
|
| 510 |
"step": 100
|
| 511 |
},
|
| 512 |
{
|
| 513 |
"epoch": 1.983132530120482,
|
| 514 |
+
"grad_norm": 0.6197902890500856,
|
| 515 |
+
"learning_rate": 7.578782576291501e-06,
|
| 516 |
+
"loss": 0.1506,
|
| 517 |
+
"mean_token_accuracy": 0.9492092207074165,
|
| 518 |
"num_tokens": 6707416.0,
|
| 519 |
"step": 102
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 2.0385542168674697,
|
| 523 |
+
"grad_norm": 0.6409344863530665,
|
| 524 |
+
"learning_rate": 7.460293388860616e-06,
|
| 525 |
+
"loss": 0.1754,
|
| 526 |
+
"mean_token_accuracy": 0.9643502771854401,
|
| 527 |
"num_tokens": 6871256.0,
|
| 528 |
"step": 104
|
| 529 |
},
|
| 530 |
{
|
| 531 |
"epoch": 2.07710843373494,
|
| 532 |
+
"grad_norm": 0.6097248296204885,
|
| 533 |
+
"learning_rate": 7.3399521696861505e-06,
|
| 534 |
+
"loss": 0.1092,
|
| 535 |
+
"mean_token_accuracy": 0.9659219309687614,
|
| 536 |
"num_tokens": 7002255.0,
|
| 537 |
"step": 106
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 2.1156626506024097,
|
| 541 |
+
"grad_norm": 0.5903613108322504,
|
| 542 |
+
"learning_rate": 7.217849507865724e-06,
|
| 543 |
+
"loss": 0.1066,
|
| 544 |
+
"mean_token_accuracy": 0.9660860486328602,
|
| 545 |
"num_tokens": 7133327.0,
|
| 546 |
"step": 108
|
| 547 |
},
|
| 548 |
{
|
| 549 |
"epoch": 2.1542168674698794,
|
| 550 |
+
"grad_norm": 0.625091072426359,
|
| 551 |
+
"learning_rate": 7.094077318455762e-06,
|
| 552 |
+
"loss": 0.1091,
|
| 553 |
+
"mean_token_accuracy": 0.9645588099956512,
|
| 554 |
"num_tokens": 7263523.0,
|
| 555 |
"step": 110
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 2.1927710843373496,
|
| 559 |
+
"grad_norm": 0.6604015968164485,
|
| 560 |
+
"learning_rate": 6.96872877328073e-06,
|
| 561 |
+
"loss": 0.1052,
|
| 562 |
+
"mean_token_accuracy": 0.9661929123103619,
|
| 563 |
"num_tokens": 7394595.0,
|
| 564 |
"step": 112
|
| 565 |
},
|
| 566 |
{
|
| 567 |
"epoch": 2.2313253012048193,
|
| 568 |
+
"grad_norm": 0.7455880093770229,
|
| 569 |
+
"learning_rate": 6.841898230796302e-06,
|
| 570 |
+
"loss": 0.1049,
|
| 571 |
+
"mean_token_accuracy": 0.9661089479923248,
|
| 572 |
"num_tokens": 7525667.0,
|
| 573 |
"step": 114
|
| 574 |
},
|
| 575 |
{
|
| 576 |
"epoch": 2.269879518072289,
|
| 577 |
+
"grad_norm": 0.6028303919109465,
|
| 578 |
+
"learning_rate": 6.713681165059271e-06,
|
| 579 |
+
"loss": 0.1127,
|
| 580 |
+
"mean_token_accuracy": 0.9631625637412071,
|
| 581 |
"num_tokens": 7656739.0,
|
| 582 |
"step": 116
|
| 583 |
},
|
| 584 |
{
|
| 585 |
"epoch": 2.3084337349397592,
|
| 586 |
+
"grad_norm": 0.6799912009709536,
|
| 587 |
+
"learning_rate": 6.584174093857676e-06,
|
| 588 |
+
"loss": 0.1035,
|
| 589 |
+
"mean_token_accuracy": 0.9669562242925167,
|
| 590 |
"num_tokens": 7787811.0,
|
| 591 |
"step": 118
|
| 592 |
},
|
| 593 |
{
|
| 594 |
"epoch": 2.346987951807229,
|
| 595 |
+
"grad_norm": 0.6255570427114552,
|
| 596 |
+
"learning_rate": 6.453474506055228e-06,
|
| 597 |
+
"loss": 0.1176,
|
| 598 |
+
"mean_token_accuracy": 0.9615787602961063,
|
| 599 |
"num_tokens": 7916616.0,
|
| 600 |
"step": 120
|
| 601 |
},
|
| 602 |
{
|
| 603 |
"epoch": 2.346987951807229,
|
| 604 |
+
"eval_loss": 0.38193774223327637,
|
| 605 |
+
"eval_mean_token_accuracy": 0.8994210568543907,
|
| 606 |
"eval_num_tokens": 7916616.0,
|
| 607 |
+
"eval_runtime": 69.6436,
|
| 608 |
+
"eval_samples_per_second": 12.277,
|
| 609 |
+
"eval_steps_per_second": 1.536,
|
| 610 |
"step": 120
|
| 611 |
},
|
| 612 |
{
|
| 613 |
"epoch": 2.3855421686746987,
|
| 614 |
+
"grad_norm": 0.6279356138996781,
|
| 615 |
+
"learning_rate": 6.3216807882047585e-06,
|
| 616 |
+
"loss": 0.0974,
|
| 617 |
+
"mean_token_accuracy": 0.968185156583786,
|
| 618 |
"num_tokens": 8047688.0,
|
| 619 |
"step": 122
|
| 620 |
},
|
| 621 |
{
|
| 622 |
"epoch": 2.4240963855421684,
|
| 623 |
+
"grad_norm": 0.6479503216427691,
|
| 624 |
+
"learning_rate": 6.188892150485904e-06,
|
| 625 |
+
"loss": 0.1087,
|
| 626 |
+
"mean_token_accuracy": 0.9651853404939175,
|
| 627 |
"num_tokens": 8178760.0,
|
| 628 |
"step": 124
|
| 629 |
},
|
| 630 |
{
|
| 631 |
"epoch": 2.4626506024096386,
|
| 632 |
+
"grad_norm": 0.7228376218883897,
|
| 633 |
+
"learning_rate": 6.0552085520227875e-06,
|
| 634 |
+
"loss": 0.1136,
|
| 635 |
+
"mean_token_accuracy": 0.9631396643817425,
|
| 636 |
"num_tokens": 8309832.0,
|
| 637 |
"step": 126
|
| 638 |
},
|
| 639 |
{
|
| 640 |
"epoch": 2.5012048192771084,
|
| 641 |
+
"grad_norm": 0.6292530226739607,
|
| 642 |
+
"learning_rate": 5.920730625637934e-06,
|
| 643 |
+
"loss": 0.1043,
|
| 644 |
+
"mean_token_accuracy": 0.9666203670203686,
|
| 645 |
"num_tokens": 8440904.0,
|
| 646 |
"step": 128
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 2.539759036144578,
|
| 650 |
+
"grad_norm": 0.6120273359022707,
|
| 651 |
+
"learning_rate": 5.785559602099019e-06,
|
| 652 |
+
"loss": 0.1073,
|
| 653 |
+
"mean_token_accuracy": 0.9648876488208771,
|
| 654 |
"num_tokens": 8571976.0,
|
| 655 |
"step": 130
|
| 656 |
},
|
| 657 |
{
|
| 658 |
"epoch": 2.5783132530120483,
|
| 659 |
+
"grad_norm": 0.6294342722298523,
|
| 660 |
+
"learning_rate": 5.649797233915539e-06,
|
| 661 |
+
"loss": 0.1092,
|
| 662 |
+
"mean_token_accuracy": 0.9644067622721195,
|
| 663 |
"num_tokens": 8703048.0,
|
| 664 |
"step": 132
|
| 665 |
},
|
| 666 |
{
|
| 667 |
"epoch": 2.616867469879518,
|
| 668 |
+
"grad_norm": 0.5665304014502571,
|
| 669 |
+
"learning_rate": 5.513545718742702e-06,
|
| 670 |
+
"loss": 0.1086,
|
| 671 |
+
"mean_token_accuracy": 0.9646815545856953,
|
| 672 |
"num_tokens": 8834120.0,
|
| 673 |
"step": 134
|
| 674 |
},
|
| 675 |
{
|
| 676 |
"epoch": 2.6554216867469878,
|
| 677 |
+
"grad_norm": 0.5673111264101424,
|
| 678 |
+
"learning_rate": 5.376907622450229e-06,
|
| 679 |
+
"loss": 0.1154,
|
| 680 |
+
"mean_token_accuracy": 0.9624109007418156,
|
| 681 |
"num_tokens": 8964375.0,
|
| 682 |
"step": 136
|
| 683 |
},
|
| 684 |
{
|
| 685 |
"epoch": 2.693975903614458,
|
| 686 |
+
"grad_norm": 0.5636466902202368,
|
| 687 |
+
"learning_rate": 5.2399858019140005e-06,
|
| 688 |
+
"loss": 0.1045,
|
| 689 |
+
"mean_token_accuracy": 0.9666311480104923,
|
| 690 |
"num_tokens": 9094906.0,
|
| 691 |
"step": 138
|
| 692 |
},
|
| 693 |
{
|
| 694 |
"epoch": 2.7325301204819277,
|
| 695 |
+
"grad_norm": 0.5754464602822424,
|
| 696 |
+
"learning_rate": 5.102883327588608e-06,
|
| 697 |
+
"loss": 0.1075,
|
| 698 |
+
"mean_token_accuracy": 0.9647044539451599,
|
| 699 |
"num_tokens": 9225978.0,
|
| 700 |
"step": 140
|
| 701 |
},
|
| 702 |
{
|
| 703 |
"epoch": 2.7325301204819277,
|
| 704 |
+
"eval_loss": 0.37826669216156006,
|
| 705 |
+
"eval_mean_token_accuracy": 0.8995784972315637,
|
| 706 |
"eval_num_tokens": 9225978.0,
|
| 707 |
+
"eval_runtime": 69.6803,
|
| 708 |
+
"eval_samples_per_second": 12.27,
|
| 709 |
+
"eval_steps_per_second": 1.536,
|
| 710 |
"step": 140
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 2.7710843373493974,
|
| 714 |
+
"grad_norm": 0.5987257906687522,
|
| 715 |
+
"learning_rate": 4.965703405919154e-06,
|
| 716 |
+
"loss": 0.1041,
|
| 717 |
+
"mean_token_accuracy": 0.9660173505544662,
|
| 718 |
"num_tokens": 9357050.0,
|
| 719 |
"step": 142
|
| 720 |
},
|
| 721 |
{
|
| 722 |
"epoch": 2.8096385542168676,
|
| 723 |
+
"grad_norm": 0.6727909756019579,
|
| 724 |
+
"learning_rate": 4.828549301650673e-06,
|
| 725 |
+
"loss": 0.1165,
|
| 726 |
+
"mean_token_accuracy": 0.9626206122338772,
|
| 727 |
"num_tokens": 9488122.0,
|
| 728 |
"step": 144
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 2.8481927710843373,
|
| 732 |
+
"grad_norm": 0.5483728501054262,
|
| 733 |
+
"learning_rate": 4.691524260093672e-06,
|
| 734 |
+
"loss": 0.1101,
|
| 735 |
+
"mean_token_accuracy": 0.9640556387603283,
|
| 736 |
"num_tokens": 9619194.0,
|
| 737 |
"step": 146
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"epoch": 2.886746987951807,
|
| 741 |
+
"grad_norm": 0.6578615356471254,
|
| 742 |
+
"learning_rate": 4.554731429404293e-06,
|
| 743 |
+
"loss": 0.1167,
|
| 744 |
+
"mean_token_accuracy": 0.9623610861599445,
|
| 745 |
"num_tokens": 9750266.0,
|
| 746 |
"step": 148
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 2.9253012048192772,
|
| 750 |
+
"grad_norm": 0.544341897970942,
|
| 751 |
+
"learning_rate": 4.4182737829376135e-06,
|
| 752 |
+
"loss": 0.1068,
|
| 753 |
+
"mean_token_accuracy": 0.965429600328207,
|
| 754 |
"num_tokens": 9881338.0,
|
| 755 |
"step": 150
|
| 756 |
},
|
| 757 |
{
|
| 758 |
"epoch": 2.963855421686747,
|
| 759 |
+
"grad_norm": 0.5807218274090602,
|
| 760 |
+
"learning_rate": 4.28225404173254e-06,
|
| 761 |
+
"loss": 0.1058,
|
| 762 |
+
"mean_token_accuracy": 0.965176422148943,
|
| 763 |
"num_tokens": 10011972.0,
|
| 764 |
"step": 152
|
| 765 |
},
|
| 766 |
{
|
| 767 |
"epoch": 3.019277108433735,
|
| 768 |
+
"grad_norm": 1.007803950038667,
|
| 769 |
+
"learning_rate": 4.146774597186622e-06,
|
| 770 |
+
"loss": 0.1488,
|
| 771 |
+
"mean_token_accuracy": 0.9695591181516647,
|
| 772 |
"num_tokens": 10175812.0,
|
| 773 |
"step": 154
|
| 774 |
},
|
| 775 |
{
|
| 776 |
"epoch": 3.057831325301205,
|
| 777 |
+
"grad_norm": 0.6613641201206724,
|
| 778 |
+
"learning_rate": 4.011937433979014e-06,
|
| 779 |
+
"loss": 0.0847,
|
| 780 |
+
"mean_token_accuracy": 0.9746656753122807,
|
| 781 |
"num_tokens": 10306884.0,
|
| 782 |
"step": 156
|
| 783 |
},
|
| 784 |
{
|
| 785 |
"epoch": 3.0963855421686746,
|
| 786 |
+
"grad_norm": 0.5427167115705699,
|
| 787 |
+
"learning_rate": 3.87784405329962e-06,
|
| 788 |
+
"loss": 0.0838,
|
| 789 |
+
"mean_token_accuracy": 0.9741344675421715,
|
| 790 |
"num_tokens": 10437883.0,
|
| 791 |
"step": 158
|
| 792 |
},
|
| 793 |
{
|
| 794 |
"epoch": 3.1349397590361447,
|
| 795 |
+
"grad_norm": 0.5059704125761413,
|
| 796 |
+
"learning_rate": 3.744595396442169e-06,
|
| 797 |
+
"loss": 0.0814,
|
| 798 |
+
"mean_token_accuracy": 0.9750473313033581,
|
| 799 |
"num_tokens": 10568955.0,
|
| 800 |
"step": 160
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 3.1349397590361447,
|
| 804 |
+
"eval_loss": 0.4201391637325287,
|
| 805 |
+
"eval_mean_token_accuracy": 0.8986482670374005,
|
| 806 |
"eval_num_tokens": 10568955.0,
|
| 807 |
+
"eval_runtime": 69.8903,
|
| 808 |
+
"eval_samples_per_second": 12.233,
|
| 809 |
+
"eval_steps_per_second": 1.531,
|
| 810 |
"step": 160
|
| 811 |
},
|
| 812 |
{
|
| 813 |
"epoch": 3.1734939759036145,
|
| 814 |
+
"grad_norm": 0.4955524619584041,
|
| 815 |
+
"learning_rate": 3.612291768818772e-06,
|
| 816 |
+
"loss": 0.0827,
|
| 817 |
+
"mean_token_accuracy": 0.9744977466762066,
|
| 818 |
"num_tokens": 10700027.0,
|
| 819 |
"step": 162
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"epoch": 3.212048192771084,
|
| 823 |
+
"grad_norm": 0.5481909266796648,
|
| 824 |
+
"learning_rate": 3.4810327644531606e-06,
|
| 825 |
+
"loss": 0.0804,
|
| 826 |
+
"mean_token_accuracy": 0.9746122434735298,
|
| 827 |
"num_tokens": 10831099.0,
|
| 828 |
"step": 164
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"epoch": 3.2506024096385544,
|
| 832 |
+
"grad_norm": 0.5869274418415635,
|
| 833 |
+
"learning_rate": 3.3509171910094162e-06,
|
| 834 |
+
"loss": 0.0849,
|
| 835 |
+
"mean_token_accuracy": 0.9735665060579777,
|
| 836 |
"num_tokens": 10962171.0,
|
| 837 |
"step": 166
|
| 838 |
},
|
| 839 |
{
|
| 840 |
"epoch": 3.289156626506024,
|
| 841 |
+
"grad_norm": 0.5997938570160334,
|
| 842 |
+
"learning_rate": 3.222042995412669e-06,
|
| 843 |
+
"loss": 0.0826,
|
| 844 |
+
"mean_token_accuracy": 0.9744274839758873,
|
| 845 |
"num_tokens": 11092367.0,
|
| 846 |
"step": 168
|
| 847 |
},
|
| 848 |
{
|
| 849 |
"epoch": 3.327710843373494,
|
| 850 |
+
"grad_norm": 0.5638967234440626,
|
| 851 |
+
"learning_rate": 3.094507190117715e-06,
|
| 852 |
+
"loss": 0.0752,
|
| 853 |
+
"mean_token_accuracy": 0.9760014712810516,
|
| 854 |
"num_tokens": 11223439.0,
|
| 855 |
"step": 170
|
| 856 |
},
|
| 857 |
{
|
| 858 |
"epoch": 3.3662650602409636,
|
| 859 |
+
"grad_norm": 0.5677450107311146,
|
| 860 |
+
"learning_rate": 2.9684057800810844e-06,
|
| 861 |
+
"loss": 0.0849,
|
| 862 |
+
"mean_token_accuracy": 0.9734520092606544,
|
| 863 |
"num_tokens": 11354511.0,
|
| 864 |
"step": 172
|
| 865 |
},
|
| 866 |
{
|
| 867 |
"epoch": 3.404819277108434,
|
| 868 |
+
"grad_norm": 0.5694190125459168,
|
| 869 |
+
"learning_rate": 2.8438336904915186e-06,
|
| 870 |
+
"loss": 0.0907,
|
| 871 |
+
"mean_token_accuracy": 0.9719940833747387,
|
| 872 |
"num_tokens": 11485583.0,
|
| 873 |
"step": 174
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 3.4433734939759035,
|
| 877 |
+
"grad_norm": 0.5008764813796651,
|
| 878 |
+
"learning_rate": 2.7208846953132685e-06,
|
| 879 |
+
"loss": 0.0782,
|
| 880 |
+
"mean_token_accuracy": 0.9755356945097446,
|
| 881 |
"num_tokens": 11616217.0,
|
| 882 |
"step": 176
|
| 883 |
},
|
| 884 |
{
|
| 885 |
"epoch": 3.4819277108433733,
|
| 886 |
+
"grad_norm": 0.5027767263738213,
|
| 887 |
+
"learning_rate": 2.599651346695979e-06,
|
| 888 |
+
"loss": 0.0773,
|
| 889 |
+
"mean_token_accuracy": 0.9762609973549843,
|
| 890 |
"num_tokens": 11747289.0,
|
| 891 |
"step": 178
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 3.5204819277108435,
|
| 895 |
+
"grad_norm": 0.5747857741850161,
|
| 896 |
+
"learning_rate": 2.4802249053043525e-06,
|
| 897 |
"loss": 0.0777,
|
| 898 |
+
"mean_token_accuracy": 0.976215198636055,
|
| 899 |
"num_tokens": 11878361.0,
|
| 900 |
"step": 180
|
| 901 |
},
|
| 902 |
{
|
| 903 |
"epoch": 3.5204819277108435,
|
| 904 |
+
"eval_loss": 0.43149346113204956,
|
| 905 |
+
"eval_mean_token_accuracy": 0.898219308563482,
|
| 906 |
"eval_num_tokens": 11878361.0,
|
| 907 |
+
"eval_runtime": 69.6743,
|
| 908 |
+
"eval_samples_per_second": 12.271,
|
| 909 |
+
"eval_steps_per_second": 1.536,
|
| 910 |
"step": 180
|
| 911 |
},
|
| 912 |
{
|
| 913 |
"epoch": 3.559036144578313,
|
| 914 |
+
"grad_norm": 0.5115273312879999,
|
| 915 |
+
"learning_rate": 2.3626952716199647e-06,
|
| 916 |
+
"loss": 0.0792,
|
| 917 |
+
"mean_token_accuracy": 0.9750167988240719,
|
| 918 |
"num_tokens": 12009433.0,
|
| 919 |
"step": 182
|
| 920 |
},
|
| 921 |
{
|
| 922 |
"epoch": 3.597590361445783,
|
| 923 |
+
"grad_norm": 0.5172911491980401,
|
| 924 |
+
"learning_rate": 2.247150918267008e-06,
|
| 925 |
+
"loss": 0.0851,
|
| 926 |
+
"mean_token_accuracy": 0.9730398207902908,
|
| 927 |
"num_tokens": 12140505.0,
|
| 928 |
"step": 184
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 3.636144578313253,
|
| 932 |
+
"grad_norm": 0.5260093719963543,
|
| 933 |
+
"learning_rate": 2.133678823412873e-06,
|
| 934 |
+
"loss": 0.0797,
|
| 935 |
+
"mean_token_accuracy": 0.9751236625015736,
|
| 936 |
"num_tokens": 12271577.0,
|
| 937 |
"step": 186
|
| 938 |
},
|
| 939 |
{
|
| 940 |
"epoch": 3.674698795180723,
|
| 941 |
+
"grad_norm": 0.5267292864138245,
|
| 942 |
+
"learning_rate": 2.022364405293703e-06,
|
| 943 |
+
"loss": 0.0832,
|
| 944 |
+
"mean_token_accuracy": 0.9738947302103043,
|
| 945 |
"num_tokens": 12402649.0,
|
| 946 |
"step": 188
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 3.7132530120481926,
|
| 950 |
+
"grad_norm": 0.5065512725199254,
|
| 951 |
+
"learning_rate": 1.913291457914234e-06,
|
| 952 |
+
"loss": 0.0856,
|
| 953 |
+
"mean_token_accuracy": 0.9732001163065434,
|
| 954 |
"num_tokens": 12533721.0,
|
| 955 |
"step": 190
|
| 956 |
},
|
| 957 |
{
|
| 958 |
"epoch": 3.7518072289156628,
|
| 959 |
+
"grad_norm": 0.5465242770321679,
|
| 960 |
+
"learning_rate": 1.8065420879702888e-06,
|
| 961 |
+
"loss": 0.0838,
|
| 962 |
+
"mean_token_accuracy": 0.9731762520968914,
|
| 963 |
"num_tokens": 12663435.0,
|
| 964 |
"step": 192
|
| 965 |
},
|
| 966 |
{
|
| 967 |
"epoch": 3.7903614457831325,
|
| 968 |
+
"grad_norm": 0.7823063875533764,
|
| 969 |
+
"learning_rate": 1.7021966530414303e-06,
|
| 970 |
+
"loss": 0.0762,
|
| 971 |
+
"mean_token_accuracy": 0.9758411757647991,
|
| 972 |
"num_tokens": 12794507.0,
|
| 973 |
"step": 194
|
| 974 |
},
|
| 975 |
{
|
| 976 |
"epoch": 3.8289156626506022,
|
| 977 |
+
"grad_norm": 0.571380544699335,
|
| 978 |
+
"learning_rate": 1.6003337011002928e-06,
|
| 979 |
+
"loss": 0.084,
|
| 980 |
+
"mean_token_accuracy": 0.9734901748597622,
|
| 981 |
"num_tokens": 12925579.0,
|
| 982 |
"step": 196
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 3.8674698795180724,
|
| 986 |
+
"grad_norm": 0.5400258981871386,
|
| 987 |
+
"learning_rate": 1.5010299113841397e-06,
|
| 988 |
+
"loss": 0.0807,
|
| 989 |
+
"mean_token_accuracy": 0.9752305261790752,
|
| 990 |
"num_tokens": 13056651.0,
|
| 991 |
"step": 198
|
| 992 |
},
|
| 993 |
{
|
| 994 |
"epoch": 3.906024096385542,
|
| 995 |
+
"grad_norm": 0.5204832843446408,
|
| 996 |
+
"learning_rate": 1.4043600366731213e-06,
|
| 997 |
+
"loss": 0.0821,
|
| 998 |
+
"mean_token_accuracy": 0.9745206460356712,
|
| 999 |
"num_tokens": 13187723.0,
|
| 1000 |
"step": 200
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 3.906024096385542,
|
| 1004 |
+
"eval_loss": 0.43459072709083557,
|
| 1005 |
+
"eval_mean_token_accuracy": 0.8980461002510285,
|
| 1006 |
"eval_num_tokens": 13187723.0,
|
| 1007 |
+
"eval_runtime": 69.6812,
|
| 1008 |
+
"eval_samples_per_second": 12.27,
|
| 1009 |
+
"eval_steps_per_second": 1.536,
|
| 1010 |
"step": 200
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 3.944578313253012,
|
| 1014 |
+
"grad_norm": 0.5732935867678565,
|
| 1015 |
+
"learning_rate": 1.3103968470187384e-06,
|
| 1016 |
+
"loss": 0.0841,
|
| 1017 |
+
"mean_token_accuracy": 0.973306454718113,
|
| 1018 |
"num_tokens": 13317686.0,
|
| 1019 |
"step": 202
|
| 1020 |
},
|
| 1021 |
{
|
| 1022 |
"epoch": 3.983132530120482,
|
| 1023 |
+
"grad_norm": 0.5049593156468802,
|
| 1024 |
+
"learning_rate": 1.2192110749648233e-06,
|
| 1025 |
+
"loss": 0.0783,
|
| 1026 |
+
"mean_token_accuracy": 0.9752342775464058,
|
| 1027 |
"num_tokens": 13447600.0,
|
| 1028 |
"step": 204
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
"epoch": 4.03855421686747,
|
| 1032 |
+
"grad_norm": 0.4900616503984239,
|
| 1033 |
+
"learning_rate": 1.1308713623022988e-06,
|
| 1034 |
+
"loss": 0.1075,
|
| 1035 |
+
"mean_token_accuracy": 0.9786272644996643,
|
| 1036 |
"num_tokens": 13611440.0,
|
| 1037 |
"step": 206
|
| 1038 |
},
|
| 1039 |
{
|
| 1040 |
"epoch": 4.0771084337349395,
|
| 1041 |
+
"grad_norm": 0.4917129834327916,
|
| 1042 |
+
"learning_rate": 1.045444208397791e-06,
|
| 1043 |
+
"loss": 0.0676,
|
| 1044 |
+
"mean_token_accuracy": 0.9801687188446522,
|
| 1045 |
"num_tokens": 13740537.0,
|
| 1046 |
"step": 208
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
"epoch": 4.11566265060241,
|
| 1050 |
+
"grad_norm": 0.47200516762524886,
|
| 1051 |
+
"learning_rate": 9.629939201349852e-07,
|
| 1052 |
+
"loss": 0.0723,
|
| 1053 |
+
"mean_token_accuracy": 0.9782837741076946,
|
| 1054 |
"num_tokens": 13871609.0,
|
| 1055 |
"step": 210
|
| 1056 |
},
|
| 1057 |
{
|
| 1058 |
"epoch": 4.15421686746988,
|
| 1059 |
+
"grad_norm": 0.44277012092487705,
|
| 1060 |
+
"learning_rate": 8.835825635064266e-07,
|
| 1061 |
+
"loss": 0.0729,
|
| 1062 |
+
"mean_token_accuracy": 0.9780853129923344,
|
| 1063 |
"num_tokens": 14002681.0,
|
| 1064 |
"step": 212
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
"epoch": 4.192771084337349,
|
| 1068 |
+
"grad_norm": 0.4753962832603972,
|
| 1069 |
+
"learning_rate": 8.072699168921827e-07,
|
| 1070 |
+
"loss": 0.0749,
|
| 1071 |
+
"mean_token_accuracy": 0.9778944849967957,
|
| 1072 |
"num_tokens": 14133753.0,
|
| 1073 |
"step": 214
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
"epoch": 4.231325301204819,
|
| 1077 |
+
"grad_norm": 0.48346978347475456,
|
| 1078 |
+
"learning_rate": 7.341134260605537e-07,
|
| 1079 |
+
"loss": 0.0692,
|
| 1080 |
+
"mean_token_accuracy": 0.9793745614588261,
|
| 1081 |
"num_tokens": 14264314.0,
|
| 1082 |
"step": 216
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
"epoch": 4.2698795180722895,
|
| 1086 |
+
"grad_norm": 0.4328206037632282,
|
| 1087 |
+
"learning_rate": 6.641681609246981e-07,
|
| 1088 |
+
"loss": 0.066,
|
| 1089 |
+
"mean_token_accuracy": 0.9801309891045094,
|
| 1090 |
"num_tokens": 14395386.0,
|
| 1091 |
"step": 218
|
| 1092 |
},
|
| 1093 |
{
|
| 1094 |
"epoch": 4.308433734939759,
|
| 1095 |
+
"grad_norm": 0.46221534542018206,
|
| 1096 |
+
"learning_rate": 5.974867740877282e-07,
|
| 1097 |
+
"loss": 0.0696,
|
| 1098 |
+
"mean_token_accuracy": 0.9789478555321693,
|
| 1099 |
"num_tokens": 14526458.0,
|
| 1100 |
"step": 220
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
"epoch": 4.308433734939759,
|
| 1104 |
+
"eval_loss": 0.4595886468887329,
|
| 1105 |
+
"eval_mean_token_accuracy": 0.897223442514366,
|
| 1106 |
"eval_num_tokens": 14526458.0,
|
| 1107 |
+
"eval_runtime": 69.6441,
|
| 1108 |
+
"eval_samples_per_second": 12.277,
|
| 1109 |
+
"eval_steps_per_second": 1.536,
|
| 1110 |
"step": 220
|
| 1111 |
},
|
| 1112 |
{
|
| 1113 |
"epoch": 4.346987951807229,
|
| 1114 |
+
"grad_norm": 0.4739286679144528,
|
| 1115 |
+
"learning_rate": 5.341194612074824e-07,
|
| 1116 |
+
"loss": 0.068,
|
| 1117 |
+
"mean_token_accuracy": 0.9796868488192558,
|
| 1118 |
"num_tokens": 14656421.0,
|
| 1119 |
"step": 222
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
"epoch": 4.385542168674699,
|
| 1123 |
+
"grad_norm": 0.43096986690967987,
|
| 1124 |
+
"learning_rate": 4.7411392321080606e-07,
|
| 1125 |
+
"loss": 0.0663,
|
| 1126 |
+
"mean_token_accuracy": 0.9802683852612972,
|
| 1127 |
"num_tokens": 14787493.0,
|
| 1128 |
"step": 224
|
| 1129 |
},
|
| 1130 |
{
|
| 1131 |
"epoch": 4.424096385542168,
|
| 1132 |
+
"grad_norm": 0.46557922408208563,
|
| 1133 |
+
"learning_rate": 4.175153303857887e-07,
|
| 1134 |
+
"loss": 0.0654,
|
| 1135 |
+
"mean_token_accuracy": 0.9804821126163006,
|
| 1136 |
"num_tokens": 14918565.0,
|
| 1137 |
"step": 226
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
"epoch": 4.462650602409639,
|
| 1141 |
+
"grad_norm": 0.5546707256189516,
|
| 1142 |
+
"learning_rate": 3.643662883789878e-07,
|
| 1143 |
+
"loss": 0.0673,
|
| 1144 |
+
"mean_token_accuracy": 0.979527972638607,
|
| 1145 |
"num_tokens": 15049637.0,
|
| 1146 |
"step": 228
|
| 1147 |
},
|
| 1148 |
{
|
| 1149 |
"epoch": 4.501204819277109,
|
| 1150 |
+
"grad_norm": 0.49021519394663,
|
| 1151 |
+
"learning_rate": 3.1470680612323503e-07,
|
| 1152 |
+
"loss": 0.07,
|
| 1153 |
+
"mean_token_accuracy": 0.9785585664212704,
|
| 1154 |
"num_tokens": 15180709.0,
|
| 1155 |
"step": 230
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
"epoch": 4.539759036144578,
|
| 1159 |
+
"grad_norm": 0.45571708386475684,
|
| 1160 |
+
"learning_rate": 2.685742657201601e-07,
|
| 1161 |
+
"loss": 0.0697,
|
| 1162 |
+
"mean_token_accuracy": 0.9785204008221626,
|
| 1163 |
"num_tokens": 15311781.0,
|
| 1164 |
"step": 232
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 4.578313253012048,
|
| 1168 |
+
"grad_norm": 0.5641008416839415,
|
| 1169 |
+
"learning_rate": 2.260033943001244e-07,
|
| 1170 |
+
"loss": 0.0663,
|
| 1171 |
+
"mean_token_accuracy": 0.9797416999936104,
|
| 1172 |
"num_tokens": 15442853.0,
|
| 1173 |
"step": 234
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
"epoch": 4.6168674698795185,
|
| 1177 |
+
"grad_norm": 0.5607141029792978,
|
| 1178 |
+
"learning_rate": 1.8702623788072028e-07,
|
| 1179 |
+
"loss": 0.0793,
|
| 1180 |
+
"mean_token_accuracy": 0.9755663834512234,
|
| 1181 |
"num_tokens": 15573925.0,
|
| 1182 |
"step": 236
|
| 1183 |
},
|
| 1184 |
{
|
| 1185 |
"epoch": 4.655421686746988,
|
| 1186 |
+
"grad_norm": 0.46095439859311127,
|
| 1187 |
+
"learning_rate": 1.5167213724353426e-07,
|
| 1188 |
+
"loss": 0.0714,
|
| 1189 |
+
"mean_token_accuracy": 0.9779479168355465,
|
| 1190 |
"num_tokens": 15704997.0,
|
| 1191 |
"step": 238
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
"epoch": 4.693975903614458,
|
| 1195 |
+
"grad_norm": 0.464368810663561,
|
| 1196 |
+
"learning_rate": 1.199677058473292e-07,
|
| 1197 |
+
"loss": 0.066,
|
| 1198 |
+
"mean_token_accuracy": 0.980153888463974,
|
| 1199 |
"num_tokens": 15836069.0,
|
| 1200 |
"step": 240
|
| 1201 |
},
|
| 1202 |
{
|
| 1203 |
"epoch": 4.693975903614458,
|
| 1204 |
+
"eval_loss": 0.46903374791145325,
|
| 1205 |
+
"eval_mean_token_accuracy": 0.8968599628065234,
|
| 1206 |
"eval_num_tokens": 15836069.0,
|
| 1207 |
+
"eval_runtime": 69.6558,
|
| 1208 |
+
"eval_samples_per_second": 12.275,
|
| 1209 |
+
"eval_steps_per_second": 1.536,
|
| 1210 |
"step": 240
|
| 1211 |
},
|
| 1212 |
{
|
| 1213 |
"epoch": 4.732530120481927,
|
| 1214 |
+
"grad_norm": 0.5162077757262011,
|
| 1215 |
+
"learning_rate": 9.193680979426189e-08,
|
| 1216 |
+
"loss": 0.0775,
|
| 1217 |
+
"mean_token_accuracy": 0.9764594584703445,
|
| 1218 |
"num_tokens": 15967141.0,
|
| 1219 |
"step": 242
|
| 1220 |
},
|
| 1221 |
{
|
| 1222 |
"epoch": 4.771084337349397,
|
| 1223 |
+
"grad_norm": 0.4482450270539155,
|
| 1224 |
+
"learning_rate": 6.760054986423459e-08,
|
| 1225 |
+
"loss": 0.0632,
|
| 1226 |
+
"mean_token_accuracy": 0.9808179698884487,
|
| 1227 |
"num_tokens": 16098213.0,
|
| 1228 |
"step": 244
|
| 1229 |
},
|
| 1230 |
{
|
| 1231 |
"epoch": 4.809638554216868,
|
| 1232 |
+
"grad_norm": 0.4698597407866022,
|
| 1233 |
+
"learning_rate": 4.697724563088646e-08,
|
| 1234 |
+
"loss": 0.0681,
|
| 1235 |
+
"mean_token_accuracy": 0.9797111675143242,
|
| 1236 |
"num_tokens": 16229285.0,
|
| 1237 |
"step": 246
|
| 1238 |
},
|
| 1239 |
{
|
| 1240 |
"epoch": 4.848192771084337,
|
| 1241 |
+
"grad_norm": 0.4662674319978425,
|
| 1242 |
+
"learning_rate": 3.0082421671192576e-08,
|
| 1243 |
+
"loss": 0.0688,
|
| 1244 |
+
"mean_token_accuracy": 0.97944400832057,
|
| 1245 |
"num_tokens": 16360357.0,
|
| 1246 |
"step": 248
|
| 1247 |
},
|
| 1248 |
{
|
| 1249 |
"epoch": 4.886746987951807,
|
| 1250 |
+
"grad_norm": 0.46327536754981147,
|
| 1251 |
+
"learning_rate": 1.692879587904983e-08,
|
| 1252 |
+
"loss": 0.0662,
|
| 1253 |
+
"mean_token_accuracy": 0.9799401611089706,
|
| 1254 |
"num_tokens": 16491429.0,
|
| 1255 |
"step": 250
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 4.925301204819277,
|
| 1259 |
+
"grad_norm": 0.4688691090714117,
|
| 1260 |
+
"learning_rate": 7.526269891646176e-09,
|
| 1261 |
+
"loss": 0.0642,
|
| 1262 |
+
"mean_token_accuracy": 0.9807046689093113,
|
| 1263 |
"num_tokens": 16621960.0,
|
| 1264 |
"step": 252
|
| 1265 |
},
|
| 1266 |
{
|
| 1267 |
"epoch": 4.9638554216867465,
|
| 1268 |
+
"grad_norm": 0.4516057398304381,
|
| 1269 |
+
"learning_rate": 1.8819216358156865e-09,
|
| 1270 |
+
"loss": 0.0688,
|
| 1271 |
+
"mean_token_accuracy": 0.9792744368314743,
|
| 1272 |
"num_tokens": 16752156.0,
|
| 1273 |
"step": 254
|
| 1274 |
},
|
| 1275 |
{
|
| 1276 |
"epoch": 4.983132530120482,
|
| 1277 |
+
"mean_token_accuracy": 0.976367861032486,
|
| 1278 |
"num_tokens": 16817692.0,
|
| 1279 |
"step": 255,
|
| 1280 |
"total_flos": 24409842647040.0,
|
| 1281 |
+
"train_loss": 0.12274208276295194,
|
| 1282 |
+
"train_runtime": 3782.9235,
|
| 1283 |
"train_samples_per_second": 2.194,
|
| 1284 |
"train_steps_per_second": 0.067
|
| 1285 |
}
|