Model save
Browse files- README.md +1 -1
- all_results.json +6 -6
- train_results.json +6 -6
- trainer_state.json +486 -262
README.md
CHANGED
|
@@ -27,7 +27,7 @@ print(output["generated_text"])
|
|
| 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/yeokch/stream-of-search-train/runs/
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with SFT.
|
|
|
|
| 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/yeokch/stream-of-search-train/runs/avxmo5b9)
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with SFT.
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"total_flos":
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"total_flos": 1.112161775475753e+17,
|
| 3 |
+
"train_loss": 0.24216800009339584,
|
| 4 |
+
"train_runtime": 6297.1418,
|
| 5 |
+
"train_samples": 10000,
|
| 6 |
+
"train_samples_per_second": 0.983,
|
| 7 |
+
"train_steps_per_second": 0.061
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"total_flos":
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 0.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"total_flos": 1.112161775475753e+17,
|
| 3 |
+
"train_loss": 0.24216800009339584,
|
| 4 |
+
"train_runtime": 6297.1418,
|
| 5 |
+
"train_samples": 10000,
|
| 6 |
+
"train_samples_per_second": 0.983,
|
| 7 |
+
"train_steps_per_second": 0.061
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -2,426 +2,650 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
-
"epoch": 0.
|
| 14 |
-
"grad_norm": 0.
|
| 15 |
-
"learning_rate":
|
| 16 |
-
"loss":
|
| 17 |
-
"mean_token_accuracy": 0.
|
| 18 |
"step": 1
|
| 19 |
},
|
| 20 |
{
|
| 21 |
-
"epoch": 0.
|
| 22 |
-
"grad_norm": 0.
|
| 23 |
-
"learning_rate":
|
| 24 |
-
"loss":
|
| 25 |
-
"mean_token_accuracy": 0.
|
| 26 |
"step": 5
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch": 0.
|
| 30 |
-
"grad_norm": 0.
|
| 31 |
-
"learning_rate":
|
| 32 |
-
"loss":
|
| 33 |
-
"mean_token_accuracy": 0.
|
| 34 |
"step": 10
|
| 35 |
},
|
| 36 |
{
|
| 37 |
-
"epoch": 0.
|
| 38 |
-
"grad_norm": 0.
|
| 39 |
-
"learning_rate":
|
| 40 |
-
"loss": 0.
|
| 41 |
-
"mean_token_accuracy": 0.
|
| 42 |
"step": 15
|
| 43 |
},
|
| 44 |
{
|
| 45 |
-
"epoch": 0.
|
| 46 |
-
"grad_norm": 0.
|
| 47 |
-
"learning_rate": 0.
|
| 48 |
-
"loss": 0.
|
| 49 |
-
"mean_token_accuracy": 0.
|
| 50 |
"step": 20
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"grad_norm": 0.
|
| 55 |
-
"learning_rate": 0.
|
| 56 |
-
"loss": 0.
|
| 57 |
-
"mean_token_accuracy": 0.
|
| 58 |
"step": 25
|
| 59 |
},
|
| 60 |
{
|
| 61 |
-
"epoch": 0.
|
| 62 |
-
"grad_norm": 0.
|
| 63 |
-
"learning_rate": 0.
|
| 64 |
-
"loss": 0.
|
| 65 |
-
"mean_token_accuracy": 0.
|
| 66 |
"step": 30
|
| 67 |
},
|
| 68 |
{
|
| 69 |
-
"epoch": 0.
|
| 70 |
-
"grad_norm": 0.
|
| 71 |
-
"learning_rate": 0.
|
| 72 |
-
"loss": 0.
|
| 73 |
-
"mean_token_accuracy": 0.
|
| 74 |
"step": 35
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 0.
|
| 78 |
-
"grad_norm": 0.
|
| 79 |
-
"learning_rate": 0.
|
| 80 |
-
"loss": 0.
|
| 81 |
-
"mean_token_accuracy": 0.
|
| 82 |
"step": 40
|
| 83 |
},
|
| 84 |
{
|
| 85 |
-
"epoch": 0.
|
| 86 |
-
"grad_norm": 0.
|
| 87 |
-
"learning_rate": 0.
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"mean_token_accuracy": 0.
|
| 90 |
"step": 45
|
| 91 |
},
|
| 92 |
{
|
| 93 |
-
"epoch": 0.
|
| 94 |
-
"grad_norm": 0.
|
| 95 |
-
"learning_rate": 0.
|
| 96 |
-
"loss": 0.
|
| 97 |
-
"mean_token_accuracy": 0.
|
| 98 |
"step": 50
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"grad_norm": 0.
|
| 103 |
-
"learning_rate": 0.
|
| 104 |
-
"loss": 0.
|
| 105 |
-
"mean_token_accuracy": 0.
|
| 106 |
"step": 55
|
| 107 |
},
|
| 108 |
{
|
| 109 |
-
"epoch": 0.
|
| 110 |
-
"grad_norm": 0.
|
| 111 |
-
"learning_rate": 0.
|
| 112 |
-
"loss": 0.
|
| 113 |
-
"mean_token_accuracy": 0.
|
| 114 |
"step": 60
|
| 115 |
},
|
| 116 |
{
|
| 117 |
-
"epoch": 0.
|
| 118 |
-
"grad_norm": 0.
|
| 119 |
-
"learning_rate": 0.
|
| 120 |
-
"loss": 0.
|
| 121 |
-
"mean_token_accuracy": 0.
|
| 122 |
"step": 65
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 0.
|
| 126 |
-
"grad_norm": 0.
|
| 127 |
-
"learning_rate": 0.
|
| 128 |
-
"loss": 0.
|
| 129 |
-
"mean_token_accuracy": 0.
|
| 130 |
"step": 70
|
| 131 |
},
|
| 132 |
{
|
| 133 |
-
"epoch": 0.
|
| 134 |
-
"grad_norm": 0.
|
| 135 |
-
"learning_rate": 0.
|
| 136 |
-
"loss": 0.
|
| 137 |
-
"mean_token_accuracy": 0.
|
| 138 |
"step": 75
|
| 139 |
},
|
| 140 |
{
|
| 141 |
-
"epoch": 0.
|
| 142 |
-
"grad_norm": 0.
|
| 143 |
-
"learning_rate": 0.
|
| 144 |
-
"loss": 0.
|
| 145 |
-
"mean_token_accuracy": 0.
|
| 146 |
"step": 80
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 0.
|
| 150 |
-
"grad_norm": 0.
|
| 151 |
-
"learning_rate": 0.
|
| 152 |
-
"loss": 0.
|
| 153 |
-
"mean_token_accuracy": 0.
|
| 154 |
"step": 85
|
| 155 |
},
|
| 156 |
{
|
| 157 |
-
"epoch": 0.
|
| 158 |
-
"grad_norm": 0.
|
| 159 |
-
"learning_rate": 0.
|
| 160 |
-
"loss": 0.
|
| 161 |
-
"mean_token_accuracy": 0.
|
| 162 |
"step": 90
|
| 163 |
},
|
| 164 |
{
|
| 165 |
-
"epoch": 0.
|
| 166 |
-
"grad_norm": 0.
|
| 167 |
-
"learning_rate": 0.
|
| 168 |
-
"loss": 0.
|
| 169 |
-
"mean_token_accuracy": 0.
|
| 170 |
"step": 95
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch": 0.
|
| 174 |
-
"grad_norm": 0.
|
| 175 |
-
"learning_rate": 0.
|
| 176 |
-
"loss": 0.
|
| 177 |
-
"mean_token_accuracy": 0.
|
| 178 |
"step": 100
|
| 179 |
},
|
| 180 |
{
|
| 181 |
-
"epoch": 0.
|
| 182 |
-
"grad_norm": 0.
|
| 183 |
-
"learning_rate": 0.
|
| 184 |
-
"loss": 0.
|
| 185 |
-
"mean_token_accuracy": 0.
|
| 186 |
"step": 105
|
| 187 |
},
|
| 188 |
{
|
| 189 |
-
"epoch": 0.
|
| 190 |
-
"grad_norm": 0.
|
| 191 |
-
"learning_rate": 0.
|
| 192 |
-
"loss": 0.
|
| 193 |
-
"mean_token_accuracy": 0.
|
| 194 |
"step": 110
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch": 0.
|
| 198 |
-
"grad_norm": 0.
|
| 199 |
-
"learning_rate": 0.
|
| 200 |
-
"loss": 0.
|
| 201 |
-
"mean_token_accuracy": 0.
|
| 202 |
"step": 115
|
| 203 |
},
|
| 204 |
{
|
| 205 |
-
"epoch": 0.
|
| 206 |
-
"grad_norm": 0.
|
| 207 |
-
"learning_rate": 0.
|
| 208 |
-
"loss": 0.
|
| 209 |
-
"mean_token_accuracy": 0.
|
| 210 |
"step": 120
|
| 211 |
},
|
| 212 |
{
|
| 213 |
-
"epoch": 0.
|
| 214 |
-
"grad_norm": 0.
|
| 215 |
-
"learning_rate": 0.
|
| 216 |
-
"loss": 0.
|
| 217 |
-
"mean_token_accuracy": 0.
|
| 218 |
"step": 125
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch": 0.
|
| 222 |
-
"grad_norm": 0.
|
| 223 |
-
"learning_rate": 0.
|
| 224 |
-
"loss": 0.
|
| 225 |
-
"mean_token_accuracy": 0.
|
| 226 |
"step": 130
|
| 227 |
},
|
| 228 |
{
|
| 229 |
-
"epoch": 0.
|
| 230 |
-
"grad_norm": 0.
|
| 231 |
-
"learning_rate": 0.
|
| 232 |
-
"loss": 0.
|
| 233 |
-
"mean_token_accuracy": 0.
|
| 234 |
"step": 135
|
| 235 |
},
|
| 236 |
{
|
| 237 |
-
"epoch": 0.
|
| 238 |
-
"grad_norm": 0.
|
| 239 |
-
"learning_rate":
|
| 240 |
-
"loss": 0.
|
| 241 |
-
"mean_token_accuracy": 0.
|
| 242 |
"step": 140
|
| 243 |
},
|
| 244 |
{
|
| 245 |
-
"epoch": 0.
|
| 246 |
-
"grad_norm": 0.
|
| 247 |
-
"learning_rate":
|
| 248 |
-
"loss": 0.
|
| 249 |
-
"mean_token_accuracy": 0.
|
| 250 |
"step": 145
|
| 251 |
},
|
| 252 |
{
|
| 253 |
-
"epoch": 0.
|
| 254 |
-
"grad_norm": 0.
|
| 255 |
-
"learning_rate":
|
| 256 |
-
"loss": 0.
|
| 257 |
-
"mean_token_accuracy": 0.
|
| 258 |
"step": 150
|
| 259 |
},
|
| 260 |
{
|
| 261 |
-
"epoch": 0.
|
| 262 |
-
"grad_norm": 0.
|
| 263 |
-
"learning_rate":
|
| 264 |
-
"loss": 0.
|
| 265 |
-
"mean_token_accuracy": 0.
|
| 266 |
"step": 155
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"grad_norm": 0.
|
| 271 |
-
"learning_rate":
|
| 272 |
-
"loss": 0.
|
| 273 |
-
"mean_token_accuracy": 0.
|
| 274 |
"step": 160
|
| 275 |
},
|
| 276 |
{
|
| 277 |
-
"epoch": 0.
|
| 278 |
-
"grad_norm": 0.
|
| 279 |
-
"learning_rate":
|
| 280 |
-
"loss": 0.
|
| 281 |
-
"mean_token_accuracy": 0.
|
| 282 |
"step": 165
|
| 283 |
},
|
| 284 |
{
|
| 285 |
-
"epoch": 0.
|
| 286 |
-
"grad_norm": 0.
|
| 287 |
-
"learning_rate":
|
| 288 |
-
"loss": 0.
|
| 289 |
-
"mean_token_accuracy": 0.
|
| 290 |
"step": 170
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch": 0.
|
| 294 |
-
"grad_norm": 0.
|
| 295 |
-
"learning_rate":
|
| 296 |
-
"loss": 0.
|
| 297 |
-
"mean_token_accuracy": 0.
|
| 298 |
"step": 175
|
| 299 |
},
|
| 300 |
{
|
| 301 |
-
"epoch": 0.
|
| 302 |
-
"grad_norm": 0.
|
| 303 |
-
"learning_rate":
|
| 304 |
-
"loss": 0.
|
| 305 |
-
"mean_token_accuracy": 0.
|
| 306 |
"step": 180
|
| 307 |
},
|
| 308 |
{
|
| 309 |
-
"epoch": 0.
|
| 310 |
-
"grad_norm": 0.
|
| 311 |
-
"learning_rate":
|
| 312 |
-
"loss": 0.
|
| 313 |
-
"mean_token_accuracy": 0.
|
| 314 |
"step": 185
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch": 0.
|
| 318 |
-
"grad_norm": 0.
|
| 319 |
-
"learning_rate":
|
| 320 |
-
"loss": 0.
|
| 321 |
-
"mean_token_accuracy": 0.
|
| 322 |
"step": 190
|
| 323 |
},
|
| 324 |
{
|
| 325 |
-
"epoch": 0.
|
| 326 |
-
"grad_norm": 0.
|
| 327 |
-
"learning_rate":
|
| 328 |
-
"loss": 0.
|
| 329 |
-
"mean_token_accuracy": 0.
|
| 330 |
"step": 195
|
| 331 |
},
|
| 332 |
{
|
| 333 |
-
"epoch": 0.
|
| 334 |
-
"grad_norm": 0.
|
| 335 |
-
"learning_rate":
|
| 336 |
-
"loss": 0.
|
| 337 |
-
"mean_token_accuracy": 0.
|
| 338 |
"step": 200
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch": 0.
|
| 342 |
-
"grad_norm": 0.
|
| 343 |
-
"learning_rate":
|
| 344 |
-
"loss": 0.
|
| 345 |
-
"mean_token_accuracy": 0.
|
| 346 |
"step": 205
|
| 347 |
},
|
| 348 |
{
|
| 349 |
-
"epoch": 0.
|
| 350 |
-
"grad_norm": 0.
|
| 351 |
-
"learning_rate":
|
| 352 |
-
"loss": 0.
|
| 353 |
-
"mean_token_accuracy": 0.
|
| 354 |
"step": 210
|
| 355 |
},
|
| 356 |
{
|
| 357 |
-
"epoch": 0.
|
| 358 |
-
"grad_norm": 0.
|
| 359 |
-
"learning_rate":
|
| 360 |
-
"loss": 0.
|
| 361 |
-
"mean_token_accuracy": 0.
|
| 362 |
"step": 215
|
| 363 |
},
|
| 364 |
{
|
| 365 |
-
"epoch": 0.
|
| 366 |
-
"grad_norm": 0.
|
| 367 |
-
"learning_rate":
|
| 368 |
-
"loss": 0.
|
| 369 |
-
"mean_token_accuracy": 0.
|
| 370 |
"step": 220
|
| 371 |
},
|
| 372 |
{
|
| 373 |
-
"epoch": 0.
|
| 374 |
-
"grad_norm": 0.
|
| 375 |
-
"learning_rate":
|
| 376 |
-
"loss": 0.
|
| 377 |
-
"mean_token_accuracy": 0.
|
| 378 |
"step": 225
|
| 379 |
},
|
| 380 |
{
|
| 381 |
-
"epoch": 0.
|
| 382 |
-
"grad_norm": 0.
|
| 383 |
-
"learning_rate":
|
| 384 |
-
"loss": 0.
|
| 385 |
-
"mean_token_accuracy": 0.
|
| 386 |
"step": 230
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch": 0.
|
| 390 |
-
"grad_norm": 0.
|
| 391 |
-
"learning_rate":
|
| 392 |
-
"loss": 0.
|
| 393 |
-
"mean_token_accuracy": 0.
|
| 394 |
"step": 235
|
| 395 |
},
|
| 396 |
{
|
| 397 |
-
"epoch": 0.
|
| 398 |
-
"grad_norm": 0.
|
| 399 |
-
"learning_rate":
|
| 400 |
-
"loss": 0.
|
| 401 |
-
"mean_token_accuracy": 0.
|
| 402 |
"step": 240
|
| 403 |
},
|
| 404 |
{
|
| 405 |
-
"epoch": 0.
|
| 406 |
-
"grad_norm": 0.
|
| 407 |
-
"learning_rate":
|
| 408 |
-
"loss": 0.
|
| 409 |
-
"mean_token_accuracy": 0.
|
| 410 |
"step": 245
|
| 411 |
},
|
| 412 |
{
|
| 413 |
-
"epoch": 0.
|
| 414 |
-
"
|
| 415 |
-
"
|
| 416 |
-
"
|
| 417 |
-
"
|
| 418 |
-
"
|
| 419 |
-
|
| 420 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 421 |
}
|
| 422 |
],
|
| 423 |
"logging_steps": 5,
|
| 424 |
-
"max_steps":
|
| 425 |
"num_input_tokens_seen": 0,
|
| 426 |
"num_train_epochs": 1,
|
| 427 |
"save_steps": 100,
|
|
@@ -437,7 +661,7 @@
|
|
| 437 |
"attributes": {}
|
| 438 |
}
|
| 439 |
},
|
| 440 |
-
"total_flos":
|
| 441 |
"train_batch_size": 1,
|
| 442 |
"trial_name": null,
|
| 443 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.9978994991113266,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 386,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
+
"epoch": 0.0025852318629827112,
|
| 14 |
+
"grad_norm": 0.19946137070655823,
|
| 15 |
+
"learning_rate": 5.128205128205128e-06,
|
| 16 |
+
"loss": 0.4855,
|
| 17 |
+
"mean_token_accuracy": 0.8692543655633926,
|
| 18 |
"step": 1
|
| 19 |
},
|
| 20 |
{
|
| 21 |
+
"epoch": 0.012926159314913557,
|
| 22 |
+
"grad_norm": 0.18048785626888275,
|
| 23 |
+
"learning_rate": 2.564102564102564e-05,
|
| 24 |
+
"loss": 0.5257,
|
| 25 |
+
"mean_token_accuracy": 0.8587683830410242,
|
| 26 |
"step": 5
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.025852318629827113,
|
| 30 |
+
"grad_norm": 0.18388661742210388,
|
| 31 |
+
"learning_rate": 5.128205128205128e-05,
|
| 32 |
+
"loss": 0.5986,
|
| 33 |
+
"mean_token_accuracy": 0.838620126247406,
|
| 34 |
"step": 10
|
| 35 |
},
|
| 36 |
{
|
| 37 |
+
"epoch": 0.038778477944740666,
|
| 38 |
+
"grad_norm": 0.13041158020496368,
|
| 39 |
+
"learning_rate": 7.692307692307693e-05,
|
| 40 |
+
"loss": 0.5464,
|
| 41 |
+
"mean_token_accuracy": 0.8496200017631054,
|
| 42 |
"step": 15
|
| 43 |
},
|
| 44 |
{
|
| 45 |
+
"epoch": 0.051704637259654226,
|
| 46 |
+
"grad_norm": 0.10899022966623306,
|
| 47 |
+
"learning_rate": 0.00010256410256410256,
|
| 48 |
+
"loss": 0.4321,
|
| 49 |
+
"mean_token_accuracy": 0.8813652157783508,
|
| 50 |
"step": 20
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 0.06463079657456779,
|
| 54 |
+
"grad_norm": 0.12952959537506104,
|
| 55 |
+
"learning_rate": 0.00012820512820512823,
|
| 56 |
+
"loss": 0.4429,
|
| 57 |
+
"mean_token_accuracy": 0.8760499283671379,
|
| 58 |
"step": 25
|
| 59 |
},
|
| 60 |
{
|
| 61 |
+
"epoch": 0.07755695588948133,
|
| 62 |
+
"grad_norm": 0.11217343807220459,
|
| 63 |
+
"learning_rate": 0.00015384615384615385,
|
| 64 |
+
"loss": 0.4426,
|
| 65 |
+
"mean_token_accuracy": 0.8728497698903084,
|
| 66 |
"step": 30
|
| 67 |
},
|
| 68 |
{
|
| 69 |
+
"epoch": 0.09048311520439489,
|
| 70 |
+
"grad_norm": 0.11570142209529877,
|
| 71 |
+
"learning_rate": 0.0001794871794871795,
|
| 72 |
+
"loss": 0.4326,
|
| 73 |
+
"mean_token_accuracy": 0.8739698991179466,
|
| 74 |
"step": 35
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 0.10340927451930845,
|
| 78 |
+
"grad_norm": 0.11375421285629272,
|
| 79 |
+
"learning_rate": 0.00019999590166142655,
|
| 80 |
+
"loss": 0.4398,
|
| 81 |
+
"mean_token_accuracy": 0.8683982357382775,
|
| 82 |
"step": 40
|
| 83 |
},
|
| 84 |
{
|
| 85 |
+
"epoch": 0.11633543383422201,
|
| 86 |
+
"grad_norm": 0.09867997467517853,
|
| 87 |
+
"learning_rate": 0.00019985249508066755,
|
| 88 |
+
"loss": 0.3217,
|
| 89 |
+
"mean_token_accuracy": 0.9007599778473377,
|
| 90 |
"step": 45
|
| 91 |
},
|
| 92 |
{
|
| 93 |
+
"epoch": 0.12926159314913557,
|
| 94 |
+
"grad_norm": 0.10740496963262558,
|
| 95 |
+
"learning_rate": 0.00019950450737506824,
|
| 96 |
+
"loss": 0.3468,
|
| 97 |
+
"mean_token_accuracy": 0.8914886720478534,
|
| 98 |
"step": 50
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.14218775246404913,
|
| 102 |
+
"grad_norm": 0.10870860517024994,
|
| 103 |
+
"learning_rate": 0.00019895265151345518,
|
| 104 |
+
"loss": 0.297,
|
| 105 |
+
"mean_token_accuracy": 0.9054083719849586,
|
| 106 |
"step": 55
|
| 107 |
},
|
| 108 |
{
|
| 109 |
+
"epoch": 0.15511391177896267,
|
| 110 |
+
"grad_norm": 0.1147221028804779,
|
| 111 |
+
"learning_rate": 0.00019819805815653768,
|
| 112 |
+
"loss": 0.2912,
|
| 113 |
+
"mean_token_accuracy": 0.9063926823437214,
|
| 114 |
"step": 60
|
| 115 |
},
|
| 116 |
{
|
| 117 |
+
"epoch": 0.16804007109387623,
|
| 118 |
+
"grad_norm": 0.1086113452911377,
|
| 119 |
+
"learning_rate": 0.00019724227334037256,
|
| 120 |
+
"loss": 0.2709,
|
| 121 |
+
"mean_token_accuracy": 0.9122237786650658,
|
| 122 |
"step": 65
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 0.18096623040878979,
|
| 126 |
+
"grad_norm": 0.11263403296470642,
|
| 127 |
+
"learning_rate": 0.00019608725530879375,
|
| 128 |
+
"loss": 0.2775,
|
| 129 |
+
"mean_token_accuracy": 0.9088832184672355,
|
| 130 |
"step": 70
|
| 131 |
},
|
| 132 |
{
|
| 133 |
+
"epoch": 0.19389238972370335,
|
| 134 |
+
"grad_norm": 0.11787361651659012,
|
| 135 |
+
"learning_rate": 0.00019473537050129704,
|
| 136 |
+
"loss": 0.2758,
|
| 137 |
+
"mean_token_accuracy": 0.909870583564043,
|
| 138 |
"step": 75
|
| 139 |
},
|
| 140 |
{
|
| 141 |
+
"epoch": 0.2068185490386169,
|
| 142 |
+
"grad_norm": 0.0966123640537262,
|
| 143 |
+
"learning_rate": 0.00019318938870459984,
|
| 144 |
+
"loss": 0.2113,
|
| 145 |
+
"mean_token_accuracy": 0.9297552116215229,
|
| 146 |
"step": 80
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 0.21974470835353047,
|
| 150 |
+
"grad_norm": 0.09169968962669373,
|
| 151 |
+
"learning_rate": 0.00019145247737780961,
|
| 152 |
+
"loss": 0.2097,
|
| 153 |
+
"mean_token_accuracy": 0.9294927291572094,
|
| 154 |
"step": 85
|
| 155 |
},
|
| 156 |
{
|
| 157 |
+
"epoch": 0.23267086766844403,
|
| 158 |
+
"grad_norm": 0.10205108672380447,
|
| 159 |
+
"learning_rate": 0.0001895281951628281,
|
| 160 |
+
"loss": 0.2354,
|
| 161 |
+
"mean_token_accuracy": 0.9218288294970989,
|
| 162 |
"step": 90
|
| 163 |
},
|
| 164 |
{
|
| 165 |
+
"epoch": 0.24559702698335756,
|
| 166 |
+
"grad_norm": 0.11714070290327072,
|
| 167 |
+
"learning_rate": 0.00018742048459328682,
|
| 168 |
+
"loss": 0.2286,
|
| 169 |
+
"mean_token_accuracy": 0.9224011048674583,
|
| 170 |
"step": 95
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 0.25852318629827115,
|
| 174 |
+
"grad_norm": 0.08743036538362503,
|
| 175 |
+
"learning_rate": 0.00018513366401695276,
|
| 176 |
+
"loss": 0.2524,
|
| 177 |
+
"mean_token_accuracy": 0.9145891763269901,
|
| 178 |
"step": 100
|
| 179 |
},
|
| 180 |
{
|
| 181 |
+
"epoch": 0.2714493456131847,
|
| 182 |
+
"grad_norm": 0.12086658179759979,
|
| 183 |
+
"learning_rate": 0.00018267241874815314,
|
| 184 |
+
"loss": 0.2355,
|
| 185 |
+
"mean_token_accuracy": 0.9212763957679272,
|
| 186 |
"step": 105
|
| 187 |
},
|
| 188 |
{
|
| 189 |
+
"epoch": 0.28437550492809827,
|
| 190 |
+
"grad_norm": 0.10271570086479187,
|
| 191 |
+
"learning_rate": 0.0001800417914683471,
|
| 192 |
+
"loss": 0.2033,
|
| 193 |
+
"mean_token_accuracy": 0.9312141306698323,
|
| 194 |
"step": 110
|
| 195 |
},
|
| 196 |
{
|
| 197 |
+
"epoch": 0.2973016642430118,
|
| 198 |
+
"grad_norm": 0.11228019744157791,
|
| 199 |
+
"learning_rate": 0.0001772471718945119,
|
| 200 |
+
"loss": 0.1578,
|
| 201 |
+
"mean_token_accuracy": 0.946045958250761,
|
| 202 |
"step": 115
|
| 203 |
},
|
| 204 |
{
|
| 205 |
+
"epoch": 0.31022782355792533,
|
| 206 |
+
"grad_norm": 0.1075127124786377,
|
| 207 |
+
"learning_rate": 0.00017429428573651024,
|
| 208 |
+
"loss": 0.2109,
|
| 209 |
+
"mean_token_accuracy": 0.928734278678894,
|
| 210 |
"step": 120
|
| 211 |
},
|
| 212 |
{
|
| 213 |
+
"epoch": 0.3231539828728389,
|
| 214 |
+
"grad_norm": 0.16245290637016296,
|
| 215 |
+
"learning_rate": 0.00017118918296606537,
|
| 216 |
+
"loss": 0.2544,
|
| 217 |
+
"mean_token_accuracy": 0.9145586542785168,
|
| 218 |
"step": 125
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 0.33608014218775245,
|
| 222 |
+
"grad_norm": 0.0984274297952652,
|
| 223 |
+
"learning_rate": 0.0001679382254213768,
|
| 224 |
+
"loss": 0.2398,
|
| 225 |
+
"mean_token_accuracy": 0.9193581290543079,
|
| 226 |
"step": 130
|
| 227 |
},
|
| 228 |
{
|
| 229 |
+
"epoch": 0.34900630150266604,
|
| 230 |
+
"grad_norm": 0.14143706858158112,
|
| 231 |
+
"learning_rate": 0.00016454807377277398,
|
| 232 |
+
"loss": 0.2187,
|
| 233 |
+
"mean_token_accuracy": 0.9259690448641777,
|
| 234 |
"step": 135
|
| 235 |
},
|
| 236 |
{
|
| 237 |
+
"epoch": 0.36193246081757957,
|
| 238 |
+
"grad_norm": 0.10752697288990021,
|
| 239 |
+
"learning_rate": 0.0001610256738761125,
|
| 240 |
+
"loss": 0.2205,
|
| 241 |
+
"mean_token_accuracy": 0.9248595975339413,
|
| 242 |
"step": 140
|
| 243 |
},
|
| 244 |
{
|
| 245 |
+
"epoch": 0.37485862013249316,
|
| 246 |
+
"grad_norm": 0.15562467277050018,
|
| 247 |
+
"learning_rate": 0.00015737824254187275,
|
| 248 |
+
"loss": 0.2399,
|
| 249 |
+
"mean_token_accuracy": 0.9187141321599483,
|
| 250 |
"step": 145
|
| 251 |
},
|
| 252 |
{
|
| 253 |
+
"epoch": 0.3877847794474067,
|
| 254 |
+
"grad_norm": 0.10149288177490234,
|
| 255 |
+
"learning_rate": 0.00015361325274911779,
|
| 256 |
+
"loss": 0.1906,
|
| 257 |
+
"mean_token_accuracy": 0.9351025439798832,
|
| 258 |
"step": 150
|
| 259 |
},
|
| 260 |
{
|
| 261 |
+
"epoch": 0.4007109387623202,
|
| 262 |
+
"grad_norm": 0.13106182217597961,
|
| 263 |
+
"learning_rate": 0.00014973841833460457,
|
| 264 |
+
"loss": 0.2366,
|
| 265 |
+
"mean_token_accuracy": 0.9191612683236599,
|
| 266 |
"step": 155
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 0.4136370980772338,
|
| 270 |
+
"grad_norm": 0.12485964596271515,
|
| 271 |
+
"learning_rate": 0.0001457616781884173,
|
| 272 |
+
"loss": 0.2532,
|
| 273 |
+
"mean_token_accuracy": 0.9145220316946506,
|
| 274 |
"step": 160
|
| 275 |
},
|
| 276 |
{
|
| 277 |
+
"epoch": 0.42656325739214734,
|
| 278 |
+
"grad_norm": 0.12415427714586258,
|
| 279 |
+
"learning_rate": 0.0001416911799885049,
|
| 280 |
+
"loss": 0.1772,
|
| 281 |
+
"mean_token_accuracy": 0.9388307243585586,
|
| 282 |
"step": 165
|
| 283 |
},
|
| 284 |
{
|
| 285 |
+
"epoch": 0.43948941670706093,
|
| 286 |
+
"grad_norm": 0.12156961858272552,
|
| 287 |
+
"learning_rate": 0.0001375352635074461,
|
| 288 |
+
"loss": 0.1974,
|
| 289 |
+
"mean_token_accuracy": 0.9322396464645862,
|
| 290 |
"step": 170
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 0.45241557602197446,
|
| 294 |
+
"grad_norm": 0.08770665526390076,
|
| 295 |
+
"learning_rate": 0.00013330244352564527,
|
| 296 |
+
"loss": 0.2159,
|
| 297 |
+
"mean_token_accuracy": 0.9263734519481659,
|
| 298 |
"step": 175
|
| 299 |
},
|
| 300 |
{
|
| 301 |
+
"epoch": 0.46534173533688805,
|
| 302 |
+
"grad_norm": 0.11970090866088867,
|
| 303 |
+
"learning_rate": 0.00012900139238596598,
|
| 304 |
+
"loss": 0.1788,
|
| 305 |
+
"mean_token_accuracy": 0.9383759558200836,
|
| 306 |
"step": 180
|
| 307 |
},
|
| 308 |
{
|
| 309 |
+
"epoch": 0.4782678946518016,
|
| 310 |
+
"grad_norm": 0.0777999758720398,
|
| 311 |
+
"learning_rate": 0.00012464092222554552,
|
| 312 |
+
"loss": 0.1921,
|
| 313 |
+
"mean_token_accuracy": 0.9340388782322406,
|
| 314 |
"step": 185
|
| 315 |
},
|
| 316 |
{
|
| 317 |
+
"epoch": 0.4911940539667151,
|
| 318 |
+
"grad_norm": 0.1088053435087204,
|
| 319 |
+
"learning_rate": 0.00012022996692119424,
|
| 320 |
+
"loss": 0.2428,
|
| 321 |
+
"mean_token_accuracy": 0.9170919217169284,
|
| 322 |
"step": 190
|
| 323 |
},
|
| 324 |
{
|
| 325 |
+
"epoch": 0.5041202132816287,
|
| 326 |
+
"grad_norm": 0.12958504259586334,
|
| 327 |
+
"learning_rate": 0.00011577756378537033,
|
| 328 |
+
"loss": 0.1805,
|
| 329 |
+
"mean_token_accuracy": 0.9373964861035347,
|
| 330 |
"step": 195
|
| 331 |
},
|
| 332 |
{
|
| 333 |
+
"epoch": 0.5170463725965423,
|
| 334 |
+
"grad_norm": 0.08942475914955139,
|
| 335 |
+
"learning_rate": 0.00011129283505023274,
|
| 336 |
+
"loss": 0.1757,
|
| 337 |
+
"mean_token_accuracy": 0.9391221977770329,
|
| 338 |
"step": 200
|
| 339 |
},
|
| 340 |
{
|
| 341 |
+
"epoch": 0.5299725319114558,
|
| 342 |
+
"grad_norm": 0.12055882811546326,
|
| 343 |
+
"learning_rate": 0.00010678496917770719,
|
| 344 |
+
"loss": 0.2274,
|
| 345 |
+
"mean_token_accuracy": 0.9221370957791806,
|
| 346 |
"step": 205
|
| 347 |
},
|
| 348 |
{
|
| 349 |
+
"epoch": 0.5428986912263694,
|
| 350 |
+
"grad_norm": 0.09946688264608383,
|
| 351 |
+
"learning_rate": 0.00010226320203385878,
|
| 352 |
+
"loss": 0.2235,
|
| 353 |
+
"mean_token_accuracy": 0.9239851593971252,
|
| 354 |
"step": 210
|
| 355 |
},
|
| 356 |
{
|
| 357 |
+
"epoch": 0.5558248505412829,
|
| 358 |
+
"grad_norm": 0.08823594450950623,
|
| 359 |
+
"learning_rate": 9.773679796614124e-05,
|
| 360 |
+
"loss": 0.1757,
|
| 361 |
+
"mean_token_accuracy": 0.9392839625477791,
|
| 362 |
"step": 215
|
| 363 |
},
|
| 364 |
{
|
| 365 |
+
"epoch": 0.5687510098561965,
|
| 366 |
+
"grad_norm": 0.09366550296545029,
|
| 367 |
+
"learning_rate": 9.321503082229282e-05,
|
| 368 |
+
"loss": 0.215,
|
| 369 |
+
"mean_token_accuracy": 0.9264558620750905,
|
| 370 |
"step": 220
|
| 371 |
},
|
| 372 |
{
|
| 373 |
+
"epoch": 0.58167716917111,
|
| 374 |
+
"grad_norm": 0.1250247210264206,
|
| 375 |
+
"learning_rate": 8.87071649497673e-05,
|
| 376 |
+
"loss": 0.203,
|
| 377 |
+
"mean_token_accuracy": 0.9299200311303139,
|
| 378 |
"step": 225
|
| 379 |
},
|
| 380 |
{
|
| 381 |
+
"epoch": 0.5946033284860236,
|
| 382 |
+
"grad_norm": 0.07920438051223755,
|
| 383 |
+
"learning_rate": 8.422243621462969e-05,
|
| 384 |
+
"loss": 0.1777,
|
| 385 |
+
"mean_token_accuracy": 0.93870100826025,
|
| 386 |
"step": 230
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 0.6075294878009372,
|
| 390 |
+
"grad_norm": 0.09220347553491592,
|
| 391 |
+
"learning_rate": 7.97700330788058e-05,
|
| 392 |
+
"loss": 0.2382,
|
| 393 |
+
"mean_token_accuracy": 0.9188789471983909,
|
| 394 |
"step": 235
|
| 395 |
},
|
| 396 |
{
|
| 397 |
+
"epoch": 0.6204556471158507,
|
| 398 |
+
"grad_norm": 0.09347136318683624,
|
| 399 |
+
"learning_rate": 7.535907777445449e-05,
|
| 400 |
+
"loss": 0.1768,
|
| 401 |
+
"mean_token_accuracy": 0.9390825219452381,
|
| 402 |
"step": 240
|
| 403 |
},
|
| 404 |
{
|
| 405 |
+
"epoch": 0.6333818064307642,
|
| 406 |
+
"grad_norm": 0.06848734617233276,
|
| 407 |
+
"learning_rate": 7.099860761403403e-05,
|
| 408 |
+
"loss": 0.1909,
|
| 409 |
+
"mean_token_accuracy": 0.9346477761864662,
|
| 410 |
"step": 245
|
| 411 |
},
|
| 412 |
{
|
| 413 |
+
"epoch": 0.6463079657456778,
|
| 414 |
+
"grad_norm": 0.08231879770755768,
|
| 415 |
+
"learning_rate": 6.669755647435474e-05,
|
| 416 |
+
"loss": 0.2094,
|
| 417 |
+
"mean_token_accuracy": 0.9281139463186264,
|
| 418 |
+
"step": 250
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.6592341250605914,
|
| 422 |
+
"grad_norm": 0.14479972422122955,
|
| 423 |
+
"learning_rate": 6.24647364925539e-05,
|
| 424 |
+
"loss": 0.1897,
|
| 425 |
+
"mean_token_accuracy": 0.9346370972692967,
|
| 426 |
+
"step": 255
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 0.6721602843755049,
|
| 430 |
+
"grad_norm": 0.10493209958076477,
|
| 431 |
+
"learning_rate": 5.830882001149517e-05,
|
| 432 |
+
"loss": 0.1977,
|
| 433 |
+
"mean_token_accuracy": 0.9318774163722991,
|
| 434 |
+
"step": 260
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.6850864436904185,
|
| 438 |
+
"grad_norm": 0.08989942073822021,
|
| 439 |
+
"learning_rate": 5.423832181158274e-05,
|
| 440 |
+
"loss": 0.1786,
|
| 441 |
+
"mean_token_accuracy": 0.9378982990980148,
|
| 442 |
+
"step": 265
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 0.6980126030053321,
|
| 446 |
+
"grad_norm": 0.0723404809832573,
|
| 447 |
+
"learning_rate": 5.0261581665395475e-05,
|
| 448 |
+
"loss": 0.1854,
|
| 449 |
+
"mean_token_accuracy": 0.9353665545582771,
|
| 450 |
+
"step": 270
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.7109387623202456,
|
| 454 |
+
"grad_norm": 0.08017778396606445,
|
| 455 |
+
"learning_rate": 4.6386747250882224e-05,
|
| 456 |
+
"loss": 0.2185,
|
| 457 |
+
"mean_token_accuracy": 0.9242644280195236,
|
| 458 |
+
"step": 275
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.7238649216351591,
|
| 462 |
+
"grad_norm": 0.09016852080821991,
|
| 463 |
+
"learning_rate": 4.2621757458127285e-05,
|
| 464 |
+
"loss": 0.1912,
|
| 465 |
+
"mean_token_accuracy": 0.9338557474315167,
|
| 466 |
+
"step": 280
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.7367910809500727,
|
| 470 |
+
"grad_norm": 0.0839182510972023,
|
| 471 |
+
"learning_rate": 3.8974326123887515e-05,
|
| 472 |
+
"loss": 0.1764,
|
| 473 |
+
"mean_token_accuracy": 0.9384186826646328,
|
| 474 |
+
"step": 285
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.7497172402649863,
|
| 478 |
+
"grad_norm": 0.08571181446313858,
|
| 479 |
+
"learning_rate": 3.5451926227225997e-05,
|
| 480 |
+
"loss": 0.1984,
|
| 481 |
+
"mean_token_accuracy": 0.931569704413414,
|
| 482 |
+
"step": 290
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.7626433995798998,
|
| 486 |
+
"grad_norm": 0.09723369032144547,
|
| 487 |
+
"learning_rate": 3.20617745786232e-05,
|
| 488 |
+
"loss": 0.2095,
|
| 489 |
+
"mean_token_accuracy": 0.9279029227793216,
|
| 490 |
+
"step": 295
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 0.7755695588948134,
|
| 494 |
+
"grad_norm": 0.08322826772928238,
|
| 495 |
+
"learning_rate": 2.8810817033934656e-05,
|
| 496 |
+
"loss": 0.214,
|
| 497 |
+
"mean_token_accuracy": 0.926049928367138,
|
| 498 |
+
"step": 300
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 0.788495718209727,
|
| 502 |
+
"grad_norm": 0.0943559929728508,
|
| 503 |
+
"learning_rate": 2.5705714263489776e-05,
|
| 504 |
+
"loss": 0.2025,
|
| 505 |
+
"mean_token_accuracy": 0.9304434671998024,
|
| 506 |
+
"step": 305
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.8014218775246404,
|
| 510 |
+
"grad_norm": 0.12426702678203583,
|
| 511 |
+
"learning_rate": 2.275282810548811e-05,
|
| 512 |
+
"loss": 0.2255,
|
| 513 |
+
"mean_token_accuracy": 0.92267579510808,
|
| 514 |
+
"step": 310
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.814348036839554,
|
| 518 |
+
"grad_norm": 0.07339806854724884,
|
| 519 |
+
"learning_rate": 1.9958208531652877e-05,
|
| 520 |
+
"loss": 0.2119,
|
| 521 |
+
"mean_token_accuracy": 0.9270739153027534,
|
| 522 |
+
"step": 315
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.8272741961544676,
|
| 526 |
+
"grad_norm": 0.07293356955051422,
|
| 527 |
+
"learning_rate": 1.73275812518469e-05,
|
| 528 |
+
"loss": 0.1728,
|
| 529 |
+
"mean_token_accuracy": 0.9397997766733169,
|
| 530 |
+
"step": 320
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.8402003554693812,
|
| 534 |
+
"grad_norm": 0.06353917717933655,
|
| 535 |
+
"learning_rate": 1.4866335983047264e-05,
|
| 536 |
+
"loss": 0.2043,
|
| 537 |
+
"mean_token_accuracy": 0.9285617105662822,
|
| 538 |
+
"step": 325
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 0.8531265147842947,
|
| 542 |
+
"grad_norm": 0.0852632075548172,
|
| 543 |
+
"learning_rate": 1.2579515406713193e-05,
|
| 544 |
+
"loss": 0.1894,
|
| 545 |
+
"mean_token_accuracy": 0.9348805241286755,
|
| 546 |
+
"step": 330
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 0.8660526740992083,
|
| 550 |
+
"grad_norm": 0.10367190837860107,
|
| 551 |
+
"learning_rate": 1.0471804837171916e-05,
|
| 552 |
+
"loss": 0.2155,
|
| 553 |
+
"mean_token_accuracy": 0.9256119452416897,
|
| 554 |
+
"step": 335
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.8789788334141219,
|
| 558 |
+
"grad_norm": 0.07647697627544403,
|
| 559 |
+
"learning_rate": 8.547522622190385e-06,
|
| 560 |
+
"loss": 0.191,
|
| 561 |
+
"mean_token_accuracy": 0.9336939886212349,
|
| 562 |
+
"step": 340
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.8919049927290353,
|
| 566 |
+
"grad_norm": 0.09329604357481003,
|
| 567 |
+
"learning_rate": 6.810611295400171e-06,
|
| 568 |
+
"loss": 0.162,
|
| 569 |
+
"mean_token_accuracy": 0.9432654656469822,
|
| 570 |
+
"step": 345
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.9048311520439489,
|
| 574 |
+
"grad_norm": 0.10007894039154053,
|
| 575 |
+
"learning_rate": 5.264629498702967e-06,
|
| 576 |
+
"loss": 0.2196,
|
| 577 |
+
"mean_token_accuracy": 0.9224743604660034,
|
| 578 |
+
"step": 350
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.9177573113588625,
|
| 582 |
+
"grad_norm": 0.09568954259157181,
|
| 583 |
+
"learning_rate": 3.91274469120626e-06,
|
| 584 |
+
"loss": 0.2326,
|
| 585 |
+
"mean_token_accuracy": 0.9203760787844658,
|
| 586 |
+
"step": 355
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 0.9306834706737761,
|
| 590 |
+
"grad_norm": 0.09242628514766693,
|
| 591 |
+
"learning_rate": 2.7577266596274576e-06,
|
| 592 |
+
"loss": 0.217,
|
| 593 |
+
"mean_token_accuracy": 0.9255997397005558,
|
| 594 |
+
"step": 360
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.9436096299886896,
|
| 598 |
+
"grad_norm": 0.08556357026100159,
|
| 599 |
+
"learning_rate": 1.8019418434623404e-06,
|
| 600 |
+
"loss": 0.1704,
|
| 601 |
+
"mean_token_accuracy": 0.9403167776763439,
|
| 602 |
+
"step": 365
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.9565357893036032,
|
| 606 |
+
"grad_norm": 0.08201264590024948,
|
| 607 |
+
"learning_rate": 1.0473484865448525e-06,
|
| 608 |
+
"loss": 0.1877,
|
| 609 |
+
"mean_token_accuracy": 0.9344814352691173,
|
| 610 |
+
"step": 370
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 0.9694619486185168,
|
| 614 |
+
"grad_norm": 0.08764708787202835,
|
| 615 |
+
"learning_rate": 4.954926249317815e-07,
|
| 616 |
+
"loss": 0.1614,
|
| 617 |
+
"mean_token_accuracy": 0.943453174829483,
|
| 618 |
+
"step": 375
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.9823881079334302,
|
| 622 |
+
"grad_norm": 0.09879707545042038,
|
| 623 |
+
"learning_rate": 1.4750491933247512e-07,
|
| 624 |
+
"loss": 0.1984,
|
| 625 |
+
"mean_token_accuracy": 0.9314216762781143,
|
| 626 |
+
"step": 380
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.9953142672483438,
|
| 630 |
+
"grad_norm": 0.0833079144358635,
|
| 631 |
+
"learning_rate": 4.0983385734660875e-09,
|
| 632 |
+
"loss": 0.1962,
|
| 633 |
+
"mean_token_accuracy": 0.9319176472723484,
|
| 634 |
+
"step": 385
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.9978994991113266,
|
| 638 |
+
"mean_token_accuracy": 0.9182563126087189,
|
| 639 |
+
"step": 386,
|
| 640 |
+
"total_flos": 1.112161775475753e+17,
|
| 641 |
+
"train_loss": 0.24216800009339584,
|
| 642 |
+
"train_runtime": 6297.1418,
|
| 643 |
+
"train_samples_per_second": 0.983,
|
| 644 |
+
"train_steps_per_second": 0.061
|
| 645 |
}
|
| 646 |
],
|
| 647 |
"logging_steps": 5,
|
| 648 |
+
"max_steps": 386,
|
| 649 |
"num_input_tokens_seen": 0,
|
| 650 |
"num_train_epochs": 1,
|
| 651 |
"save_steps": 100,
|
|
|
|
| 661 |
"attributes": {}
|
| 662 |
}
|
| 663 |
},
|
| 664 |
+
"total_flos": 1.112161775475753e+17,
|
| 665 |
"train_batch_size": 1,
|
| 666 |
"trial_name": null,
|
| 667 |
"trial_params": null
|