Commit
•
278086c
1
Parent(s):
2203e51
huggingartists
Browse files- README.md +3 -3
- config.json +1 -1
- evaluation.txt +1 -1
- flax_model.msgpack +1 -1
- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +2 -2
- scheduler.pt +1 -1
- trainer_state.json +441 -2099
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -45,15 +45,15 @@ from datasets import load_dataset
|
|
| 45 |
dataset = load_dataset("huggingartists/eminem")
|
| 46 |
```
|
| 47 |
|
| 48 |
-
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/
|
| 49 |
|
| 50 |
## Training procedure
|
| 51 |
|
| 52 |
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on Eminem's lyrics.
|
| 53 |
|
| 54 |
-
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/
|
| 55 |
|
| 56 |
-
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/
|
| 57 |
|
| 58 |
## How to use
|
| 59 |
|
|
|
|
| 45 |
dataset = load_dataset("huggingartists/eminem")
|
| 46 |
```
|
| 47 |
|
| 48 |
+
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/3ugwysdf/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
|
| 49 |
|
| 50 |
## Training procedure
|
| 51 |
|
| 52 |
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on Eminem's lyrics.
|
| 53 |
|
| 54 |
+
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/3ppfpolr) for full transparency and reproducibility.
|
| 55 |
|
| 56 |
+
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/3ppfpolr/artifacts) is logged and versioned.
|
| 57 |
|
| 58 |
## How to use
|
| 59 |
|
config.json
CHANGED
|
@@ -35,7 +35,7 @@
|
|
| 35 |
}
|
| 36 |
},
|
| 37 |
"torch_dtype": "float32",
|
| 38 |
-
"transformers_version": "4.10.
|
| 39 |
"use_cache": true,
|
| 40 |
"vocab_size": 50257
|
| 41 |
}
|
|
|
|
| 35 |
}
|
| 36 |
},
|
| 37 |
"torch_dtype": "float32",
|
| 38 |
+
"transformers_version": "4.10.2",
|
| 39 |
"use_cache": true,
|
| 40 |
"vocab_size": 50257
|
| 41 |
}
|
evaluation.txt
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"eval_loss": 2.
|
|
|
|
| 1 |
+
{"eval_loss": 2.4912426471710205, "eval_runtime": 31.0278, "eval_samples_per_second": 20.659, "eval_steps_per_second": 2.611, "epoch": 6.0}
|
flax_model.msgpack
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 497764120
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba86a7379d53933ac5e5218639ca8f842f5e002bf6366e9a3fd24bed24f28122
|
| 3 |
size 497764120
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 995604017
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7afdd4d66f7a5fb461c1f5449c0a88b116edb6baa5c5a18d3f1fc451d2c39538
|
| 3 |
size 995604017
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 510403817
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:511ec31b47c42ac20592874aafd4053e011c0a6029d1762336507dee879f4b62
|
| 3 |
size 510403817
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:706c9e2b7a9643251c7e54913d71cd38a14dc9ea9c2c37f5a6c67a08aae9d978
|
| 3 |
+
size 14503
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c0538d80c9e5b4ba41140beacd1f45a01b77629e237c5c71b0ff1ae309ea5115
|
| 3 |
size 623
|
trainer_state.json
CHANGED
|
@@ -1,2782 +1,1124 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 2.
|
| 3 |
-
"best_model_checkpoint": "output/eminem/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
-
"learning_rate":
|
| 13 |
-
"loss":
|
| 14 |
"step": 5
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss": 2.
|
| 20 |
"step": 10
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.03,
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss":
|
| 26 |
"step": 15
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.04,
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss":
|
| 32 |
"step": 20
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.05,
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss":
|
| 38 |
"step": 25
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.07,
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss":
|
| 44 |
"step": 30
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.08,
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss":
|
| 50 |
"step": 35
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.09,
|
| 54 |
-
"learning_rate":
|
| 55 |
-
"loss":
|
| 56 |
"step": 40
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.1,
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss":
|
| 62 |
"step": 45
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.11,
|
| 66 |
-
"learning_rate":
|
| 67 |
-
"loss":
|
| 68 |
"step": 50
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.12,
|
| 72 |
-
"learning_rate":
|
| 73 |
-
"loss": 2.
|
| 74 |
"step": 55
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.13,
|
| 78 |
-
"learning_rate":
|
| 79 |
-
"loss":
|
| 80 |
"step": 60
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.14,
|
| 84 |
-
"learning_rate":
|
| 85 |
-
"loss":
|
| 86 |
"step": 65
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.15,
|
| 90 |
-
"learning_rate":
|
| 91 |
-
"loss":
|
| 92 |
"step": 70
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.16,
|
| 96 |
-
"learning_rate":
|
| 97 |
-
"loss":
|
| 98 |
"step": 75
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate":
|
| 103 |
-
"loss":
|
| 104 |
"step": 80
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.19,
|
| 108 |
-
"learning_rate":
|
| 109 |
-
"loss":
|
| 110 |
"step": 85
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.2,
|
| 114 |
-
"learning_rate":
|
| 115 |
-
"loss":
|
| 116 |
"step": 90
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.21,
|
| 120 |
-
"learning_rate":
|
| 121 |
-
"loss":
|
| 122 |
"step": 95
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.22,
|
| 126 |
-
"learning_rate":
|
| 127 |
-
"loss":
|
| 128 |
"step": 100
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.23,
|
| 132 |
-
"learning_rate":
|
| 133 |
-
"loss":
|
| 134 |
"step": 105
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.24,
|
| 138 |
-
"learning_rate":
|
| 139 |
-
"loss": 2.
|
| 140 |
"step": 110
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.25,
|
| 144 |
-
"learning_rate":
|
| 145 |
-
"loss":
|
| 146 |
"step": 115
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.26,
|
| 150 |
-
"learning_rate":
|
| 151 |
-
"loss":
|
| 152 |
"step": 120
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.27,
|
| 156 |
-
"learning_rate":
|
| 157 |
-
"loss": 2.
|
| 158 |
"step": 125
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch": 0.
|
| 162 |
-
"learning_rate":
|
| 163 |
-
"loss":
|
| 164 |
"step": 130
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch": 0.
|
| 168 |
-
"learning_rate":
|
| 169 |
-
"loss":
|
| 170 |
"step": 135
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.31,
|
| 174 |
-
"learning_rate":
|
| 175 |
-
"loss":
|
| 176 |
"step": 140
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.32,
|
| 180 |
-
"learning_rate":
|
| 181 |
-
"loss":
|
| 182 |
"step": 145
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.33,
|
| 186 |
-
"learning_rate":
|
| 187 |
-
"loss":
|
| 188 |
"step": 150
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.34,
|
| 192 |
-
"learning_rate":
|
| 193 |
-
"loss":
|
| 194 |
"step": 155
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.35,
|
| 198 |
-
"learning_rate":
|
| 199 |
-
"loss":
|
| 200 |
"step": 160
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.36,
|
| 204 |
-
"learning_rate":
|
| 205 |
-
"loss":
|
| 206 |
"step": 165
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.37,
|
| 210 |
-
"learning_rate":
|
| 211 |
-
"loss": 2.
|
| 212 |
"step": 170
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.38,
|
| 216 |
-
"learning_rate":
|
| 217 |
-
"loss":
|
| 218 |
"step": 175
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch": 0.
|
| 222 |
-
"learning_rate":
|
| 223 |
-
"loss":
|
| 224 |
"step": 180
|
| 225 |
},
|
| 226 |
{
|
| 227 |
-
"epoch": 0.
|
| 228 |
-
"learning_rate":
|
| 229 |
-
"loss":
|
| 230 |
"step": 185
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch": 0.
|
| 234 |
-
"learning_rate":
|
| 235 |
-
"loss":
|
| 236 |
"step": 190
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.43,
|
| 240 |
-
"learning_rate":
|
| 241 |
-
"loss":
|
| 242 |
"step": 195
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.44,
|
| 246 |
-
"learning_rate":
|
| 247 |
-
"loss":
|
| 248 |
"step": 200
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.45,
|
| 252 |
-
"learning_rate":
|
| 253 |
-
"loss":
|
| 254 |
"step": 205
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.46,
|
| 258 |
-
"learning_rate":
|
| 259 |
-
"loss":
|
| 260 |
"step": 210
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 0.47,
|
| 264 |
-
"learning_rate":
|
| 265 |
-
"loss":
|
| 266 |
"step": 215
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 0.48,
|
| 270 |
-
"learning_rate": 7.
|
| 271 |
-
"loss":
|
| 272 |
"step": 220
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.49,
|
| 276 |
-
"learning_rate": 7.
|
| 277 |
-
"loss":
|
| 278 |
"step": 225
|
| 279 |
},
|
| 280 |
{
|
| 281 |
-
"epoch": 0.
|
| 282 |
-
"learning_rate":
|
| 283 |
-
"loss": 2.
|
| 284 |
"step": 230
|
| 285 |
},
|
| 286 |
{
|
| 287 |
-
"epoch": 0.
|
| 288 |
-
"learning_rate":
|
| 289 |
-
"loss":
|
| 290 |
"step": 235
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch": 0.
|
| 294 |
-
"learning_rate":
|
| 295 |
-
"loss":
|
| 296 |
"step": 240
|
| 297 |
},
|
| 298 |
{
|
| 299 |
-
"epoch": 0.
|
| 300 |
-
"learning_rate":
|
| 301 |
-
"loss": 2.
|
| 302 |
"step": 245
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.55,
|
| 306 |
-
"learning_rate":
|
| 307 |
-
"loss":
|
| 308 |
"step": 250
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.56,
|
| 312 |
-
"learning_rate":
|
| 313 |
-
"loss":
|
| 314 |
"step": 255
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 0.57,
|
| 318 |
-
"learning_rate":
|
| 319 |
-
"loss":
|
| 320 |
"step": 260
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 0.58,
|
| 324 |
-
"learning_rate":
|
| 325 |
-
"loss": 2.
|
| 326 |
"step": 265
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.59,
|
| 330 |
-
"learning_rate":
|
| 331 |
-
"loss":
|
| 332 |
"step": 270
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 0.6,
|
| 336 |
-
"learning_rate":
|
| 337 |
-
"loss":
|
| 338 |
"step": 275
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch": 0.
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss":
|
| 344 |
"step": 280
|
| 345 |
},
|
| 346 |
{
|
| 347 |
-
"epoch": 0.
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss": 2.
|
| 350 |
"step": 285
|
| 351 |
},
|
| 352 |
{
|
| 353 |
-
"epoch": 0.
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss":
|
| 356 |
"step": 290
|
| 357 |
},
|
| 358 |
{
|
| 359 |
-
"epoch": 0.
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss":
|
| 362 |
"step": 295
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 0.66,
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss":
|
| 368 |
"step": 300
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 0.67,
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss":
|
| 374 |
"step": 305
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 0.68,
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss": 2.
|
| 380 |
"step": 310
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.69,
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss":
|
| 386 |
"step": 315
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 0.7,
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss": 2.
|
| 392 |
"step": 320
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 0.71,
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss": 2.
|
| 398 |
"step": 325
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch": 0.
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss": 2.
|
| 404 |
"step": 330
|
| 405 |
},
|
| 406 |
{
|
| 407 |
-
"epoch": 0.
|
| 408 |
-
"learning_rate":
|
| 409 |
-
"loss":
|
| 410 |
"step": 335
|
| 411 |
},
|
| 412 |
{
|
| 413 |
-
"epoch": 0.
|
| 414 |
-
"learning_rate":
|
| 415 |
-
"loss":
|
| 416 |
"step": 340
|
| 417 |
},
|
| 418 |
{
|
| 419 |
-
"epoch": 0.
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"loss":
|
| 422 |
"step": 345
|
| 423 |
},
|
| 424 |
{
|
| 425 |
-
"epoch": 0.
|
| 426 |
-
"learning_rate":
|
| 427 |
-
"loss":
|
| 428 |
"step": 350
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 0.78,
|
| 432 |
-
"learning_rate":
|
| 433 |
-
"loss": 2.
|
| 434 |
"step": 355
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 0.79,
|
| 438 |
-
"learning_rate":
|
| 439 |
-
"loss": 2.
|
| 440 |
"step": 360
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 0.8,
|
| 444 |
-
"learning_rate":
|
| 445 |
-
"loss":
|
| 446 |
"step": 365
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 0.81,
|
| 450 |
-
"learning_rate":
|
| 451 |
-
"loss":
|
| 452 |
"step": 370
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 0.82,
|
| 456 |
-
"learning_rate":
|
| 457 |
-
"loss":
|
| 458 |
"step": 375
|
| 459 |
},
|
| 460 |
{
|
| 461 |
-
"epoch": 0.
|
| 462 |
-
"learning_rate":
|
| 463 |
-
"loss":
|
| 464 |
"step": 380
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch": 0.
|
| 468 |
-
"learning_rate":
|
| 469 |
-
"loss":
|
| 470 |
"step": 385
|
| 471 |
},
|
| 472 |
{
|
| 473 |
-
"epoch": 0.
|
| 474 |
-
"learning_rate":
|
| 475 |
-
"loss":
|
| 476 |
"step": 390
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch": 0.
|
| 480 |
-
"learning_rate":
|
| 481 |
-
"loss":
|
| 482 |
"step": 395
|
| 483 |
},
|
| 484 |
{
|
| 485 |
-
"epoch": 0.
|
| 486 |
-
"learning_rate":
|
| 487 |
-
"loss": 2.
|
| 488 |
"step": 400
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch": 0.
|
| 492 |
-
"learning_rate":
|
| 493 |
-
"loss": 2.
|
| 494 |
"step": 405
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 0.9,
|
| 498 |
-
"learning_rate":
|
| 499 |
-
"loss":
|
| 500 |
"step": 410
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 0.91,
|
| 504 |
-
"learning_rate":
|
| 505 |
-
"loss":
|
| 506 |
"step": 415
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.92,
|
| 510 |
-
"learning_rate":
|
| 511 |
-
"loss":
|
| 512 |
"step": 420
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 0.93,
|
| 516 |
-
"learning_rate":
|
| 517 |
-
"loss":
|
| 518 |
"step": 425
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch": 0.
|
| 522 |
-
"learning_rate":
|
| 523 |
-
"loss":
|
| 524 |
"step": 430
|
| 525 |
},
|
| 526 |
{
|
| 527 |
-
"epoch": 0.
|
| 528 |
-
"learning_rate":
|
| 529 |
-
"loss":
|
| 530 |
"step": 435
|
| 531 |
},
|
| 532 |
{
|
| 533 |
-
"epoch": 0.
|
| 534 |
-
"learning_rate":
|
| 535 |
-
"loss":
|
| 536 |
"step": 440
|
| 537 |
},
|
| 538 |
{
|
| 539 |
-
"epoch": 0.
|
| 540 |
-
"learning_rate":
|
| 541 |
-
"loss": 2.
|
| 542 |
"step": 445
|
| 543 |
},
|
| 544 |
{
|
| 545 |
-
"epoch": 0.
|
| 546 |
-
"learning_rate":
|
| 547 |
-
"loss": 2.
|
| 548 |
"step": 450
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch": 0
|
| 552 |
-
"learning_rate":
|
| 553 |
-
"loss":
|
| 554 |
"step": 455
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 1.0,
|
| 558 |
-
"eval_loss":
|
| 559 |
-
"eval_runtime":
|
| 560 |
-
"eval_samples_per_second":
|
| 561 |
-
"eval_steps_per_second": 2.
|
| 562 |
-
"step":
|
| 563 |
},
|
| 564 |
{
|
| 565 |
"epoch": 1.01,
|
| 566 |
-
"learning_rate":
|
| 567 |
-
"loss": 2.
|
| 568 |
"step": 460
|
| 569 |
},
|
| 570 |
{
|
| 571 |
"epoch": 1.02,
|
| 572 |
-
"learning_rate":
|
| 573 |
-
"loss": 2.
|
| 574 |
"step": 465
|
| 575 |
},
|
| 576 |
{
|
| 577 |
"epoch": 1.03,
|
| 578 |
-
"learning_rate":
|
| 579 |
-
"loss": 2.
|
| 580 |
"step": 470
|
| 581 |
},
|
| 582 |
{
|
| 583 |
"epoch": 1.04,
|
| 584 |
-
"learning_rate":
|
| 585 |
-
"loss":
|
| 586 |
"step": 475
|
| 587 |
},
|
| 588 |
{
|
| 589 |
"epoch": 1.05,
|
| 590 |
-
"learning_rate":
|
| 591 |
-
"loss": 2.
|
| 592 |
"step": 480
|
| 593 |
},
|
| 594 |
{
|
| 595 |
-
"epoch": 1.
|
| 596 |
-
"learning_rate":
|
| 597 |
-
"loss":
|
| 598 |
"step": 485
|
| 599 |
},
|
| 600 |
{
|
| 601 |
-
"epoch": 1.
|
| 602 |
-
"learning_rate":
|
| 603 |
-
"loss":
|
| 604 |
"step": 490
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 1.09,
|
| 608 |
-
"learning_rate":
|
| 609 |
-
"loss": 2.
|
| 610 |
"step": 495
|
| 611 |
},
|
| 612 |
{
|
| 613 |
"epoch": 1.1,
|
| 614 |
-
"learning_rate":
|
| 615 |
-
"loss":
|
| 616 |
"step": 500
|
| 617 |
},
|
| 618 |
{
|
| 619 |
"epoch": 1.11,
|
| 620 |
-
"learning_rate":
|
| 621 |
-
"loss": 2.
|
| 622 |
"step": 505
|
| 623 |
},
|
| 624 |
{
|
| 625 |
"epoch": 1.12,
|
| 626 |
-
"learning_rate":
|
| 627 |
-
"loss": 2.
|
| 628 |
"step": 510
|
| 629 |
},
|
| 630 |
{
|
| 631 |
"epoch": 1.13,
|
| 632 |
-
"learning_rate":
|
| 633 |
-
"loss":
|
| 634 |
"step": 515
|
| 635 |
},
|
| 636 |
{
|
| 637 |
"epoch": 1.14,
|
| 638 |
-
"learning_rate":
|
| 639 |
-
"loss": 2.
|
| 640 |
"step": 520
|
| 641 |
},
|
| 642 |
{
|
| 643 |
"epoch": 1.15,
|
| 644 |
-
"learning_rate":
|
| 645 |
-
"loss":
|
| 646 |
"step": 525
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 1.16,
|
| 650 |
-
"learning_rate":
|
| 651 |
-
"loss":
|
| 652 |
"step": 530
|
| 653 |
},
|
| 654 |
{
|
| 655 |
-
"epoch": 1.
|
| 656 |
-
"learning_rate":
|
| 657 |
-
"loss":
|
| 658 |
"step": 535
|
| 659 |
},
|
| 660 |
{
|
| 661 |
-
"epoch": 1.
|
| 662 |
-
"learning_rate":
|
| 663 |
-
"loss":
|
| 664 |
"step": 540
|
| 665 |
},
|
| 666 |
{
|
| 667 |
"epoch": 1.2,
|
| 668 |
-
"learning_rate":
|
| 669 |
-
"loss": 2.
|
| 670 |
"step": 545
|
| 671 |
},
|
| 672 |
{
|
| 673 |
"epoch": 1.21,
|
| 674 |
-
"learning_rate":
|
| 675 |
-
"loss": 2.
|
| 676 |
"step": 550
|
| 677 |
},
|
| 678 |
{
|
| 679 |
"epoch": 1.22,
|
| 680 |
-
"learning_rate":
|
| 681 |
-
"loss":
|
| 682 |
"step": 555
|
| 683 |
},
|
| 684 |
{
|
| 685 |
"epoch": 1.23,
|
| 686 |
-
"learning_rate":
|
| 687 |
-
"loss":
|
| 688 |
"step": 560
|
| 689 |
},
|
| 690 |
{
|
| 691 |
"epoch": 1.24,
|
| 692 |
-
"learning_rate":
|
| 693 |
-
"loss":
|
| 694 |
"step": 565
|
| 695 |
},
|
| 696 |
{
|
| 697 |
"epoch": 1.25,
|
| 698 |
-
"learning_rate":
|
| 699 |
-
"loss": 2.
|
| 700 |
"step": 570
|
| 701 |
},
|
| 702 |
{
|
| 703 |
"epoch": 1.26,
|
| 704 |
-
"learning_rate":
|
| 705 |
-
"loss":
|
| 706 |
"step": 575
|
| 707 |
},
|
| 708 |
{
|
| 709 |
"epoch": 1.27,
|
| 710 |
-
"learning_rate":
|
| 711 |
-
"loss": 2.
|
| 712 |
"step": 580
|
| 713 |
},
|
| 714 |
{
|
| 715 |
-
"epoch": 1.
|
| 716 |
-
"learning_rate":
|
| 717 |
-
"loss": 2.
|
| 718 |
"step": 585
|
| 719 |
},
|
| 720 |
{
|
| 721 |
-
"epoch": 1.
|
| 722 |
-
"learning_rate":
|
| 723 |
-
"loss": 2.
|
| 724 |
"step": 590
|
| 725 |
},
|
| 726 |
{
|
| 727 |
-
"epoch": 1.
|
| 728 |
-
"learning_rate":
|
| 729 |
-
"loss":
|
| 730 |
"step": 595
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 1.32,
|
| 734 |
-
"learning_rate":
|
| 735 |
-
"loss":
|
| 736 |
"step": 600
|
| 737 |
},
|
| 738 |
{
|
| 739 |
"epoch": 1.33,
|
| 740 |
-
"learning_rate":
|
| 741 |
-
"loss": 2.
|
| 742 |
"step": 605
|
| 743 |
},
|
| 744 |
{
|
| 745 |
"epoch": 1.34,
|
| 746 |
-
"learning_rate":
|
| 747 |
-
"loss":
|
| 748 |
"step": 610
|
| 749 |
},
|
| 750 |
{
|
| 751 |
"epoch": 1.35,
|
| 752 |
-
"learning_rate":
|
| 753 |
-
"loss": 2.
|
| 754 |
"step": 615
|
| 755 |
},
|
| 756 |
{
|
| 757 |
"epoch": 1.36,
|
| 758 |
-
"learning_rate":
|
| 759 |
-
"loss": 2.
|
| 760 |
"step": 620
|
| 761 |
},
|
| 762 |
{
|
| 763 |
"epoch": 1.37,
|
| 764 |
-
"learning_rate":
|
| 765 |
-
"loss":
|
| 766 |
"step": 625
|
| 767 |
},
|
| 768 |
{
|
| 769 |
"epoch": 1.38,
|
| 770 |
-
"learning_rate":
|
| 771 |
-
"loss":
|
| 772 |
"step": 630
|
| 773 |
},
|
| 774 |
{
|
| 775 |
-
"epoch": 1.
|
| 776 |
-
"learning_rate":
|
| 777 |
-
"loss": 2.
|
| 778 |
"step": 635
|
| 779 |
},
|
| 780 |
{
|
| 781 |
-
"epoch": 1.
|
| 782 |
-
"learning_rate":
|
| 783 |
-
"loss":
|
| 784 |
"step": 640
|
| 785 |
},
|
| 786 |
{
|
| 787 |
-
"epoch": 1.
|
| 788 |
-
"learning_rate":
|
| 789 |
-
"loss": 2.
|
| 790 |
"step": 645
|
| 791 |
},
|
| 792 |
{
|
| 793 |
"epoch": 1.43,
|
| 794 |
-
"learning_rate":
|
| 795 |
-
"loss":
|
| 796 |
"step": 650
|
| 797 |
},
|
| 798 |
{
|
| 799 |
"epoch": 1.44,
|
| 800 |
-
"learning_rate":
|
| 801 |
-
"loss":
|
| 802 |
"step": 655
|
| 803 |
},
|
| 804 |
{
|
| 805 |
"epoch": 1.45,
|
| 806 |
-
"learning_rate":
|
| 807 |
-
"loss":
|
| 808 |
"step": 660
|
| 809 |
},
|
| 810 |
{
|
| 811 |
"epoch": 1.46,
|
| 812 |
-
"learning_rate":
|
| 813 |
-
"loss":
|
| 814 |
"step": 665
|
| 815 |
},
|
| 816 |
{
|
| 817 |
"epoch": 1.47,
|
| 818 |
-
"learning_rate": 6.
|
| 819 |
-
"loss": 2.
|
| 820 |
"step": 670
|
| 821 |
},
|
| 822 |
{
|
| 823 |
"epoch": 1.48,
|
| 824 |
-
"learning_rate": 6.
|
| 825 |
-
"loss": 2.
|
| 826 |
"step": 675
|
| 827 |
},
|
| 828 |
{
|
| 829 |
"epoch": 1.49,
|
| 830 |
-
"learning_rate":
|
| 831 |
-
"loss":
|
| 832 |
"step": 680
|
| 833 |
},
|
| 834 |
{
|
| 835 |
-
"epoch": 1.
|
| 836 |
-
"learning_rate":
|
| 837 |
-
"loss": 2.
|
| 838 |
"step": 685
|
| 839 |
},
|
| 840 |
{
|
| 841 |
-
"epoch": 1.
|
| 842 |
-
"learning_rate":
|
| 843 |
-
"loss":
|
| 844 |
"step": 690
|
| 845 |
},
|
| 846 |
{
|
| 847 |
-
"epoch": 1.
|
| 848 |
-
"learning_rate":
|
| 849 |
-
"loss":
|
| 850 |
"step": 695
|
| 851 |
},
|
| 852 |
{
|
| 853 |
"epoch": 1.54,
|
| 854 |
-
"learning_rate":
|
| 855 |
-
"loss":
|
| 856 |
"step": 700
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 1.55,
|
| 860 |
-
"learning_rate":
|
| 861 |
-
"loss":
|
| 862 |
"step": 705
|
| 863 |
},
|
| 864 |
{
|
| 865 |
"epoch": 1.56,
|
| 866 |
-
"learning_rate":
|
| 867 |
-
"loss":
|
| 868 |
"step": 710
|
| 869 |
},
|
| 870 |
{
|
| 871 |
"epoch": 1.57,
|
| 872 |
-
"learning_rate":
|
| 873 |
-
"loss":
|
| 874 |
"step": 715
|
| 875 |
},
|
| 876 |
{
|
| 877 |
"epoch": 1.58,
|
| 878 |
-
"learning_rate":
|
| 879 |
-
"loss": 2.
|
| 880 |
"step": 720
|
| 881 |
},
|
| 882 |
{
|
| 883 |
"epoch": 1.59,
|
| 884 |
-
"learning_rate":
|
| 885 |
-
"loss":
|
| 886 |
"step": 725
|
| 887 |
},
|
| 888 |
{
|
| 889 |
"epoch": 1.6,
|
| 890 |
-
"learning_rate":
|
| 891 |
-
"loss":
|
| 892 |
"step": 730
|
| 893 |
},
|
| 894 |
{
|
| 895 |
-
"epoch": 1.
|
| 896 |
-
"learning_rate":
|
| 897 |
-
"loss":
|
| 898 |
"step": 735
|
| 899 |
},
|
| 900 |
{
|
| 901 |
-
"epoch": 1.
|
| 902 |
-
"learning_rate":
|
| 903 |
-
"loss":
|
| 904 |
"step": 740
|
| 905 |
},
|
| 906 |
{
|
| 907 |
-
"epoch": 1.
|
| 908 |
-
"learning_rate":
|
| 909 |
-
"loss":
|
| 910 |
"step": 745
|
| 911 |
},
|
| 912 |
{
|
| 913 |
-
"epoch": 1.
|
| 914 |
-
"learning_rate":
|
| 915 |
-
"loss":
|
| 916 |
"step": 750
|
| 917 |
},
|
| 918 |
{
|
| 919 |
"epoch": 1.66,
|
| 920 |
-
"learning_rate":
|
| 921 |
-
"loss": 2.
|
| 922 |
"step": 755
|
| 923 |
},
|
| 924 |
{
|
| 925 |
"epoch": 1.67,
|
| 926 |
-
"learning_rate":
|
| 927 |
-
"loss":
|
| 928 |
"step": 760
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 1.68,
|
| 932 |
-
"learning_rate":
|
| 933 |
-
"loss": 2.
|
| 934 |
"step": 765
|
| 935 |
},
|
| 936 |
{
|
| 937 |
"epoch": 1.69,
|
| 938 |
-
"learning_rate":
|
| 939 |
-
"loss":
|
| 940 |
"step": 770
|
| 941 |
},
|
| 942 |
{
|
| 943 |
"epoch": 1.7,
|
| 944 |
-
"learning_rate":
|
| 945 |
-
"loss":
|
| 946 |
"step": 775
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 1.71,
|
| 950 |
-
"learning_rate":
|
| 951 |
-
"loss":
|
| 952 |
"step": 780
|
| 953 |
},
|
| 954 |
{
|
| 955 |
-
"epoch": 1.
|
| 956 |
-
"learning_rate":
|
| 957 |
-
"loss":
|
| 958 |
"step": 785
|
| 959 |
},
|
| 960 |
{
|
| 961 |
-
"epoch": 1.
|
| 962 |
-
"learning_rate":
|
| 963 |
-
"loss":
|
| 964 |
"step": 790
|
| 965 |
},
|
| 966 |
{
|
| 967 |
-
"epoch": 1.
|
| 968 |
-
"learning_rate":
|
| 969 |
-
"loss":
|
| 970 |
"step": 795
|
| 971 |
},
|
| 972 |
{
|
| 973 |
-
"epoch": 1.
|
| 974 |
-
"learning_rate":
|
| 975 |
-
"loss": 2.
|
| 976 |
"step": 800
|
| 977 |
},
|
| 978 |
{
|
| 979 |
"epoch": 1.77,
|
| 980 |
-
"learning_rate":
|
| 981 |
-
"loss":
|
| 982 |
"step": 805
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 1.78,
|
| 986 |
-
"learning_rate":
|
| 987 |
-
"loss":
|
| 988 |
"step": 810
|
| 989 |
},
|
| 990 |
{
|
| 991 |
"epoch": 1.79,
|
| 992 |
-
"learning_rate":
|
| 993 |
-
"loss":
|
| 994 |
"step": 815
|
| 995 |
},
|
| 996 |
{
|
| 997 |
"epoch": 1.8,
|
| 998 |
-
"learning_rate":
|
| 999 |
-
"loss":
|
| 1000 |
"step": 820
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 1.81,
|
| 1004 |
-
"learning_rate":
|
| 1005 |
-
"loss": 2.
|
| 1006 |
"step": 825
|
| 1007 |
},
|
| 1008 |
{
|
| 1009 |
"epoch": 1.82,
|
| 1010 |
-
"learning_rate":
|
| 1011 |
-
"loss":
|
| 1012 |
"step": 830
|
| 1013 |
},
|
| 1014 |
{
|
| 1015 |
-
"epoch": 1.
|
| 1016 |
-
"learning_rate":
|
| 1017 |
-
"loss": 2.
|
| 1018 |
"step": 835
|
| 1019 |
},
|
| 1020 |
{
|
| 1021 |
-
"epoch": 1.
|
| 1022 |
-
"learning_rate":
|
| 1023 |
-
"loss":
|
| 1024 |
"step": 840
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
-
"epoch": 1.
|
| 1028 |
-
"learning_rate":
|
| 1029 |
-
"loss":
|
| 1030 |
"step": 845
|
| 1031 |
},
|
| 1032 |
{
|
| 1033 |
-
"epoch": 1.
|
| 1034 |
-
"learning_rate":
|
| 1035 |
-
"loss":
|
| 1036 |
"step": 850
|
| 1037 |
},
|
| 1038 |
{
|
| 1039 |
"epoch": 1.88,
|
| 1040 |
-
"learning_rate":
|
| 1041 |
-
"loss": 2.
|
| 1042 |
"step": 855
|
| 1043 |
},
|
| 1044 |
{
|
| 1045 |
"epoch": 1.89,
|
| 1046 |
-
"learning_rate":
|
| 1047 |
-
"loss":
|
| 1048 |
"step": 860
|
| 1049 |
},
|
| 1050 |
{
|
| 1051 |
"epoch": 1.9,
|
| 1052 |
-
"learning_rate":
|
| 1053 |
-
"loss":
|
| 1054 |
"step": 865
|
| 1055 |
},
|
| 1056 |
{
|
| 1057 |
"epoch": 1.91,
|
| 1058 |
-
"learning_rate":
|
| 1059 |
-
"loss":
|
| 1060 |
"step": 870
|
| 1061 |
},
|
| 1062 |
{
|
| 1063 |
"epoch": 1.92,
|
| 1064 |
-
"learning_rate":
|
| 1065 |
-
"loss":
|
| 1066 |
"step": 875
|
| 1067 |
},
|
| 1068 |
{
|
| 1069 |
"epoch": 1.93,
|
| 1070 |
-
"learning_rate":
|
| 1071 |
-
"loss":
|
| 1072 |
"step": 880
|
| 1073 |
},
|
| 1074 |
{
|
| 1075 |
-
"epoch": 1.
|
| 1076 |
-
"learning_rate":
|
| 1077 |
-
"loss":
|
| 1078 |
"step": 885
|
| 1079 |
},
|
| 1080 |
{
|
| 1081 |
-
"epoch": 1.
|
| 1082 |
-
"learning_rate":
|
| 1083 |
-
"loss":
|
| 1084 |
"step": 890
|
| 1085 |
},
|
| 1086 |
{
|
| 1087 |
-
"epoch": 1.
|
| 1088 |
-
"learning_rate":
|
| 1089 |
-
"loss": 2.
|
| 1090 |
"step": 895
|
| 1091 |
},
|
| 1092 |
{
|
| 1093 |
-
"epoch": 1.
|
| 1094 |
-
"learning_rate":
|
| 1095 |
-
"loss": 2.
|
| 1096 |
"step": 900
|
| 1097 |
},
|
| 1098 |
{
|
| 1099 |
-
"epoch": 1.
|
| 1100 |
-
"learning_rate":
|
| 1101 |
-
"loss":
|
| 1102 |
"step": 905
|
| 1103 |
},
|
| 1104 |
{
|
| 1105 |
"epoch": 2.0,
|
| 1106 |
-
"learning_rate":
|
| 1107 |
-
"loss":
|
| 1108 |
"step": 910
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 2.0,
|
| 1112 |
-
"eval_loss": 2.
|
| 1113 |
-
"eval_runtime":
|
| 1114 |
-
"eval_samples_per_second":
|
| 1115 |
-
"eval_steps_per_second": 2.
|
| 1116 |
-
"step":
|
| 1117 |
-
},
|
| 1118 |
-
{
|
| 1119 |
-
"epoch": 2.02,
|
| 1120 |
-
"learning_rate": 0.00013518019096478567,
|
| 1121 |
-
"loss": 2.9623,
|
| 1122 |
-
"step": 915
|
| 1123 |
-
},
|
| 1124 |
-
{
|
| 1125 |
-
"epoch": 2.04,
|
| 1126 |
-
"learning_rate": 0.00013456586620543129,
|
| 1127 |
-
"loss": 2.8352,
|
| 1128 |
-
"step": 920
|
| 1129 |
-
},
|
| 1130 |
-
{
|
| 1131 |
-
"epoch": 2.05,
|
| 1132 |
-
"learning_rate": 0.000133871881786694,
|
| 1133 |
-
"loss": 3.0422,
|
| 1134 |
-
"step": 925
|
| 1135 |
-
},
|
| 1136 |
-
{
|
| 1137 |
-
"epoch": 2.06,
|
| 1138 |
-
"learning_rate": 0.00013309907575657638,
|
| 1139 |
-
"loss": 3.1173,
|
| 1140 |
-
"step": 930
|
| 1141 |
-
},
|
| 1142 |
-
{
|
| 1143 |
-
"epoch": 2.07,
|
| 1144 |
-
"learning_rate": 0.00013224838134719792,
|
| 1145 |
-
"loss": 2.8456,
|
| 1146 |
-
"step": 935
|
| 1147 |
-
},
|
| 1148 |
-
{
|
| 1149 |
-
"epoch": 2.08,
|
| 1150 |
-
"learning_rate": 0.0001313208258478339,
|
| 1151 |
-
"loss": 2.9064,
|
| 1152 |
-
"step": 940
|
| 1153 |
-
},
|
| 1154 |
-
{
|
| 1155 |
-
"epoch": 2.09,
|
| 1156 |
-
"learning_rate": 0.00013031752936437234,
|
| 1157 |
-
"loss": 3.0691,
|
| 1158 |
-
"step": 945
|
| 1159 |
-
},
|
| 1160 |
-
{
|
| 1161 |
-
"epoch": 2.1,
|
| 1162 |
-
"learning_rate": 0.0001292397034666859,
|
| 1163 |
-
"loss": 2.8686,
|
| 1164 |
-
"step": 950
|
| 1165 |
-
},
|
| 1166 |
-
{
|
| 1167 |
-
"epoch": 2.11,
|
| 1168 |
-
"learning_rate": 0.00012808864972555387,
|
| 1169 |
-
"loss": 3.1551,
|
| 1170 |
-
"step": 955
|
| 1171 |
-
},
|
| 1172 |
-
{
|
| 1173 |
-
"epoch": 2.12,
|
| 1174 |
-
"learning_rate": 0.00012686575814089955,
|
| 1175 |
-
"loss": 2.9915,
|
| 1176 |
-
"step": 960
|
| 1177 |
-
},
|
| 1178 |
-
{
|
| 1179 |
-
"epoch": 2.13,
|
| 1180 |
-
"learning_rate": 0.00012557250546324104,
|
| 1181 |
-
"loss": 2.9484,
|
| 1182 |
-
"step": 965
|
| 1183 |
-
},
|
| 1184 |
-
{
|
| 1185 |
-
"epoch": 2.15,
|
| 1186 |
-
"learning_rate": 0.00012421045341038344,
|
| 1187 |
-
"loss": 2.9243,
|
| 1188 |
-
"step": 970
|
| 1189 |
-
},
|
| 1190 |
-
{
|
| 1191 |
-
"epoch": 2.16,
|
| 1192 |
-
"learning_rate": 0.0001227812467815061,
|
| 1193 |
-
"loss": 3.0377,
|
| 1194 |
-
"step": 975
|
| 1195 |
-
},
|
| 1196 |
-
{
|
| 1197 |
-
"epoch": 2.17,
|
| 1198 |
-
"learning_rate": 0.00012128661147092065,
|
| 1199 |
-
"loss": 2.9919,
|
| 1200 |
-
"step": 980
|
| 1201 |
-
},
|
| 1202 |
-
{
|
| 1203 |
-
"epoch": 2.18,
|
| 1204 |
-
"learning_rate": 0.00011972835238389994,
|
| 1205 |
-
"loss": 2.9963,
|
| 1206 |
-
"step": 985
|
| 1207 |
-
},
|
| 1208 |
-
{
|
| 1209 |
-
"epoch": 2.19,
|
| 1210 |
-
"learning_rate": 0.00011810835125709417,
|
| 1211 |
-
"loss": 2.8212,
|
| 1212 |
-
"step": 990
|
| 1213 |
-
},
|
| 1214 |
-
{
|
| 1215 |
-
"epoch": 2.2,
|
| 1216 |
-
"learning_rate": 0.00011642856438616635,
|
| 1217 |
-
"loss": 2.8691,
|
| 1218 |
-
"step": 995
|
| 1219 |
-
},
|
| 1220 |
-
{
|
| 1221 |
-
"epoch": 2.21,
|
| 1222 |
-
"learning_rate": 0.00011469102026339105,
|
| 1223 |
-
"loss": 2.827,
|
| 1224 |
-
"step": 1000
|
| 1225 |
-
},
|
| 1226 |
-
{
|
| 1227 |
-
"epoch": 2.22,
|
| 1228 |
-
"learning_rate": 0.00011289781712806871,
|
| 1229 |
-
"loss": 3.0833,
|
| 1230 |
-
"step": 1005
|
| 1231 |
-
},
|
| 1232 |
-
{
|
| 1233 |
-
"epoch": 2.23,
|
| 1234 |
-
"learning_rate": 0.00011105112043271582,
|
| 1235 |
-
"loss": 3.1229,
|
| 1236 |
-
"step": 1010
|
| 1237 |
-
},
|
| 1238 |
-
{
|
| 1239 |
-
"epoch": 2.25,
|
| 1240 |
-
"learning_rate": 0.00010915316022808737,
|
| 1241 |
-
"loss": 2.762,
|
| 1242 |
-
"step": 1015
|
| 1243 |
-
},
|
| 1244 |
-
{
|
| 1245 |
-
"epoch": 2.26,
|
| 1246 |
-
"learning_rate": 0.00010720622847019211,
|
| 1247 |
-
"loss": 3.0876,
|
| 1248 |
-
"step": 1020
|
| 1249 |
-
},
|
| 1250 |
-
{
|
| 1251 |
-
"epoch": 2.27,
|
| 1252 |
-
"learning_rate": 0.00010521267625255156,
|
| 1253 |
-
"loss": 2.8371,
|
| 1254 |
-
"step": 1025
|
| 1255 |
-
},
|
| 1256 |
-
{
|
| 1257 |
-
"epoch": 2.28,
|
| 1258 |
-
"learning_rate": 0.0001031749109670451,
|
| 1259 |
-
"loss": 2.6792,
|
| 1260 |
-
"step": 1030
|
| 1261 |
-
},
|
| 1262 |
-
{
|
| 1263 |
-
"epoch": 2.29,
|
| 1264 |
-
"learning_rate": 0.00010109539339677007,
|
| 1265 |
-
"loss": 3.03,
|
| 1266 |
-
"step": 1035
|
| 1267 |
-
},
|
| 1268 |
-
{
|
| 1269 |
-
"epoch": 2.3,
|
| 1270 |
-
"learning_rate": 9.897663474442611e-05,
|
| 1271 |
-
"loss": 2.9339,
|
| 1272 |
-
"step": 1040
|
| 1273 |
-
},
|
| 1274 |
-
{
|
| 1275 |
-
"epoch": 2.31,
|
| 1276 |
-
"learning_rate": 9.682119359981534e-05,
|
| 1277 |
-
"loss": 3.0358,
|
| 1278 |
-
"step": 1045
|
| 1279 |
-
},
|
| 1280 |
-
{
|
| 1281 |
-
"epoch": 2.32,
|
| 1282 |
-
"learning_rate": 9.463167285011724e-05,
|
| 1283 |
-
"loss": 3.0776,
|
| 1284 |
-
"step": 1050
|
| 1285 |
-
},
|
| 1286 |
-
{
|
| 1287 |
-
"epoch": 2.33,
|
| 1288 |
-
"learning_rate": 9.241071653666963e-05,
|
| 1289 |
-
"loss": 3.0847,
|
| 1290 |
-
"step": 1055
|
| 1291 |
-
},
|
| 1292 |
-
{
|
| 1293 |
-
"epoch": 2.35,
|
| 1294 |
-
"learning_rate": 9.016100666205323e-05,
|
| 1295 |
-
"loss": 2.97,
|
| 1296 |
-
"step": 1060
|
| 1297 |
-
},
|
| 1298 |
-
{
|
| 1299 |
-
"epoch": 2.36,
|
| 1300 |
-
"learning_rate": 8.788525995133582e-05,
|
| 1301 |
-
"loss": 2.889,
|
| 1302 |
-
"step": 1065
|
| 1303 |
-
},
|
| 1304 |
-
{
|
| 1305 |
-
"epoch": 2.37,
|
| 1306 |
-
"learning_rate": 8.55862245713838e-05,
|
| 1307 |
-
"loss": 3.1711,
|
| 1308 |
-
"step": 1070
|
| 1309 |
-
},
|
| 1310 |
-
{
|
| 1311 |
-
"epoch": 2.38,
|
| 1312 |
-
"learning_rate": 8.326667681220608e-05,
|
| 1313 |
-
"loss": 2.8987,
|
| 1314 |
-
"step": 1075
|
| 1315 |
-
},
|
| 1316 |
-
{
|
| 1317 |
-
"epoch": 2.39,
|
| 1318 |
-
"learning_rate": 8.092941773433665e-05,
|
| 1319 |
-
"loss": 2.8936,
|
| 1320 |
-
"step": 1080
|
| 1321 |
-
},
|
| 1322 |
-
{
|
| 1323 |
-
"epoch": 2.4,
|
| 1324 |
-
"learning_rate": 7.857726978630442e-05,
|
| 1325 |
-
"loss": 2.8497,
|
| 1326 |
-
"step": 1085
|
| 1327 |
-
},
|
| 1328 |
-
{
|
| 1329 |
-
"epoch": 2.41,
|
| 1330 |
-
"learning_rate": 7.621307339627561e-05,
|
| 1331 |
-
"loss": 2.8509,
|
| 1332 |
-
"step": 1090
|
| 1333 |
-
},
|
| 1334 |
-
{
|
| 1335 |
-
"epoch": 2.42,
|
| 1336 |
-
"learning_rate": 7.383968354198253e-05,
|
| 1337 |
-
"loss": 3.0417,
|
| 1338 |
-
"step": 1095
|
| 1339 |
-
},
|
| 1340 |
-
{
|
| 1341 |
-
"epoch": 2.43,
|
| 1342 |
-
"learning_rate": 7.14599663030853e-05,
|
| 1343 |
-
"loss": 3.0553,
|
| 1344 |
-
"step": 1100
|
| 1345 |
-
},
|
| 1346 |
-
{
|
| 1347 |
-
"epoch": 2.44,
|
| 1348 |
-
"learning_rate": 6.907679540012426e-05,
|
| 1349 |
-
"loss": 3.0356,
|
| 1350 |
-
"step": 1105
|
| 1351 |
-
},
|
| 1352 |
-
{
|
| 1353 |
-
"epoch": 2.46,
|
| 1354 |
-
"learning_rate": 6.669304872424684e-05,
|
| 1355 |
-
"loss": 3.0422,
|
| 1356 |
-
"step": 1110
|
| 1357 |
-
},
|
| 1358 |
-
{
|
| 1359 |
-
"epoch": 2.47,
|
| 1360 |
-
"learning_rate": 6.431160486189743e-05,
|
| 1361 |
-
"loss": 2.9538,
|
| 1362 |
-
"step": 1115
|
| 1363 |
-
},
|
| 1364 |
-
{
|
| 1365 |
-
"epoch": 2.48,
|
| 1366 |
-
"learning_rate": 6.193533961866804e-05,
|
| 1367 |
-
"loss": 2.7657,
|
| 1368 |
-
"step": 1120
|
| 1369 |
-
},
|
| 1370 |
-
{
|
| 1371 |
-
"epoch": 2.49,
|
| 1372 |
-
"learning_rate": 5.95671225465068e-05,
|
| 1373 |
-
"loss": 2.9534,
|
| 1374 |
-
"step": 1125
|
| 1375 |
-
},
|
| 1376 |
-
{
|
| 1377 |
-
"epoch": 2.5,
|
| 1378 |
-
"learning_rate": 5.7209813478478596e-05,
|
| 1379 |
-
"loss": 2.8748,
|
| 1380 |
-
"step": 1130
|
| 1381 |
-
},
|
| 1382 |
-
{
|
| 1383 |
-
"epoch": 2.51,
|
| 1384 |
-
"learning_rate": 5.486625907526076e-05,
|
| 1385 |
-
"loss": 2.9115,
|
| 1386 |
-
"step": 1135
|
| 1387 |
-
},
|
| 1388 |
-
{
|
| 1389 |
-
"epoch": 2.52,
|
| 1390 |
-
"learning_rate": 5.2539289387549066e-05,
|
| 1391 |
-
"loss": 2.8815,
|
| 1392 |
-
"step": 1140
|
| 1393 |
-
},
|
| 1394 |
-
{
|
| 1395 |
-
"epoch": 2.53,
|
| 1396 |
-
"learning_rate": 5.023171443851731e-05,
|
| 1397 |
-
"loss": 2.8293,
|
| 1398 |
-
"step": 1145
|
| 1399 |
-
},
|
| 1400 |
-
{
|
| 1401 |
-
"epoch": 2.54,
|
| 1402 |
-
"learning_rate": 4.79463208304643e-05,
|
| 1403 |
-
"loss": 2.9858,
|
| 1404 |
-
"step": 1150
|
| 1405 |
-
},
|
| 1406 |
-
{
|
| 1407 |
-
"epoch": 2.56,
|
| 1408 |
-
"learning_rate": 4.5685868379744455e-05,
|
| 1409 |
-
"loss": 2.9581,
|
| 1410 |
-
"step": 1155
|
| 1411 |
-
},
|
| 1412 |
-
{
|
| 1413 |
-
"epoch": 2.57,
|
| 1414 |
-
"learning_rate": 4.3453086784042875e-05,
|
| 1415 |
-
"loss": 2.7478,
|
| 1416 |
-
"step": 1160
|
| 1417 |
-
},
|
| 1418 |
-
{
|
| 1419 |
-
"epoch": 2.58,
|
| 1420 |
-
"learning_rate": 4.125067232602242e-05,
|
| 1421 |
-
"loss": 2.8651,
|
| 1422 |
-
"step": 1165
|
| 1423 |
-
},
|
| 1424 |
-
{
|
| 1425 |
-
"epoch": 2.59,
|
| 1426 |
-
"learning_rate": 3.908128461732221e-05,
|
| 1427 |
-
"loss": 2.7991,
|
| 1428 |
-
"step": 1170
|
| 1429 |
-
},
|
| 1430 |
-
{
|
| 1431 |
-
"epoch": 2.6,
|
| 1432 |
-
"learning_rate": 3.6947543386839726e-05,
|
| 1433 |
-
"loss": 3.1513,
|
| 1434 |
-
"step": 1175
|
| 1435 |
-
},
|
| 1436 |
-
{
|
| 1437 |
-
"epoch": 2.61,
|
| 1438 |
-
"learning_rate": 3.485202531717497e-05,
|
| 1439 |
-
"loss": 3.0578,
|
| 1440 |
-
"step": 1180
|
| 1441 |
-
},
|
| 1442 |
-
{
|
| 1443 |
-
"epoch": 2.62,
|
| 1444 |
-
"learning_rate": 3.2797260933055854e-05,
|
| 1445 |
-
"loss": 3.0007,
|
| 1446 |
-
"step": 1185
|
| 1447 |
-
},
|
| 1448 |
-
{
|
| 1449 |
-
"epoch": 2.63,
|
| 1450 |
-
"learning_rate": 3.0785731545505505e-05,
|
| 1451 |
-
"loss": 3.0579,
|
| 1452 |
-
"step": 1190
|
| 1453 |
-
},
|
| 1454 |
-
{
|
| 1455 |
-
"epoch": 2.64,
|
| 1456 |
-
"learning_rate": 2.88198662554373e-05,
|
| 1457 |
-
"loss": 2.965,
|
| 1458 |
-
"step": 1195
|
| 1459 |
-
},
|
| 1460 |
-
{
|
| 1461 |
-
"epoch": 2.65,
|
| 1462 |
-
"learning_rate": 2.6902039020299345e-05,
|
| 1463 |
-
"loss": 2.8752,
|
| 1464 |
-
"step": 1200
|
| 1465 |
-
},
|
| 1466 |
-
{
|
| 1467 |
-
"epoch": 2.67,
|
| 1468 |
-
"learning_rate": 2.503456578730942e-05,
|
| 1469 |
-
"loss": 2.9304,
|
| 1470 |
-
"step": 1205
|
| 1471 |
-
},
|
| 1472 |
-
{
|
| 1473 |
-
"epoch": 2.68,
|
| 1474 |
-
"learning_rate": 2.3219701696742384e-05,
|
| 1475 |
-
"loss": 3.0212,
|
| 1476 |
-
"step": 1210
|
| 1477 |
-
},
|
| 1478 |
-
{
|
| 1479 |
-
"epoch": 2.69,
|
| 1480 |
-
"learning_rate": 2.1459638358647452e-05,
|
| 1481 |
-
"loss": 2.9375,
|
| 1482 |
-
"step": 1215
|
| 1483 |
-
},
|
| 1484 |
-
{
|
| 1485 |
-
"epoch": 2.7,
|
| 1486 |
-
"learning_rate": 1.9756501206284008e-05,
|
| 1487 |
-
"loss": 2.7038,
|
| 1488 |
-
"step": 1220
|
| 1489 |
-
},
|
| 1490 |
-
{
|
| 1491 |
-
"epoch": 2.71,
|
| 1492 |
-
"learning_rate": 1.8112346929470726e-05,
|
| 1493 |
-
"loss": 2.8753,
|
| 1494 |
-
"step": 1225
|
| 1495 |
-
},
|
| 1496 |
-
{
|
| 1497 |
-
"epoch": 2.72,
|
| 1498 |
-
"learning_rate": 1.652916099095118e-05,
|
| 1499 |
-
"loss": 3.1553,
|
| 1500 |
-
"step": 1230
|
| 1501 |
-
},
|
| 1502 |
-
{
|
| 1503 |
-
"epoch": 2.73,
|
| 1504 |
-
"learning_rate": 1.5008855228769123e-05,
|
| 1505 |
-
"loss": 3.0094,
|
| 1506 |
-
"step": 1235
|
| 1507 |
-
},
|
| 1508 |
-
{
|
| 1509 |
-
"epoch": 2.74,
|
| 1510 |
-
"learning_rate": 1.3553265547554012e-05,
|
| 1511 |
-
"loss": 3.0194,
|
| 1512 |
-
"step": 1240
|
| 1513 |
-
},
|
| 1514 |
-
{
|
| 1515 |
-
"epoch": 2.75,
|
| 1516 |
-
"learning_rate": 1.21641497015034e-05,
|
| 1517 |
-
"loss": 2.8742,
|
| 1518 |
-
"step": 1245
|
| 1519 |
-
},
|
| 1520 |
-
{
|
| 1521 |
-
"epoch": 2.77,
|
| 1522 |
-
"learning_rate": 1.084318517173752e-05,
|
| 1523 |
-
"loss": 2.9059,
|
| 1524 |
-
"step": 1250
|
| 1525 |
-
},
|
| 1526 |
-
{
|
| 1527 |
-
"epoch": 2.78,
|
| 1528 |
-
"learning_rate": 9.59196714059158e-06,
|
| 1529 |
-
"loss": 3.0479,
|
| 1530 |
-
"step": 1255
|
| 1531 |
-
},
|
| 1532 |
-
{
|
| 1533 |
-
"epoch": 2.79,
|
| 1534 |
-
"learning_rate": 8.41200656529103e-06,
|
| 1535 |
-
"loss": 3.0663,
|
| 1536 |
-
"step": 1260
|
| 1537 |
-
},
|
| 1538 |
-
{
|
| 1539 |
-
"epoch": 2.8,
|
| 1540 |
-
"learning_rate": 7.304728353336053e-06,
|
| 1541 |
-
"loss": 2.9088,
|
| 1542 |
-
"step": 1265
|
| 1543 |
-
},
|
| 1544 |
-
{
|
| 1545 |
-
"epoch": 2.81,
|
| 1546 |
-
"learning_rate": 6.271469641799014e-06,
|
| 1547 |
-
"loss": 3.0761,
|
| 1548 |
-
"step": 1270
|
| 1549 |
-
},
|
| 1550 |
-
{
|
| 1551 |
-
"epoch": 2.82,
|
| 1552 |
-
"learning_rate": 5.31347818261188e-06,
|
| 1553 |
-
"loss": 2.666,
|
| 1554 |
-
"step": 1275
|
| 1555 |
-
},
|
| 1556 |
-
{
|
| 1557 |
-
"epoch": 2.83,
|
| 1558 |
-
"learning_rate": 4.4319108357952465e-06,
|
| 1559 |
-
"loss": 2.7844,
|
| 1560 |
-
"step": 1280
|
| 1561 |
-
},
|
| 1562 |
-
{
|
| 1563 |
-
"epoch": 2.84,
|
| 1564 |
-
"learning_rate": 3.627832172446345e-06,
|
| 1565 |
-
"loss": 2.9375,
|
| 1566 |
-
"step": 1285
|
| 1567 |
-
},
|
| 1568 |
-
{
|
| 1569 |
-
"epoch": 2.85,
|
| 1570 |
-
"learning_rate": 2.9022131891750718e-06,
|
| 1571 |
-
"loss": 2.6596,
|
| 1572 |
-
"step": 1290
|
| 1573 |
-
},
|
| 1574 |
-
{
|
| 1575 |
-
"epoch": 2.87,
|
| 1576 |
-
"learning_rate": 2.2559301355383432e-06,
|
| 1577 |
-
"loss": 2.8171,
|
| 1578 |
-
"step": 1295
|
| 1579 |
-
},
|
| 1580 |
-
{
|
| 1581 |
-
"epoch": 2.88,
|
| 1582 |
-
"learning_rate": 1.689763455891467e-06,
|
| 1583 |
-
"loss": 3.0183,
|
| 1584 |
-
"step": 1300
|
| 1585 |
-
},
|
| 1586 |
-
{
|
| 1587 |
-
"epoch": 2.89,
|
| 1588 |
-
"learning_rate": 1.2043968469316554e-06,
|
| 1589 |
-
"loss": 2.7782,
|
| 1590 |
-
"step": 1305
|
| 1591 |
-
},
|
| 1592 |
-
{
|
| 1593 |
-
"epoch": 2.9,
|
| 1594 |
-
"learning_rate": 8.004164320735402e-07,
|
| 1595 |
-
"loss": 3.051,
|
| 1596 |
-
"step": 1310
|
| 1597 |
-
},
|
| 1598 |
-
{
|
| 1599 |
-
"epoch": 2.91,
|
| 1600 |
-
"learning_rate": 4.783100536526791e-07,
|
| 1601 |
-
"loss": 2.4571,
|
| 1602 |
-
"step": 1315
|
| 1603 |
-
},
|
| 1604 |
-
{
|
| 1605 |
-
"epoch": 2.92,
|
| 1606 |
-
"learning_rate": 2.384666838129012e-07,
|
| 1607 |
-
"loss": 2.8637,
|
| 1608 |
-
"step": 1320
|
| 1609 |
-
},
|
| 1610 |
-
{
|
| 1611 |
-
"epoch": 2.93,
|
| 1612 |
-
"learning_rate": 8.117595478742552e-08,
|
| 1613 |
-
"loss": 2.9795,
|
| 1614 |
-
"step": 1325
|
| 1615 |
-
},
|
| 1616 |
-
{
|
| 1617 |
-
"epoch": 2.94,
|
| 1618 |
-
"learning_rate": 6.62780914226615e-09,
|
| 1619 |
-
"loss": 2.9055,
|
| 1620 |
-
"step": 1330
|
| 1621 |
-
},
|
| 1622 |
-
{
|
| 1623 |
-
"epoch": 2.95,
|
| 1624 |
-
"learning_rate": 1.4912270403706995e-08,
|
| 1625 |
-
"loss": 2.7752,
|
| 1626 |
-
"step": 1335
|
| 1627 |
-
},
|
| 1628 |
-
{
|
| 1629 |
-
"epoch": 2.96,
|
| 1630 |
-
"learning_rate": 1.0601933434684325e-07,
|
| 1631 |
-
"loss": 2.6349,
|
| 1632 |
-
"step": 1340
|
| 1633 |
-
},
|
| 1634 |
-
{
|
| 1635 |
-
"epoch": 2.98,
|
| 1636 |
-
"learning_rate": 2.7983898107656093e-07,
|
| 1637 |
-
"loss": 2.8498,
|
| 1638 |
-
"step": 1345
|
| 1639 |
-
},
|
| 1640 |
-
{
|
| 1641 |
-
"epoch": 2.99,
|
| 1642 |
-
"learning_rate": 5.361613078863541e-07,
|
| 1643 |
-
"loss": 2.7323,
|
| 1644 |
-
"step": 1350
|
| 1645 |
-
},
|
| 1646 |
-
{
|
| 1647 |
-
"epoch": 3.0,
|
| 1648 |
-
"learning_rate": 8.746767827345921e-07,
|
| 1649 |
-
"loss": 2.728,
|
| 1650 |
-
"step": 1355
|
| 1651 |
-
},
|
| 1652 |
-
{
|
| 1653 |
-
"epoch": 3.0,
|
| 1654 |
-
"eval_loss": 2.8375096321105957,
|
| 1655 |
-
"eval_runtime": 31.5132,
|
| 1656 |
-
"eval_samples_per_second": 21.197,
|
| 1657 |
-
"eval_steps_per_second": 2.666,
|
| 1658 |
-
"step": 1356
|
| 1659 |
-
},
|
| 1660 |
-
{
|
| 1661 |
-
"epoch": 3.01,
|
| 1662 |
-
"learning_rate": 1.294976618031915e-06,
|
| 1663 |
-
"loss": 2.8229,
|
| 1664 |
-
"step": 1360
|
| 1665 |
-
},
|
| 1666 |
-
{
|
| 1667 |
-
"epoch": 3.02,
|
| 1668 |
-
"learning_rate": 1.7965532642888837e-06,
|
| 1669 |
-
"loss": 2.875,
|
| 1670 |
-
"step": 1365
|
| 1671 |
-
},
|
| 1672 |
-
{
|
| 1673 |
-
"epoch": 3.03,
|
| 1674 |
-
"learning_rate": 2.37880102302694e-06,
|
| 1675 |
-
"loss": 2.6679,
|
| 1676 |
-
"step": 1370
|
| 1677 |
-
},
|
| 1678 |
-
{
|
| 1679 |
-
"epoch": 3.04,
|
| 1680 |
-
"learning_rate": 3.0410167782134987e-06,
|
| 1681 |
-
"loss": 2.6421,
|
| 1682 |
-
"step": 1375
|
| 1683 |
-
},
|
| 1684 |
-
{
|
| 1685 |
-
"epoch": 3.05,
|
| 1686 |
-
"learning_rate": 3.7824008453367663e-06,
|
| 1687 |
-
"loss": 2.8491,
|
| 1688 |
-
"step": 1380
|
| 1689 |
-
},
|
| 1690 |
-
{
|
| 1691 |
-
"epoch": 3.06,
|
| 1692 |
-
"learning_rate": 4.602057937096792e-06,
|
| 1693 |
-
"loss": 2.5251,
|
| 1694 |
-
"step": 1385
|
| 1695 |
-
},
|
| 1696 |
-
{
|
| 1697 |
-
"epoch": 3.08,
|
| 1698 |
-
"learning_rate": 5.498998244543872e-06,
|
| 1699 |
-
"loss": 2.8046,
|
| 1700 |
-
"step": 1390
|
| 1701 |
-
},
|
| 1702 |
-
{
|
| 1703 |
-
"epoch": 3.09,
|
| 1704 |
-
"learning_rate": 6.472138632361396e-06,
|
| 1705 |
-
"loss": 2.8098,
|
| 1706 |
-
"step": 1395
|
| 1707 |
-
},
|
| 1708 |
-
{
|
| 1709 |
-
"epoch": 3.1,
|
| 1710 |
-
"learning_rate": 7.520303946848027e-06,
|
| 1711 |
-
"loss": 2.8847,
|
| 1712 |
-
"step": 1400
|
| 1713 |
-
},
|
| 1714 |
-
{
|
| 1715 |
-
"epoch": 3.11,
|
| 1716 |
-
"learning_rate": 8.642228435021056e-06,
|
| 1717 |
-
"loss": 2.9209,
|
| 1718 |
-
"step": 1405
|
| 1719 |
-
},
|
| 1720 |
-
{
|
| 1721 |
-
"epoch": 3.12,
|
| 1722 |
-
"learning_rate": 9.836557273124741e-06,
|
| 1723 |
-
"loss": 2.7372,
|
| 1724 |
-
"step": 1410
|
| 1725 |
-
},
|
| 1726 |
-
{
|
| 1727 |
-
"epoch": 3.13,
|
| 1728 |
-
"learning_rate": 1.1101848202701654e-05,
|
| 1729 |
-
"loss": 2.6496,
|
| 1730 |
-
"step": 1415
|
| 1731 |
-
},
|
| 1732 |
-
{
|
| 1733 |
-
"epoch": 3.14,
|
| 1734 |
-
"learning_rate": 1.243657327224822e-05,
|
| 1735 |
-
"loss": 2.9148,
|
| 1736 |
-
"step": 1420
|
| 1737 |
-
},
|
| 1738 |
-
{
|
| 1739 |
-
"epoch": 3.15,
|
| 1740 |
-
"learning_rate": 1.38391206823516e-05,
|
| 1741 |
-
"loss": 3.0211,
|
| 1742 |
-
"step": 1425
|
| 1743 |
-
},
|
| 1744 |
-
{
|
| 1745 |
-
"epoch": 3.16,
|
| 1746 |
-
"learning_rate": 1.5307796732081518e-05,
|
| 1747 |
-
"loss": 2.6395,
|
| 1748 |
-
"step": 1430
|
| 1749 |
-
},
|
| 1750 |
-
{
|
| 1751 |
-
"epoch": 3.17,
|
| 1752 |
-
"learning_rate": 1.6840827864284945e-05,
|
| 1753 |
-
"loss": 2.7042,
|
| 1754 |
-
"step": 1435
|
| 1755 |
-
},
|
| 1756 |
-
{
|
| 1757 |
-
"epoch": 3.19,
|
| 1758 |
-
"learning_rate": 1.8436362807314293e-05,
|
| 1759 |
-
"loss": 2.8092,
|
| 1760 |
-
"step": 1440
|
| 1761 |
-
},
|
| 1762 |
-
{
|
| 1763 |
-
"epoch": 3.2,
|
| 1764 |
-
"learning_rate": 2.0092474810602836e-05,
|
| 1765 |
-
"loss": 2.7596,
|
| 1766 |
-
"step": 1445
|
| 1767 |
-
},
|
| 1768 |
-
{
|
| 1769 |
-
"epoch": 3.21,
|
| 1770 |
-
"learning_rate": 2.180716397138768e-05,
|
| 1771 |
-
"loss": 2.7119,
|
| 1772 |
-
"step": 1450
|
| 1773 |
-
},
|
| 1774 |
-
{
|
| 1775 |
-
"epoch": 3.22,
|
| 1776 |
-
"learning_rate": 2.3578359649770255e-05,
|
| 1777 |
-
"loss": 2.7985,
|
| 1778 |
-
"step": 1455
|
| 1779 |
-
},
|
| 1780 |
-
{
|
| 1781 |
-
"epoch": 3.23,
|
| 1782 |
-
"learning_rate": 2.5403922969199407e-05,
|
| 1783 |
-
"loss": 2.9927,
|
| 1784 |
-
"step": 1460
|
| 1785 |
-
},
|
| 1786 |
-
{
|
| 1787 |
-
"epoch": 3.24,
|
| 1788 |
-
"learning_rate": 2.728164939935347e-05,
|
| 1789 |
-
"loss": 2.8457,
|
| 1790 |
-
"step": 1465
|
| 1791 |
-
},
|
| 1792 |
-
{
|
| 1793 |
-
"epoch": 3.25,
|
| 1794 |
-
"learning_rate": 2.9209271418308684e-05,
|
| 1795 |
-
"loss": 2.781,
|
| 1796 |
-
"step": 1470
|
| 1797 |
-
},
|
| 1798 |
-
{
|
| 1799 |
-
"epoch": 3.26,
|
| 1800 |
-
"learning_rate": 3.118446125077373e-05,
|
| 1801 |
-
"loss": 2.9769,
|
| 1802 |
-
"step": 1475
|
| 1803 |
-
},
|
| 1804 |
-
{
|
| 1805 |
-
"epoch": 3.27,
|
| 1806 |
-
"learning_rate": 3.320483367908478e-05,
|
| 1807 |
-
"loss": 2.8904,
|
| 1808 |
-
"step": 1480
|
| 1809 |
-
},
|
| 1810 |
-
{
|
| 1811 |
-
"epoch": 3.29,
|
| 1812 |
-
"learning_rate": 3.526794892356904e-05,
|
| 1813 |
-
"loss": 2.8189,
|
| 1814 |
-
"step": 1485
|
| 1815 |
-
},
|
| 1816 |
-
{
|
| 1817 |
-
"epoch": 3.3,
|
| 1818 |
-
"learning_rate": 3.737131558879585e-05,
|
| 1819 |
-
"loss": 2.9244,
|
| 1820 |
-
"step": 1490
|
| 1821 |
-
},
|
| 1822 |
-
{
|
| 1823 |
-
"epoch": 3.31,
|
| 1824 |
-
"learning_rate": 3.951239367215861e-05,
|
| 1825 |
-
"loss": 2.7045,
|
| 1826 |
-
"step": 1495
|
| 1827 |
-
},
|
| 1828 |
-
{
|
| 1829 |
-
"epoch": 3.32,
|
| 1830 |
-
"learning_rate": 4.168859763115413e-05,
|
| 1831 |
-
"loss": 2.9866,
|
| 1832 |
-
"step": 1500
|
| 1833 |
-
},
|
| 1834 |
-
{
|
| 1835 |
-
"epoch": 3.33,
|
| 1836 |
-
"learning_rate": 4.389729950565535e-05,
|
| 1837 |
-
"loss": 3.0244,
|
| 1838 |
-
"step": 1505
|
| 1839 |
-
},
|
| 1840 |
-
{
|
| 1841 |
-
"epoch": 3.34,
|
| 1842 |
-
"learning_rate": 4.613583209140809e-05,
|
| 1843 |
-
"loss": 2.901,
|
| 1844 |
-
"step": 1510
|
| 1845 |
-
},
|
| 1846 |
-
{
|
| 1847 |
-
"epoch": 3.35,
|
| 1848 |
-
"learning_rate": 4.8401492160916547e-05,
|
| 1849 |
-
"loss": 2.7344,
|
| 1850 |
-
"step": 1515
|
| 1851 |
-
},
|
| 1852 |
-
{
|
| 1853 |
-
"epoch": 3.36,
|
| 1854 |
-
"learning_rate": 5.0691543727832066e-05,
|
| 1855 |
-
"loss": 2.8102,
|
| 1856 |
-
"step": 1520
|
| 1857 |
-
},
|
| 1858 |
-
{
|
| 1859 |
-
"epoch": 3.37,
|
| 1860 |
-
"learning_rate": 5.3003221350900195e-05,
|
| 1861 |
-
"loss": 2.8373,
|
| 1862 |
-
"step": 1525
|
| 1863 |
-
},
|
| 1864 |
-
{
|
| 1865 |
-
"epoch": 3.38,
|
| 1866 |
-
"learning_rate": 5.533373347347734e-05,
|
| 1867 |
-
"loss": 2.7037,
|
| 1868 |
-
"step": 1530
|
| 1869 |
-
},
|
| 1870 |
-
{
|
| 1871 |
-
"epoch": 3.4,
|
| 1872 |
-
"learning_rate": 5.7680265794584045e-05,
|
| 1873 |
-
"loss": 2.8285,
|
| 1874 |
-
"step": 1535
|
| 1875 |
-
},
|
| 1876 |
-
{
|
| 1877 |
-
"epoch": 3.41,
|
| 1878 |
-
"learning_rate": 6.0039984667424135e-05,
|
| 1879 |
-
"loss": 2.8403,
|
| 1880 |
-
"step": 1540
|
| 1881 |
-
},
|
| 1882 |
-
{
|
| 1883 |
-
"epoch": 3.42,
|
| 1884 |
-
"learning_rate": 6.241004052126537e-05,
|
| 1885 |
-
"loss": 2.8711,
|
| 1886 |
-
"step": 1545
|
| 1887 |
-
},
|
| 1888 |
-
{
|
| 1889 |
-
"epoch": 3.43,
|
| 1890 |
-
"learning_rate": 6.478757130255114e-05,
|
| 1891 |
-
"loss": 2.9325,
|
| 1892 |
-
"step": 1550
|
| 1893 |
-
},
|
| 1894 |
-
{
|
| 1895 |
-
"epoch": 3.44,
|
| 1896 |
-
"learning_rate": 6.716970593108268e-05,
|
| 1897 |
-
"loss": 2.7888,
|
| 1898 |
-
"step": 1555
|
| 1899 |
-
},
|
| 1900 |
-
{
|
| 1901 |
-
"epoch": 3.45,
|
| 1902 |
-
"learning_rate": 6.95535677671063e-05,
|
| 1903 |
-
"loss": 2.8757,
|
| 1904 |
-
"step": 1560
|
| 1905 |
-
},
|
| 1906 |
-
{
|
| 1907 |
-
"epoch": 3.46,
|
| 1908 |
-
"learning_rate": 7.193627808511189e-05,
|
| 1909 |
-
"loss": 2.9177,
|
| 1910 |
-
"step": 1565
|
| 1911 |
-
},
|
| 1912 |
-
{
|
| 1913 |
-
"epoch": 3.47,
|
| 1914 |
-
"learning_rate": 7.431495955014918e-05,
|
| 1915 |
-
"loss": 2.5954,
|
| 1916 |
-
"step": 1570
|
| 1917 |
-
},
|
| 1918 |
-
{
|
| 1919 |
-
"epoch": 3.48,
|
| 1920 |
-
"learning_rate": 7.668673969246706e-05,
|
| 1921 |
-
"loss": 2.8505,
|
| 1922 |
-
"step": 1575
|
| 1923 |
-
},
|
| 1924 |
-
{
|
| 1925 |
-
"epoch": 3.5,
|
| 1926 |
-
"learning_rate": 7.90487543762762e-05,
|
| 1927 |
-
"loss": 2.8186,
|
| 1928 |
-
"step": 1580
|
| 1929 |
-
},
|
| 1930 |
-
{
|
| 1931 |
-
"epoch": 3.51,
|
| 1932 |
-
"learning_rate": 8.13981512584484e-05,
|
| 1933 |
-
"loss": 2.614,
|
| 1934 |
-
"step": 1585
|
| 1935 |
-
},
|
| 1936 |
-
{
|
| 1937 |
-
"epoch": 3.52,
|
| 1938 |
-
"learning_rate": 8.373209323297482e-05,
|
| 1939 |
-
"loss": 2.8168,
|
| 1940 |
-
"step": 1590
|
| 1941 |
-
},
|
| 1942 |
-
{
|
| 1943 |
-
"epoch": 3.53,
|
| 1944 |
-
"learning_rate": 8.604776185702404e-05,
|
| 1945 |
-
"loss": 2.9398,
|
| 1946 |
-
"step": 1595
|
| 1947 |
-
},
|
| 1948 |
-
{
|
| 1949 |
-
"epoch": 3.54,
|
| 1950 |
-
"learning_rate": 8.834236075446371e-05,
|
| 1951 |
-
"loss": 2.6622,
|
| 1952 |
-
"step": 1600
|
| 1953 |
-
},
|
| 1954 |
-
{
|
| 1955 |
-
"epoch": 3.55,
|
| 1956 |
-
"learning_rate": 9.061311899273253e-05,
|
| 1957 |
-
"loss": 2.9225,
|
| 1958 |
-
"step": 1605
|
| 1959 |
-
},
|
| 1960 |
-
{
|
| 1961 |
-
"epoch": 3.56,
|
| 1962 |
-
"learning_rate": 9.285729442898916e-05,
|
| 1963 |
-
"loss": 2.9019,
|
| 1964 |
-
"step": 1610
|
| 1965 |
-
},
|
| 1966 |
-
{
|
| 1967 |
-
"epoch": 3.57,
|
| 1968 |
-
"learning_rate": 9.507217702149268e-05,
|
| 1969 |
-
"loss": 2.924,
|
| 1970 |
-
"step": 1615
|
| 1971 |
-
},
|
| 1972 |
-
{
|
| 1973 |
-
"epoch": 3.58,
|
| 1974 |
-
"learning_rate": 9.725509210222166e-05,
|
| 1975 |
-
"loss": 2.8929,
|
| 1976 |
-
"step": 1620
|
| 1977 |
-
},
|
| 1978 |
-
{
|
| 1979 |
-
"epoch": 3.6,
|
| 1980 |
-
"learning_rate": 9.940340360677169e-05,
|
| 1981 |
-
"loss": 2.916,
|
| 1982 |
-
"step": 1625
|
| 1983 |
-
},
|
| 1984 |
-
{
|
| 1985 |
-
"epoch": 3.61,
|
| 1986 |
-
"learning_rate": 0.0001015145172576392,
|
| 1987 |
-
"loss": 2.995,
|
| 1988 |
-
"step": 1630
|
| 1989 |
-
},
|
| 1990 |
-
{
|
| 1991 |
-
"epoch": 3.62,
|
| 1992 |
-
"learning_rate": 0.00010358588369704218,
|
| 1993 |
-
"loss": 3.0979,
|
| 1994 |
-
"step": 1635
|
| 1995 |
-
},
|
| 1996 |
-
{
|
| 1997 |
-
"epoch": 3.63,
|
| 1998 |
-
"learning_rate": 0.00010561500156549773,
|
| 1999 |
-
"loss": 2.9044,
|
| 2000 |
-
"step": 1640
|
| 2001 |
-
},
|
| 2002 |
-
{
|
| 2003 |
-
"epoch": 3.64,
|
| 2004 |
-
"learning_rate": 0.00010759942052243444,
|
| 2005 |
-
"loss": 2.8281,
|
| 2006 |
-
"step": 1645
|
| 2007 |
-
},
|
| 2008 |
-
{
|
| 2009 |
-
"epoch": 3.65,
|
| 2010 |
-
"learning_rate": 0.00010953674420519884,
|
| 2011 |
-
"loss": 2.9313,
|
| 2012 |
-
"step": 1650
|
| 2013 |
-
},
|
| 2014 |
-
{
|
| 2015 |
-
"epoch": 3.66,
|
| 2016 |
-
"learning_rate": 0.00011142463312287635,
|
| 2017 |
-
"loss": 2.7283,
|
| 2018 |
-
"step": 1655
|
| 2019 |
-
},
|
| 2020 |
-
{
|
| 2021 |
-
"epoch": 3.67,
|
| 2022 |
-
"learning_rate": 0.00011326080748143409,
|
| 2023 |
-
"loss": 2.9097,
|
| 2024 |
-
"step": 1660
|
| 2025 |
-
},
|
| 2026 |
-
{
|
| 2027 |
-
"epoch": 3.68,
|
| 2028 |
-
"learning_rate": 0.00011504304993677608,
|
| 2029 |
-
"loss": 2.6674,
|
| 2030 |
-
"step": 1665
|
| 2031 |
-
},
|
| 2032 |
-
{
|
| 2033 |
-
"epoch": 3.69,
|
| 2034 |
-
"learning_rate": 0.00011676920827238338,
|
| 2035 |
-
"loss": 2.8365,
|
| 2036 |
-
"step": 1670
|
| 2037 |
-
},
|
| 2038 |
-
{
|
| 2039 |
-
"epoch": 3.71,
|
| 2040 |
-
"learning_rate": 0.00011843719799830742,
|
| 2041 |
-
"loss": 2.8295,
|
| 2042 |
-
"step": 1675
|
| 2043 |
-
},
|
| 2044 |
-
{
|
| 2045 |
-
"epoch": 3.72,
|
| 2046 |
-
"learning_rate": 0.0001200450048683772,
|
| 2047 |
-
"loss": 2.9448,
|
| 2048 |
-
"step": 1680
|
| 2049 |
-
},
|
| 2050 |
-
{
|
| 2051 |
-
"epoch": 3.73,
|
| 2052 |
-
"learning_rate": 0.00012159068731258081,
|
| 2053 |
-
"loss": 2.84,
|
| 2054 |
-
"step": 1685
|
| 2055 |
-
},
|
| 2056 |
-
{
|
| 2057 |
-
"epoch": 3.74,
|
| 2058 |
-
"learning_rate": 0.0001230723787816848,
|
| 2059 |
-
"loss": 3.0295,
|
| 2060 |
-
"step": 1690
|
| 2061 |
-
},
|
| 2062 |
-
{
|
| 2063 |
-
"epoch": 3.75,
|
| 2064 |
-
"learning_rate": 0.00012448829000125775,
|
| 2065 |
-
"loss": 2.8701,
|
| 2066 |
-
"step": 1695
|
| 2067 |
-
},
|
| 2068 |
-
{
|
| 2069 |
-
"epoch": 3.76,
|
| 2070 |
-
"learning_rate": 0.0001258367111323792,
|
| 2071 |
-
"loss": 3.0611,
|
| 2072 |
-
"step": 1700
|
| 2073 |
-
},
|
| 2074 |
-
{
|
| 2075 |
-
"epoch": 3.77,
|
| 2076 |
-
"learning_rate": 0.0001271160138364217,
|
| 2077 |
-
"loss": 2.9251,
|
| 2078 |
-
"step": 1705
|
| 2079 |
-
},
|
| 2080 |
-
{
|
| 2081 |
-
"epoch": 3.78,
|
| 2082 |
-
"learning_rate": 0.000128324653241416,
|
| 2083 |
-
"loss": 2.7988,
|
| 2084 |
-
"step": 1710
|
| 2085 |
-
},
|
| 2086 |
-
{
|
| 2087 |
-
"epoch": 3.79,
|
| 2088 |
-
"learning_rate": 0.00012946116980762042,
|
| 2089 |
-
"loss": 2.6729,
|
| 2090 |
-
"step": 1715
|
| 2091 |
-
},
|
| 2092 |
-
{
|
| 2093 |
-
"epoch": 3.81,
|
| 2094 |
-
"learning_rate": 0.00013052419109004545,
|
| 2095 |
-
"loss": 2.9239,
|
| 2096 |
-
"step": 1720
|
| 2097 |
-
},
|
| 2098 |
-
{
|
| 2099 |
-
"epoch": 3.82,
|
| 2100 |
-
"learning_rate": 0.00013151243339580335,
|
| 2101 |
-
"loss": 2.7932,
|
| 2102 |
-
"step": 1725
|
| 2103 |
-
},
|
| 2104 |
-
{
|
| 2105 |
-
"epoch": 3.83,
|
| 2106 |
-
"learning_rate": 0.0001324247033342821,
|
| 2107 |
-
"loss": 2.9398,
|
| 2108 |
-
"step": 1730
|
| 2109 |
-
},
|
| 2110 |
-
{
|
| 2111 |
-
"epoch": 3.84,
|
| 2112 |
-
"learning_rate": 0.00013325989925826984,
|
| 2113 |
-
"loss": 2.8449,
|
| 2114 |
-
"step": 1735
|
| 2115 |
-
},
|
| 2116 |
-
{
|
| 2117 |
-
"epoch": 3.85,
|
| 2118 |
-
"learning_rate": 0.0001340170125942929,
|
| 2119 |
-
"loss": 2.8704,
|
| 2120 |
-
"step": 1740
|
| 2121 |
-
},
|
| 2122 |
-
{
|
| 2123 |
-
"epoch": 3.86,
|
| 2124 |
-
"learning_rate": 0.000134695129060558,
|
| 2125 |
-
"loss": 2.9972,
|
| 2126 |
-
"step": 1745
|
| 2127 |
-
},
|
| 2128 |
-
{
|
| 2129 |
-
"epoch": 3.87,
|
| 2130 |
-
"learning_rate": 0.00013529342977102845,
|
| 2131 |
-
"loss": 2.8523,
|
| 2132 |
-
"step": 1750
|
| 2133 |
-
},
|
| 2134 |
-
{
|
| 2135 |
-
"epoch": 3.88,
|
| 2136 |
-
"learning_rate": 0.00013581119222430253,
|
| 2137 |
-
"loss": 2.6293,
|
| 2138 |
-
"step": 1755
|
| 2139 |
-
},
|
| 2140 |
-
{
|
| 2141 |
-
"epoch": 3.89,
|
| 2142 |
-
"learning_rate": 0.00013624779117609787,
|
| 2143 |
-
"loss": 2.8462,
|
| 2144 |
-
"step": 1760
|
| 2145 |
-
},
|
| 2146 |
-
{
|
| 2147 |
-
"epoch": 3.9,
|
| 2148 |
-
"learning_rate": 0.00013660269939429003,
|
| 2149 |
-
"loss": 3.1364,
|
| 2150 |
-
"step": 1765
|
| 2151 |
-
},
|
| 2152 |
-
{
|
| 2153 |
-
"epoch": 3.92,
|
| 2154 |
-
"learning_rate": 0.00013687548829559187,
|
| 2155 |
-
"loss": 2.8377,
|
| 2156 |
-
"step": 1770
|
| 2157 |
-
},
|
| 2158 |
-
{
|
| 2159 |
-
"epoch": 3.93,
|
| 2160 |
-
"learning_rate": 0.00013706582846310632,
|
| 2161 |
-
"loss": 2.5993,
|
| 2162 |
-
"step": 1775
|
| 2163 |
-
},
|
| 2164 |
-
{
|
| 2165 |
-
"epoch": 3.94,
|
| 2166 |
-
"learning_rate": 0.0001371734900441264,
|
| 2167 |
-
"loss": 2.7971,
|
| 2168 |
-
"step": 1780
|
| 2169 |
-
},
|
| 2170 |
-
{
|
| 2171 |
-
"epoch": 3.95,
|
| 2172 |
-
"learning_rate": 0.00013719834302770308,
|
| 2173 |
-
"loss": 3.0011,
|
| 2174 |
-
"step": 1785
|
| 2175 |
-
},
|
| 2176 |
-
{
|
| 2177 |
-
"epoch": 3.96,
|
| 2178 |
-
"learning_rate": 0.00013714035740164488,
|
| 2179 |
-
"loss": 2.5719,
|
| 2180 |
-
"step": 1790
|
| 2181 |
-
},
|
| 2182 |
-
{
|
| 2183 |
-
"epoch": 3.97,
|
| 2184 |
-
"learning_rate": 0.0001369996031887604,
|
| 2185 |
-
"loss": 2.8305,
|
| 2186 |
-
"step": 1795
|
| 2187 |
-
},
|
| 2188 |
-
{
|
| 2189 |
-
"epoch": 3.98,
|
| 2190 |
-
"learning_rate": 0.0001367762503622994,
|
| 2191 |
-
"loss": 3.0508,
|
| 2192 |
-
"step": 1800
|
| 2193 |
-
},
|
| 2194 |
-
{
|
| 2195 |
-
"epoch": 3.99,
|
| 2196 |
-
"learning_rate": 0.00013647056864069505,
|
| 2197 |
-
"loss": 2.8931,
|
| 2198 |
-
"step": 1805
|
| 2199 |
-
},
|
| 2200 |
-
{
|
| 2201 |
-
"epoch": 4.0,
|
| 2202 |
-
"eval_loss": 2.668595790863037,
|
| 2203 |
-
"eval_runtime": 32.3526,
|
| 2204 |
-
"eval_samples_per_second": 20.617,
|
| 2205 |
-
"eval_steps_per_second": 2.596,
|
| 2206 |
-
"step": 1808
|
| 2207 |
-
},
|
| 2208 |
-
{
|
| 2209 |
-
"epoch": 4.01,
|
| 2210 |
-
"learning_rate": 0.0001352850051421873,
|
| 2211 |
-
"loss": 3.0347,
|
| 2212 |
-
"step": 1810
|
| 2213 |
-
},
|
| 2214 |
-
{
|
| 2215 |
-
"epoch": 4.02,
|
| 2216 |
-
"learning_rate": 0.00013468407714712873,
|
| 2217 |
-
"loss": 2.8731,
|
| 2218 |
-
"step": 1815
|
| 2219 |
-
},
|
| 2220 |
-
{
|
| 2221 |
-
"epoch": 4.04,
|
| 2222 |
-
"learning_rate": 0.00013400299249496825,
|
| 2223 |
-
"loss": 2.6963,
|
| 2224 |
-
"step": 1820
|
| 2225 |
-
},
|
| 2226 |
-
{
|
| 2227 |
-
"epoch": 4.05,
|
| 2228 |
-
"learning_rate": 0.00013324257730708462,
|
| 2229 |
-
"loss": 2.7722,
|
| 2230 |
-
"step": 1825
|
| 2231 |
-
},
|
| 2232 |
-
{
|
| 2233 |
-
"epoch": 4.06,
|
| 2234 |
-
"learning_rate": 0.00013240375392880089,
|
| 2235 |
-
"loss": 2.7651,
|
| 2236 |
-
"step": 1830
|
| 2237 |
-
},
|
| 2238 |
-
{
|
| 2239 |
-
"epoch": 4.07,
|
| 2240 |
-
"learning_rate": 0.00013148753981062527,
|
| 2241 |
-
"loss": 2.7735,
|
| 2242 |
-
"step": 1835
|
| 2243 |
-
},
|
| 2244 |
-
{
|
| 2245 |
-
"epoch": 4.08,
|
| 2246 |
-
"learning_rate": 0.00013049504627413603,
|
| 2247 |
-
"loss": 2.9064,
|
| 2248 |
-
"step": 1840
|
| 2249 |
-
},
|
| 2250 |
-
{
|
| 2251 |
-
"epoch": 4.09,
|
| 2252 |
-
"learning_rate": 0.0001294274771640039,
|
| 2253 |
-
"loss": 2.7531,
|
| 2254 |
-
"step": 1845
|
| 2255 |
-
},
|
| 2256 |
-
{
|
| 2257 |
-
"epoch": 4.1,
|
| 2258 |
-
"learning_rate": 0.0001282861273877885,
|
| 2259 |
-
"loss": 2.5696,
|
| 2260 |
-
"step": 1850
|
| 2261 |
-
},
|
| 2262 |
-
{
|
| 2263 |
-
"epoch": 4.11,
|
| 2264 |
-
"learning_rate": 0.00012707238134528176,
|
| 2265 |
-
"loss": 2.9475,
|
| 2266 |
-
"step": 1855
|
| 2267 |
-
},
|
| 2268 |
-
{
|
| 2269 |
-
"epoch": 4.12,
|
| 2270 |
-
"learning_rate": 0.00012578771124930035,
|
| 2271 |
-
"loss": 2.6126,
|
| 2272 |
-
"step": 1860
|
| 2273 |
-
},
|
| 2274 |
-
{
|
| 2275 |
-
"epoch": 4.14,
|
| 2276 |
-
"learning_rate": 0.000124433675339965,
|
| 2277 |
-
"loss": 2.7006,
|
| 2278 |
-
"step": 1865
|
| 2279 |
-
},
|
| 2280 |
-
{
|
| 2281 |
-
"epoch": 4.15,
|
| 2282 |
-
"learning_rate": 0.00012301191599463465,
|
| 2283 |
-
"loss": 2.7261,
|
| 2284 |
-
"step": 1870
|
| 2285 |
-
},
|
| 2286 |
-
{
|
| 2287 |
-
"epoch": 4.16,
|
| 2288 |
-
"learning_rate": 0.00012152415773578541,
|
| 2289 |
-
"loss": 2.9439,
|
| 2290 |
-
"step": 1875
|
| 2291 |
-
},
|
| 2292 |
-
{
|
| 2293 |
-
"epoch": 4.17,
|
| 2294 |
-
"learning_rate": 0.0001199722051392502,
|
| 2295 |
-
"loss": 2.8357,
|
| 2296 |
-
"step": 1880
|
| 2297 |
-
},
|
| 2298 |
-
{
|
| 2299 |
-
"epoch": 4.18,
|
| 2300 |
-
"learning_rate": 0.00011835794064536121,
|
| 2301 |
-
"loss": 2.8717,
|
| 2302 |
-
"step": 1885
|
| 2303 |
-
},
|
| 2304 |
-
{
|
| 2305 |
-
"epoch": 4.19,
|
| 2306 |
-
"learning_rate": 0.00011668332227564373,
|
| 2307 |
-
"loss": 2.8137,
|
| 2308 |
-
"step": 1890
|
| 2309 |
-
},
|
| 2310 |
-
{
|
| 2311 |
-
"epoch": 4.2,
|
| 2312 |
-
"learning_rate": 0.00011495038125783349,
|
| 2313 |
-
"loss": 2.5525,
|
| 2314 |
-
"step": 1895
|
| 2315 |
-
},
|
| 2316 |
-
{
|
| 2317 |
-
"epoch": 4.21,
|
| 2318 |
-
"learning_rate": 0.00011316121956210187,
|
| 2319 |
-
"loss": 2.7496,
|
| 2320 |
-
"step": 1900
|
| 2321 |
-
},
|
| 2322 |
-
{
|
| 2323 |
-
"epoch": 4.22,
|
| 2324 |
-
"learning_rate": 0.00011131800735147087,
|
| 2325 |
-
"loss": 2.5725,
|
| 2326 |
-
"step": 1905
|
| 2327 |
-
},
|
| 2328 |
-
{
|
| 2329 |
-
"epoch": 4.24,
|
| 2330 |
-
"learning_rate": 0.00010942298034951362,
|
| 2331 |
-
"loss": 3.0666,
|
| 2332 |
-
"step": 1910
|
| 2333 |
-
},
|
| 2334 |
-
{
|
| 2335 |
-
"epoch": 4.25,
|
| 2336 |
-
"learning_rate": 0.00010747843712853636,
|
| 2337 |
-
"loss": 2.7999,
|
| 2338 |
-
"step": 1915
|
| 2339 |
-
},
|
| 2340 |
-
{
|
| 2341 |
-
"epoch": 4.26,
|
| 2342 |
-
"learning_rate": 0.000105486736321525,
|
| 2343 |
-
"loss": 2.7631,
|
| 2344 |
-
"step": 1920
|
| 2345 |
-
},
|
| 2346 |
-
{
|
| 2347 |
-
"epoch": 4.27,
|
| 2348 |
-
"learning_rate": 0.00010345029376124139,
|
| 2349 |
-
"loss": 2.7083,
|
| 2350 |
-
"step": 1925
|
| 2351 |
-
},
|
| 2352 |
-
{
|
| 2353 |
-
"epoch": 4.28,
|
| 2354 |
-
"learning_rate": 0.00010137157954994116,
|
| 2355 |
-
"loss": 2.6497,
|
| 2356 |
-
"step": 1930
|
| 2357 |
-
},
|
| 2358 |
-
{
|
| 2359 |
-
"epoch": 4.29,
|
| 2360 |
-
"learning_rate": 9.92531150632634e-05,
|
| 2361 |
-
"loss": 2.6362,
|
| 2362 |
-
"step": 1935
|
| 2363 |
-
},
|
| 2364 |
-
{
|
| 2365 |
-
"epoch": 4.3,
|
| 2366 |
-
"learning_rate": 9.709746989192951e-05,
|
| 2367 |
-
"loss": 2.8674,
|
| 2368 |
-
"step": 1940
|
| 2369 |
-
},
|
| 2370 |
-
{
|
| 2371 |
-
"epoch": 4.31,
|
| 2372 |
-
"learning_rate": 9.49072587249574e-05,
|
| 2373 |
-
"loss": 2.9979,
|
| 2374 |
-
"step": 1945
|
| 2375 |
-
},
|
| 2376 |
-
{
|
| 2377 |
-
"epoch": 4.32,
|
| 2378 |
-
"learning_rate": 9.268513817817794e-05,
|
| 2379 |
-
"loss": 2.5657,
|
| 2380 |
-
"step": 1950
|
| 2381 |
-
},
|
| 2382 |
-
{
|
| 2383 |
-
"epoch": 4.33,
|
| 2384 |
-
"learning_rate": 9.043380357189261e-05,
|
| 2385 |
-
"loss": 2.603,
|
| 2386 |
-
"step": 1955
|
| 2387 |
-
},
|
| 2388 |
-
{
|
| 2389 |
-
"epoch": 4.35,
|
| 2390 |
-
"learning_rate": 8.815598566158422e-05,
|
| 2391 |
-
"loss": 2.8067,
|
| 2392 |
-
"step": 1960
|
| 2393 |
-
},
|
| 2394 |
-
{
|
| 2395 |
-
"epoch": 4.36,
|
| 2396 |
-
"learning_rate": 8.585444732565087e-05,
|
| 2397 |
-
"loss": 2.6835,
|
| 2398 |
-
"step": 1965
|
| 2399 |
-
},
|
| 2400 |
-
{
|
| 2401 |
-
"epoch": 4.37,
|
| 2402 |
-
"learning_rate": 8.353198021417222e-05,
|
| 2403 |
-
"loss": 2.7185,
|
| 2404 |
-
"step": 1970
|
| 2405 |
-
},
|
| 2406 |
-
{
|
| 2407 |
-
"epoch": 4.38,
|
| 2408 |
-
"learning_rate": 8.119140136277707e-05,
|
| 2409 |
-
"loss": 2.7122,
|
| 2410 |
-
"step": 1975
|
| 2411 |
-
},
|
| 2412 |
-
{
|
| 2413 |
-
"epoch": 4.39,
|
| 2414 |
-
"learning_rate": 7.883554977572397e-05,
|
| 2415 |
-
"loss": 2.7984,
|
| 2416 |
-
"step": 1980
|
| 2417 |
-
},
|
| 2418 |
-
{
|
| 2419 |
-
"epoch": 4.4,
|
| 2420 |
-
"learning_rate": 7.646728298233036e-05,
|
| 2421 |
-
"loss": 2.8378,
|
| 2422 |
-
"step": 1985
|
| 2423 |
-
},
|
| 2424 |
-
{
|
| 2425 |
-
"epoch": 4.41,
|
| 2426 |
-
"learning_rate": 7.408947357093288e-05,
|
| 2427 |
-
"loss": 2.7695,
|
| 2428 |
-
"step": 1990
|
| 2429 |
-
},
|
| 2430 |
-
{
|
| 2431 |
-
"epoch": 4.42,
|
| 2432 |
-
"learning_rate": 7.170500570458389e-05,
|
| 2433 |
-
"loss": 2.6919,
|
| 2434 |
-
"step": 1995
|
| 2435 |
-
},
|
| 2436 |
-
{
|
| 2437 |
-
"epoch": 4.43,
|
| 2438 |
-
"learning_rate": 6.931677162270657e-05,
|
| 2439 |
-
"loss": 2.9963,
|
| 2440 |
-
"step": 2000
|
| 2441 |
-
},
|
| 2442 |
-
{
|
| 2443 |
-
"epoch": 4.45,
|
| 2444 |
-
"learning_rate": 6.692766813295225e-05,
|
| 2445 |
-
"loss": 2.6891,
|
| 2446 |
-
"step": 2005
|
| 2447 |
-
},
|
| 2448 |
-
{
|
| 2449 |
-
"epoch": 4.46,
|
| 2450 |
-
"learning_rate": 6.454059309752181e-05,
|
| 2451 |
-
"loss": 2.6475,
|
| 2452 |
-
"step": 2010
|
| 2453 |
-
},
|
| 2454 |
-
{
|
| 2455 |
-
"epoch": 4.47,
|
| 2456 |
-
"learning_rate": 6.215844191820312e-05,
|
| 2457 |
-
"loss": 2.558,
|
| 2458 |
-
"step": 2015
|
| 2459 |
-
},
|
| 2460 |
-
{
|
| 2461 |
-
"epoch": 4.48,
|
| 2462 |
-
"learning_rate": 5.978410402439316e-05,
|
| 2463 |
-
"loss": 2.7152,
|
| 2464 |
-
"step": 2020
|
| 2465 |
-
},
|
| 2466 |
-
{
|
| 2467 |
-
"epoch": 4.49,
|
| 2468 |
-
"learning_rate": 5.742045936836858e-05,
|
| 2469 |
-
"loss": 2.9234,
|
| 2470 |
-
"step": 2025
|
| 2471 |
-
},
|
| 2472 |
-
{
|
| 2473 |
-
"epoch": 4.5,
|
| 2474 |
-
"learning_rate": 5.5070374932047194e-05,
|
| 2475 |
-
"loss": 2.6887,
|
| 2476 |
-
"step": 2030
|
| 2477 |
-
},
|
| 2478 |
-
{
|
| 2479 |
-
"epoch": 4.51,
|
| 2480 |
-
"learning_rate": 5.273670124948182e-05,
|
| 2481 |
-
"loss": 2.8225,
|
| 2482 |
-
"step": 2035
|
| 2483 |
-
},
|
| 2484 |
-
{
|
| 2485 |
-
"epoch": 4.52,
|
| 2486 |
-
"learning_rate": 5.042226894930904e-05,
|
| 2487 |
-
"loss": 2.6082,
|
| 2488 |
-
"step": 2040
|
| 2489 |
-
},
|
| 2490 |
-
{
|
| 2491 |
-
"epoch": 4.53,
|
| 2492 |
-
"learning_rate": 4.812988532133743e-05,
|
| 2493 |
-
"loss": 2.6766,
|
| 2494 |
-
"step": 2045
|
| 2495 |
-
},
|
| 2496 |
-
{
|
| 2497 |
-
"epoch": 4.55,
|
| 2498 |
-
"learning_rate": 4.5862330911446034e-05,
|
| 2499 |
-
"loss": 2.6366,
|
| 2500 |
-
"step": 2050
|
| 2501 |
-
},
|
| 2502 |
-
{
|
| 2503 |
-
"epoch": 4.56,
|
| 2504 |
-
"learning_rate": 4.3622356148923677e-05,
|
| 2505 |
-
"loss": 2.7639,
|
| 2506 |
-
"step": 2055
|
| 2507 |
-
},
|
| 2508 |
-
{
|
| 2509 |
-
"epoch": 4.57,
|
| 2510 |
-
"learning_rate": 4.141267801033621e-05,
|
| 2511 |
-
"loss": 2.6465,
|
| 2512 |
-
"step": 2060
|
| 2513 |
-
},
|
| 2514 |
-
{
|
| 2515 |
-
"epoch": 4.58,
|
| 2516 |
-
"learning_rate": 3.9235976723969275e-05,
|
| 2517 |
-
"loss": 2.8069,
|
| 2518 |
-
"step": 2065
|
| 2519 |
-
},
|
| 2520 |
-
{
|
| 2521 |
-
"epoch": 4.59,
|
| 2522 |
-
"learning_rate": 3.70948925188491e-05,
|
| 2523 |
-
"loss": 2.7196,
|
| 2524 |
-
"step": 2070
|
| 2525 |
-
},
|
| 2526 |
-
{
|
| 2527 |
-
"epoch": 4.6,
|
| 2528 |
-
"learning_rate": 3.4992022422276234e-05,
|
| 2529 |
-
"loss": 2.8907,
|
| 2530 |
-
"step": 2075
|
| 2531 |
-
},
|
| 2532 |
-
{
|
| 2533 |
-
"epoch": 4.61,
|
| 2534 |
-
"learning_rate": 3.292991710976105e-05,
|
| 2535 |
-
"loss": 2.7258,
|
| 2536 |
-
"step": 2080
|
| 2537 |
-
},
|
| 2538 |
-
{
|
| 2539 |
-
"epoch": 4.62,
|
| 2540 |
-
"learning_rate": 3.0911077811185354e-05,
|
| 2541 |
-
"loss": 2.8221,
|
| 2542 |
-
"step": 2085
|
| 2543 |
-
},
|
| 2544 |
-
{
|
| 2545 |
-
"epoch": 4.63,
|
| 2546 |
-
"learning_rate": 2.8937953276935187e-05,
|
| 2547 |
-
"loss": 3.0361,
|
| 2548 |
-
"step": 2090
|
| 2549 |
-
},
|
| 2550 |
-
{
|
| 2551 |
-
"epoch": 4.65,
|
| 2552 |
-
"learning_rate": 2.7012936807688804e-05,
|
| 2553 |
-
"loss": 2.7882,
|
| 2554 |
-
"step": 2095
|
| 2555 |
-
},
|
| 2556 |
-
{
|
| 2557 |
-
"epoch": 4.66,
|
| 2558 |
-
"learning_rate": 2.5138363351465895e-05,
|
| 2559 |
-
"loss": 2.6547,
|
| 2560 |
-
"step": 2100
|
| 2561 |
-
},
|
| 2562 |
-
{
|
| 2563 |
-
"epoch": 4.67,
|
| 2564 |
-
"learning_rate": 2.3316506671451792e-05,
|
| 2565 |
-
"loss": 2.6357,
|
| 2566 |
-
"step": 2105
|
| 2567 |
-
},
|
| 2568 |
-
{
|
| 2569 |
-
"epoch": 4.68,
|
| 2570 |
-
"learning_rate": 2.1549576588036973e-05,
|
| 2571 |
-
"loss": 2.5398,
|
| 2572 |
-
"step": 2110
|
| 2573 |
-
},
|
| 2574 |
-
{
|
| 2575 |
-
"epoch": 4.69,
|
| 2576 |
-
"learning_rate": 1.9839716298417252e-05,
|
| 2577 |
-
"loss": 2.6769,
|
| 2578 |
-
"step": 2115
|
| 2579 |
-
},
|
| 2580 |
-
{
|
| 2581 |
-
"epoch": 4.7,
|
| 2582 |
-
"learning_rate": 1.8188999777003054e-05,
|
| 2583 |
-
"loss": 2.7592,
|
| 2584 |
-
"step": 2120
|
| 2585 |
-
},
|
| 2586 |
-
{
|
| 2587 |
-
"epoch": 4.71,
|
| 2588 |
-
"learning_rate": 1.6599429259792494e-05,
|
| 2589 |
-
"loss": 2.7318,
|
| 2590 |
-
"step": 2125
|
| 2591 |
-
},
|
| 2592 |
-
{
|
| 2593 |
-
"epoch": 4.72,
|
| 2594 |
-
"learning_rate": 1.5072932815760403e-05,
|
| 2595 |
-
"loss": 2.6568,
|
| 2596 |
-
"step": 2130
|
| 2597 |
-
},
|
| 2598 |
-
{
|
| 2599 |
-
"epoch": 4.73,
|
| 2600 |
-
"learning_rate": 1.3611362008208023e-05,
|
| 2601 |
-
"loss": 2.5581,
|
| 2602 |
-
"step": 2135
|
| 2603 |
-
},
|
| 2604 |
-
{
|
| 2605 |
-
"epoch": 4.75,
|
| 2606 |
-
"learning_rate": 1.2216489648909795e-05,
|
| 2607 |
-
"loss": 2.9316,
|
| 2608 |
-
"step": 2140
|
| 2609 |
-
},
|
| 2610 |
-
{
|
| 2611 |
-
"epoch": 4.76,
|
| 2612 |
-
"learning_rate": 1.0890007647781e-05,
|
| 2613 |
-
"loss": 2.7289,
|
| 2614 |
-
"step": 2145
|
| 2615 |
-
},
|
| 2616 |
-
{
|
| 2617 |
-
"epoch": 4.77,
|
| 2618 |
-
"learning_rate": 9.63352496067796e-06,
|
| 2619 |
-
"loss": 2.7816,
|
| 2620 |
-
"step": 2150
|
| 2621 |
-
},
|
| 2622 |
-
{
|
| 2623 |
-
"epoch": 4.78,
|
| 2624 |
-
"learning_rate": 8.448565637814693e-06,
|
| 2625 |
-
"loss": 2.7996,
|
| 2626 |
-
"step": 2155
|
| 2627 |
-
},
|
| 2628 |
-
{
|
| 2629 |
-
"epoch": 4.79,
|
| 2630 |
-
"learning_rate": 7.33656697516614e-06,
|
| 2631 |
-
"loss": 2.9528,
|
| 2632 |
-
"step": 2160
|
| 2633 |
-
},
|
| 2634 |
-
{
|
| 2635 |
-
"epoch": 4.8,
|
| 2636 |
-
"learning_rate": 6.298877771101947e-06,
|
| 2637 |
-
"loss": 2.7164,
|
| 2638 |
-
"step": 2165
|
| 2639 |
-
},
|
| 2640 |
-
{
|
| 2641 |
-
"epoch": 4.81,
|
| 2642 |
-
"learning_rate": 5.336756690361118e-06,
|
| 2643 |
-
"loss": 2.5886,
|
| 2644 |
-
"step": 2170
|
| 2645 |
-
},
|
| 2646 |
-
{
|
| 2647 |
-
"epoch": 4.82,
|
| 2648 |
-
"learning_rate": 4.451370737354938e-06,
|
| 2649 |
-
"loss": 2.8503,
|
| 2650 |
-
"step": 2175
|
| 2651 |
-
},
|
| 2652 |
-
{
|
| 2653 |
-
"epoch": 4.83,
|
| 2654 |
-
"learning_rate": 3.6437938406500714e-06,
|
| 2655 |
-
"loss": 2.8792,
|
| 2656 |
-
"step": 2180
|
| 2657 |
-
},
|
| 2658 |
-
{
|
| 2659 |
-
"epoch": 4.84,
|
| 2660 |
-
"learning_rate": 2.9150055503466386e-06,
|
| 2661 |
-
"loss": 2.7267,
|
| 2662 |
-
"step": 2185
|
| 2663 |
-
},
|
| 2664 |
-
{
|
| 2665 |
-
"epoch": 4.86,
|
| 2666 |
-
"learning_rate": 2.2658898499336053e-06,
|
| 2667 |
-
"loss": 2.6496,
|
| 2668 |
-
"step": 2190
|
| 2669 |
-
},
|
| 2670 |
-
{
|
| 2671 |
-
"epoch": 4.87,
|
| 2672 |
-
"learning_rate": 1.6972340840614438e-06,
|
| 2673 |
-
"loss": 2.6994,
|
| 2674 |
-
"step": 2195
|
| 2675 |
-
},
|
| 2676 |
-
{
|
| 2677 |
-
"epoch": 4.88,
|
| 2678 |
-
"learning_rate": 1.209728003532575e-06,
|
| 2679 |
-
"loss": 2.825,
|
| 2680 |
-
"step": 2200
|
| 2681 |
-
},
|
| 2682 |
-
{
|
| 2683 |
-
"epoch": 4.89,
|
| 2684 |
-
"learning_rate": 8.039629286681063e-07,
|
| 2685 |
-
"loss": 2.7805,
|
| 2686 |
-
"step": 2205
|
| 2687 |
-
},
|
| 2688 |
-
{
|
| 2689 |
-
"epoch": 4.9,
|
| 2690 |
-
"learning_rate": 4.804310320666616e-07,
|
| 2691 |
-
"loss": 2.5982,
|
| 2692 |
-
"step": 2210
|
| 2693 |
-
},
|
| 2694 |
-
{
|
| 2695 |
-
"epoch": 4.91,
|
| 2696 |
-
"learning_rate": 2.3952474162342595e-07,
|
| 2697 |
-
"loss": 2.7046,
|
| 2698 |
-
"step": 2215
|
| 2699 |
-
},
|
| 2700 |
-
{
|
| 2701 |
-
"epoch": 4.92,
|
| 2702 |
-
"learning_rate": 8.153626453480383e-08,
|
| 2703 |
-
"loss": 2.8443,
|
| 2704 |
-
"step": 2220
|
| 2705 |
-
},
|
| 2706 |
-
{
|
| 2707 |
-
"epoch": 4.93,
|
| 2708 |
-
"learning_rate": 6.657232866107043e-09,
|
| 2709 |
-
"loss": 2.7825,
|
| 2710 |
-
"step": 2225
|
| 2711 |
-
},
|
| 2712 |
-
{
|
| 2713 |
-
"epoch": 4.94,
|
| 2714 |
-
"learning_rate": 1.4978471111282743e-08,
|
| 2715 |
-
"loss": 2.6962,
|
| 2716 |
-
"step": 2230
|
| 2717 |
-
},
|
| 2718 |
-
{
|
| 2719 |
-
"epoch": 4.96,
|
| 2720 |
-
"learning_rate": 1.0648988602746714e-07,
|
| 2721 |
-
"loss": 2.6028,
|
| 2722 |
-
"step": 2235
|
| 2723 |
-
},
|
| 2724 |
-
{
|
| 2725 |
-
"epoch": 4.97,
|
| 2726 |
-
"learning_rate": 2.810804788775565e-07,
|
| 2727 |
-
"loss": 2.8954,
|
| 2728 |
-
"step": 2240
|
| 2729 |
-
},
|
| 2730 |
-
{
|
| 2731 |
-
"epoch": 4.98,
|
| 2732 |
-
"learning_rate": 5.385384800660121e-07,
|
| 2733 |
-
"loss": 2.8426,
|
| 2734 |
-
"step": 2245
|
| 2735 |
-
},
|
| 2736 |
-
{
|
| 2737 |
-
"epoch": 4.99,
|
| 2738 |
-
"learning_rate": 8.785516060049755e-07,
|
| 2739 |
-
"loss": 2.6331,
|
| 2740 |
-
"step": 2250
|
| 2741 |
-
},
|
| 2742 |
-
{
|
| 2743 |
-
"epoch": 5.0,
|
| 2744 |
-
"learning_rate": 1.3007074378981807e-06,
|
| 2745 |
-
"loss": 2.63,
|
| 2746 |
-
"step": 2255
|
| 2747 |
-
},
|
| 2748 |
-
{
|
| 2749 |
-
"epoch": 5.0,
|
| 2750 |
-
"eval_loss": 2.480816602706909,
|
| 2751 |
-
"eval_runtime": 32.1241,
|
| 2752 |
-
"eval_samples_per_second": 21.012,
|
| 2753 |
-
"eval_steps_per_second": 2.646,
|
| 2754 |
-
"step": 2255
|
| 2755 |
-
},
|
| 2756 |
-
{
|
| 2757 |
-
"epoch": 4.99,
|
| 2758 |
-
"learning_rate": 3.708414773111306e-07,
|
| 2759 |
-
"loss": 2.5263,
|
| 2760 |
-
"step": 2260
|
| 2761 |
-
},
|
| 2762 |
-
{
|
| 2763 |
-
"epoch": 5.0,
|
| 2764 |
-
"learning_rate": 6.58811375157305e-07,
|
| 2765 |
-
"loss": 2.8915,
|
| 2766 |
-
"step": 2265
|
| 2767 |
-
},
|
| 2768 |
-
{
|
| 2769 |
-
"epoch": 5.0,
|
| 2770 |
-
"eval_loss": 2.342907667160034,
|
| 2771 |
-
"eval_runtime": 31.416,
|
| 2772 |
-
"eval_samples_per_second": 21.008,
|
| 2773 |
-
"eval_steps_per_second": 2.642,
|
| 2774 |
-
"step": 2265
|
| 2775 |
}
|
| 2776 |
],
|
| 2777 |
-
"max_steps":
|
| 2778 |
"num_train_epochs": 6,
|
| 2779 |
-
"total_flos":
|
| 2780 |
"trial_name": null,
|
| 2781 |
"trial_params": null
|
| 2782 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 2.4912426471710205,
|
| 3 |
+
"best_model_checkpoint": "output/eminem/checkpoint-910",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 910,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 4.7203068149785564e-07,
|
| 13 |
+
"loss": 2.6483,
|
| 14 |
"step": 5
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 7.899165231781454e-07,
|
| 19 |
+
"loss": 2.495,
|
| 20 |
"step": 10
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.03,
|
| 24 |
+
"learning_rate": 1.1886129228032032e-06,
|
| 25 |
+
"loss": 2.7138,
|
| 26 |
"step": 15
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.04,
|
| 30 |
+
"learning_rate": 1.6676447463390023e-06,
|
| 31 |
+
"loss": 2.5829,
|
| 32 |
"step": 20
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.05,
|
| 36 |
+
"learning_rate": 2.2264411225088625e-06,
|
| 37 |
+
"loss": 2.7096,
|
| 38 |
"step": 25
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 2.8643361231111207e-06,
|
| 43 |
+
"loss": 2.4466,
|
| 44 |
"step": 30
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.08,
|
| 48 |
+
"learning_rate": 3.580569556618851e-06,
|
| 49 |
+
"loss": 2.7,
|
| 50 |
"step": 35
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.09,
|
| 54 |
+
"learning_rate": 4.3742878741139976e-06,
|
| 55 |
+
"loss": 2.718,
|
| 56 |
"step": 40
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.1,
|
| 60 |
+
"learning_rate": 5.244545186477934e-06,
|
| 61 |
+
"loss": 2.6071,
|
| 62 |
"step": 45
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.11,
|
| 66 |
+
"learning_rate": 6.190304391625697e-06,
|
| 67 |
+
"loss": 2.5909,
|
| 68 |
"step": 50
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.12,
|
| 72 |
+
"learning_rate": 7.210438410439488e-06,
|
| 73 |
+
"loss": 2.5627,
|
| 74 |
"step": 55
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.13,
|
| 78 |
+
"learning_rate": 8.303731529930944e-06,
|
| 79 |
+
"loss": 2.4577,
|
| 80 |
"step": 60
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.14,
|
| 84 |
+
"learning_rate": 9.468880852028573e-06,
|
| 85 |
+
"loss": 2.5309,
|
| 86 |
"step": 65
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.15,
|
| 90 |
+
"learning_rate": 1.0704497846266275e-05,
|
| 91 |
+
"loss": 2.7378,
|
| 92 |
"step": 70
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.16,
|
| 96 |
+
"learning_rate": 1.200911000452175e-05,
|
| 97 |
+
"loss": 2.6599,
|
| 98 |
"step": 75
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.18,
|
| 102 |
+
"learning_rate": 1.338116259583136e-05,
|
| 103 |
+
"loss": 2.3352,
|
| 104 |
"step": 80
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.19,
|
| 108 |
+
"learning_rate": 1.481902051919347e-05,
|
| 109 |
+
"loss": 2.7494,
|
| 110 |
"step": 85
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.2,
|
| 114 |
+
"learning_rate": 1.6320970252148078e-05,
|
| 115 |
+
"loss": 2.395,
|
| 116 |
"step": 90
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.21,
|
| 120 |
+
"learning_rate": 1.7885221892815092e-05,
|
| 121 |
+
"loss": 2.5425,
|
| 122 |
"step": 95
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.22,
|
| 126 |
+
"learning_rate": 1.950991129295317e-05,
|
| 127 |
+
"loss": 2.6284,
|
| 128 |
"step": 100
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.23,
|
| 132 |
+
"learning_rate": 2.11931022795012e-05,
|
| 133 |
+
"loss": 2.4998,
|
| 134 |
"step": 105
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.24,
|
| 138 |
+
"learning_rate": 2.2932788961951275e-05,
|
| 139 |
+
"loss": 2.5979,
|
| 140 |
"step": 110
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.25,
|
| 144 |
+
"learning_rate": 2.4726898122808088e-05,
|
| 145 |
+
"loss": 2.7143,
|
| 146 |
"step": 115
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.26,
|
| 150 |
+
"learning_rate": 2.6573291688279658e-05,
|
| 151 |
+
"loss": 2.4124,
|
| 152 |
"step": 120
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.27,
|
| 156 |
+
"learning_rate": 2.8469769276260912e-05,
|
| 157 |
+
"loss": 2.6573,
|
| 158 |
"step": 125
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 0.29,
|
| 162 |
+
"learning_rate": 3.0414070818571297e-05,
|
| 163 |
+
"loss": 2.6071,
|
| 164 |
"step": 130
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 0.3,
|
| 168 |
+
"learning_rate": 3.24038792543193e-05,
|
| 169 |
+
"loss": 2.8709,
|
| 170 |
"step": 135
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.31,
|
| 174 |
+
"learning_rate": 3.4436823291189305e-05,
|
| 175 |
+
"loss": 2.6786,
|
| 176 |
"step": 140
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.32,
|
| 180 |
+
"learning_rate": 3.651048023135391e-05,
|
| 181 |
+
"loss": 2.6154,
|
| 182 |
"step": 145
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.33,
|
| 186 |
+
"learning_rate": 3.8622378858650226e-05,
|
| 187 |
+
"loss": 2.6546,
|
| 188 |
"step": 150
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.34,
|
| 192 |
+
"learning_rate": 4.077000238357357e-05,
|
| 193 |
+
"loss": 2.6476,
|
| 194 |
"step": 155
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.35,
|
| 198 |
+
"learning_rate": 4.295079144258355e-05,
|
| 199 |
+
"loss": 2.5445,
|
| 200 |
"step": 160
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.36,
|
| 204 |
+
"learning_rate": 4.516214714814676e-05,
|
| 205 |
+
"loss": 2.4619,
|
| 206 |
"step": 165
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.37,
|
| 210 |
+
"learning_rate": 4.740143418587853e-05,
|
| 211 |
+
"loss": 2.5966,
|
| 212 |
"step": 170
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.38,
|
| 216 |
+
"learning_rate": 4.9665983955098734e-05,
|
| 217 |
+
"loss": 2.4876,
|
| 218 |
"step": 175
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 0.4,
|
| 222 |
+
"learning_rate": 5.1953097749051775e-05,
|
| 223 |
+
"loss": 2.6759,
|
| 224 |
"step": 180
|
| 225 |
},
|
| 226 |
{
|
| 227 |
+
"epoch": 0.41,
|
| 228 |
+
"learning_rate": 5.426004997100832e-05,
|
| 229 |
+
"loss": 2.4991,
|
| 230 |
"step": 185
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 0.42,
|
| 234 |
+
"learning_rate": 5.658409138240877e-05,
|
| 235 |
+
"loss": 2.5614,
|
| 236 |
"step": 190
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.43,
|
| 240 |
+
"learning_rate": 5.892245237918353e-05,
|
| 241 |
+
"loss": 2.6733,
|
| 242 |
"step": 195
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.44,
|
| 246 |
+
"learning_rate": 6.127234629234317e-05,
|
| 247 |
+
"loss": 2.7397,
|
| 248 |
"step": 200
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.45,
|
| 252 |
+
"learning_rate": 6.363097270890295e-05,
|
| 253 |
+
"loss": 2.7079,
|
| 254 |
"step": 205
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.46,
|
| 258 |
+
"learning_rate": 6.599552080918974e-05,
|
| 259 |
+
"loss": 2.5227,
|
| 260 |
"step": 210
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 0.47,
|
| 264 |
+
"learning_rate": 6.836317271654733e-05,
|
| 265 |
+
"loss": 2.6247,
|
| 266 |
"step": 215
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 0.48,
|
| 270 |
+
"learning_rate": 7.073110685545387e-05,
|
| 271 |
+
"loss": 2.5805,
|
| 272 |
"step": 220
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.49,
|
| 276 |
+
"learning_rate": 7.309650131404767e-05,
|
| 277 |
+
"loss": 2.7595,
|
| 278 |
"step": 225
|
| 279 |
},
|
| 280 |
{
|
| 281 |
+
"epoch": 0.51,
|
| 282 |
+
"learning_rate": 7.545653720705101e-05,
|
| 283 |
+
"loss": 2.5378,
|
| 284 |
"step": 230
|
| 285 |
},
|
| 286 |
{
|
| 287 |
+
"epoch": 0.52,
|
| 288 |
+
"learning_rate": 7.780840203509109e-05,
|
| 289 |
+
"loss": 2.6695,
|
| 290 |
"step": 235
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 0.53,
|
| 294 |
+
"learning_rate": 8.014929303640662e-05,
|
| 295 |
+
"loss": 2.6016,
|
| 296 |
"step": 240
|
| 297 |
},
|
| 298 |
{
|
| 299 |
+
"epoch": 0.54,
|
| 300 |
+
"learning_rate": 8.2476420526954e-05,
|
| 301 |
+
"loss": 2.8115,
|
| 302 |
"step": 245
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.55,
|
| 306 |
+
"learning_rate": 8.478701122492476e-05,
|
| 307 |
+
"loss": 2.8291,
|
| 308 |
"step": 250
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.56,
|
| 312 |
+
"learning_rate": 8.707831155571868e-05,
|
| 313 |
+
"loss": 2.4987,
|
| 314 |
"step": 255
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 0.57,
|
| 318 |
+
"learning_rate": 8.934759093343136e-05,
|
| 319 |
+
"loss": 2.5696,
|
| 320 |
"step": 260
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 0.58,
|
| 324 |
+
"learning_rate": 9.15921450149434e-05,
|
| 325 |
+
"loss": 2.6025,
|
| 326 |
"step": 265
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.59,
|
| 330 |
+
"learning_rate": 9.380929892273904e-05,
|
| 331 |
+
"loss": 2.6397,
|
| 332 |
"step": 270
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 0.6,
|
| 336 |
+
"learning_rate": 9.599641043260666e-05,
|
| 337 |
+
"loss": 2.5424,
|
| 338 |
"step": 275
|
| 339 |
},
|
| 340 |
{
|
| 341 |
+
"epoch": 0.62,
|
| 342 |
+
"learning_rate": 9.815087312242802e-05,
|
| 343 |
+
"loss": 2.7226,
|
| 344 |
"step": 280
|
| 345 |
},
|
| 346 |
{
|
| 347 |
+
"epoch": 0.63,
|
| 348 |
+
"learning_rate": 0.00010027011947830137,
|
| 349 |
+
"loss": 2.6259,
|
| 350 |
"step": 285
|
| 351 |
},
|
| 352 |
{
|
| 353 |
+
"epoch": 0.64,
|
| 354 |
+
"learning_rate": 0.00010235162395429476,
|
| 355 |
+
"loss": 2.6201,
|
| 356 |
"step": 290
|
| 357 |
},
|
| 358 |
{
|
| 359 |
+
"epoch": 0.65,
|
| 360 |
+
"learning_rate": 0.0001043929059821886,
|
| 361 |
+
"loss": 2.7705,
|
| 362 |
"step": 295
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 0.66,
|
| 366 |
+
"learning_rate": 0.00010639153292761396,
|
| 367 |
+
"loss": 2.7333,
|
| 368 |
"step": 300
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 0.67,
|
| 372 |
+
"learning_rate": 0.00010834512298907064,
|
| 373 |
+
"loss": 2.6623,
|
| 374 |
"step": 305
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 0.68,
|
| 378 |
+
"learning_rate": 0.00011025134803636347,
|
| 379 |
+
"loss": 2.8,
|
| 380 |
"step": 310
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.69,
|
| 384 |
+
"learning_rate": 0.00011210793638507988,
|
| 385 |
+
"loss": 2.6572,
|
| 386 |
"step": 315
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 0.7,
|
| 390 |
+
"learning_rate": 0.00011391267550379957,
|
| 391 |
+
"loss": 2.8413,
|
| 392 |
"step": 320
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 0.71,
|
| 396 |
+
"learning_rate": 0.00011566341465080861,
|
| 397 |
+
"loss": 2.6584,
|
| 398 |
"step": 325
|
| 399 |
},
|
| 400 |
{
|
| 401 |
+
"epoch": 0.73,
|
| 402 |
+
"learning_rate": 0.00011735806743718085,
|
| 403 |
+
"loss": 2.743,
|
| 404 |
"step": 330
|
| 405 |
},
|
| 406 |
{
|
| 407 |
+
"epoch": 0.74,
|
| 408 |
+
"learning_rate": 0.00011899461431316495,
|
| 409 |
+
"loss": 2.6499,
|
| 410 |
"step": 335
|
| 411 |
},
|
| 412 |
{
|
| 413 |
+
"epoch": 0.75,
|
| 414 |
+
"learning_rate": 0.00012057110497492171,
|
| 415 |
+
"loss": 2.6371,
|
| 416 |
"step": 340
|
| 417 |
},
|
| 418 |
{
|
| 419 |
+
"epoch": 0.76,
|
| 420 |
+
"learning_rate": 0.00012208566068873709,
|
| 421 |
+
"loss": 2.7359,
|
| 422 |
"step": 345
|
| 423 |
},
|
| 424 |
{
|
| 425 |
+
"epoch": 0.77,
|
| 426 |
+
"learning_rate": 0.00012353647652994535,
|
| 427 |
+
"loss": 2.5658,
|
| 428 |
"step": 350
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 0.78,
|
| 432 |
+
"learning_rate": 0.00012492182353389096,
|
| 433 |
+
"loss": 2.7984,
|
| 434 |
"step": 355
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 0.79,
|
| 438 |
+
"learning_rate": 0.0001262400507563695,
|
| 439 |
+
"loss": 2.6983,
|
| 440 |
"step": 360
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 0.8,
|
| 444 |
+
"learning_rate": 0.0001274895872410881,
|
| 445 |
+
"loss": 2.7025,
|
| 446 |
"step": 365
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 0.81,
|
| 450 |
+
"learning_rate": 0.00012866894389180423,
|
| 451 |
+
"loss": 2.6162,
|
| 452 |
"step": 370
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 0.82,
|
| 456 |
+
"learning_rate": 0.00012977671524691021,
|
| 457 |
+
"loss": 2.7679,
|
| 458 |
"step": 375
|
| 459 |
},
|
| 460 |
{
|
| 461 |
+
"epoch": 0.84,
|
| 462 |
+
"learning_rate": 0.0001308115811543474,
|
| 463 |
+
"loss": 2.408,
|
| 464 |
"step": 380
|
| 465 |
},
|
| 466 |
{
|
| 467 |
+
"epoch": 0.85,
|
| 468 |
+
"learning_rate": 0.00013177230834485735,
|
| 469 |
+
"loss": 2.7832,
|
| 470 |
"step": 385
|
| 471 |
},
|
| 472 |
{
|
| 473 |
+
"epoch": 0.86,
|
| 474 |
+
"learning_rate": 0.00013265775190169125,
|
| 475 |
+
"loss": 2.7911,
|
| 476 |
"step": 390
|
| 477 |
},
|
| 478 |
{
|
| 479 |
+
"epoch": 0.87,
|
| 480 |
+
"learning_rate": 0.00013346685662502964,
|
| 481 |
+
"loss": 2.7389,
|
| 482 |
"step": 395
|
| 483 |
},
|
| 484 |
{
|
| 485 |
+
"epoch": 0.88,
|
| 486 |
+
"learning_rate": 0.00013419865828948326,
|
| 487 |
+
"loss": 2.6844,
|
| 488 |
"step": 400
|
| 489 |
},
|
| 490 |
{
|
| 491 |
+
"epoch": 0.89,
|
| 492 |
+
"learning_rate": 0.00013485228479317903,
|
| 493 |
+
"loss": 2.8377,
|
| 494 |
"step": 405
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 0.9,
|
| 498 |
+
"learning_rate": 0.00013542695719706048,
|
| 499 |
+
"loss": 2.4116,
|
| 500 |
"step": 410
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 0.91,
|
| 504 |
+
"learning_rate": 0.00013592199065316375,
|
| 505 |
+
"loss": 2.5933,
|
| 506 |
"step": 415
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.92,
|
| 510 |
+
"learning_rate": 0.00013633679522076416,
|
| 511 |
+
"loss": 2.7723,
|
| 512 |
"step": 420
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 0.93,
|
| 516 |
+
"learning_rate": 0.0001366708765694196,
|
| 517 |
+
"loss": 2.7647,
|
| 518 |
"step": 425
|
| 519 |
},
|
| 520 |
{
|
| 521 |
+
"epoch": 0.95,
|
| 522 |
+
"learning_rate": 0.0001369238365680734,
|
| 523 |
+
"loss": 2.6368,
|
| 524 |
"step": 430
|
| 525 |
},
|
| 526 |
{
|
| 527 |
+
"epoch": 0.96,
|
| 528 |
+
"learning_rate": 0.0001370953737595148,
|
| 529 |
+
"loss": 2.3958,
|
| 530 |
"step": 435
|
| 531 |
},
|
| 532 |
{
|
| 533 |
+
"epoch": 0.97,
|
| 534 |
+
"learning_rate": 0.00013718528371963105,
|
| 535 |
+
"loss": 2.7726,
|
| 536 |
"step": 440
|
| 537 |
},
|
| 538 |
{
|
| 539 |
+
"epoch": 0.98,
|
| 540 |
+
"learning_rate": 0.00013719345930102362,
|
| 541 |
+
"loss": 2.6424,
|
| 542 |
"step": 445
|
| 543 |
},
|
| 544 |
{
|
| 545 |
+
"epoch": 0.99,
|
| 546 |
+
"learning_rate": 0.00013711989076069762,
|
| 547 |
+
"loss": 2.807,
|
| 548 |
"step": 450
|
| 549 |
},
|
| 550 |
{
|
| 551 |
+
"epoch": 1.0,
|
| 552 |
+
"learning_rate": 0.00013696466577167284,
|
| 553 |
+
"loss": 2.5944,
|
| 554 |
"step": 455
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 1.0,
|
| 558 |
+
"eval_loss": 2.5120232105255127,
|
| 559 |
+
"eval_runtime": 30.8025,
|
| 560 |
+
"eval_samples_per_second": 20.81,
|
| 561 |
+
"eval_steps_per_second": 2.63,
|
| 562 |
+
"step": 455
|
| 563 |
},
|
| 564 |
{
|
| 565 |
"epoch": 1.01,
|
| 566 |
+
"learning_rate": 0.00013672796931850215,
|
| 567 |
+
"loss": 2.5617,
|
| 568 |
"step": 460
|
| 569 |
},
|
| 570 |
{
|
| 571 |
"epoch": 1.02,
|
| 572 |
+
"learning_rate": 0.00013641008347682184,
|
| 573 |
+
"loss": 2.3941,
|
| 574 |
"step": 465
|
| 575 |
},
|
| 576 |
{
|
| 577 |
"epoch": 1.03,
|
| 578 |
+
"learning_rate": 0.0001360113870771968,
|
| 579 |
+
"loss": 2.609,
|
| 580 |
"step": 470
|
| 581 |
},
|
| 582 |
{
|
| 583 |
"epoch": 1.04,
|
| 584 |
+
"learning_rate": 0.000135532355253661,
|
| 585 |
+
"loss": 2.6935,
|
| 586 |
"step": 475
|
| 587 |
},
|
| 588 |
{
|
| 589 |
"epoch": 1.05,
|
| 590 |
+
"learning_rate": 0.00013497355887749114,
|
| 591 |
+
"loss": 2.4847,
|
| 592 |
"step": 480
|
| 593 |
},
|
| 594 |
{
|
| 595 |
+
"epoch": 1.07,
|
| 596 |
+
"learning_rate": 0.00013433566387688887,
|
| 597 |
+
"loss": 2.5804,
|
| 598 |
"step": 485
|
| 599 |
},
|
| 600 |
{
|
| 601 |
+
"epoch": 1.08,
|
| 602 |
+
"learning_rate": 0.00013361943044338116,
|
| 603 |
+
"loss": 2.493,
|
| 604 |
"step": 490
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 1.09,
|
| 608 |
+
"learning_rate": 0.000132825712125886,
|
| 609 |
+
"loss": 2.432,
|
| 610 |
"step": 495
|
| 611 |
},
|
| 612 |
{
|
| 613 |
"epoch": 1.1,
|
| 614 |
+
"learning_rate": 0.00013195545481352208,
|
| 615 |
+
"loss": 2.3443,
|
| 616 |
"step": 500
|
| 617 |
},
|
| 618 |
{
|
| 619 |
"epoch": 1.11,
|
| 620 |
+
"learning_rate": 0.0001310096956083743,
|
| 621 |
+
"loss": 2.5815,
|
| 622 |
"step": 505
|
| 623 |
},
|
| 624 |
{
|
| 625 |
"epoch": 1.12,
|
| 626 |
+
"learning_rate": 0.0001299895615895605,
|
| 627 |
+
"loss": 2.5337,
|
| 628 |
"step": 510
|
| 629 |
},
|
| 630 |
{
|
| 631 |
"epoch": 1.13,
|
| 632 |
+
"learning_rate": 0.00012889626847006907,
|
| 633 |
+
"loss": 2.3796,
|
| 634 |
"step": 515
|
| 635 |
},
|
| 636 |
{
|
| 637 |
"epoch": 1.14,
|
| 638 |
+
"learning_rate": 0.00012773111914797142,
|
| 639 |
+
"loss": 2.5178,
|
| 640 |
"step": 520
|
| 641 |
},
|
| 642 |
{
|
| 643 |
"epoch": 1.15,
|
| 644 |
+
"learning_rate": 0.00012649550215373373,
|
| 645 |
+
"loss": 2.4299,
|
| 646 |
"step": 525
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 1.16,
|
| 650 |
+
"learning_rate": 0.0001251908899954784,
|
| 651 |
+
"loss": 2.4451,
|
| 652 |
"step": 530
|
| 653 |
},
|
| 654 |
{
|
| 655 |
+
"epoch": 1.18,
|
| 656 |
+
"learning_rate": 0.00012381883740416864,
|
| 657 |
+
"loss": 2.6521,
|
| 658 |
"step": 535
|
| 659 |
},
|
| 660 |
{
|
| 661 |
+
"epoch": 1.19,
|
| 662 |
+
"learning_rate": 0.00012238097948080652,
|
| 663 |
+
"loss": 2.403,
|
| 664 |
"step": 540
|
| 665 |
},
|
| 666 |
{
|
| 667 |
"epoch": 1.2,
|
| 668 |
+
"learning_rate": 0.00012087902974785178,
|
| 669 |
+
"loss": 2.601,
|
| 670 |
"step": 545
|
| 671 |
},
|
| 672 |
{
|
| 673 |
"epoch": 1.21,
|
| 674 |
+
"learning_rate": 0.00011931477810718475,
|
| 675 |
+
"loss": 2.4694,
|
| 676 |
"step": 550
|
| 677 |
},
|
| 678 |
{
|
| 679 |
"epoch": 1.22,
|
| 680 |
+
"learning_rate": 0.00011769008870704702,
|
| 681 |
+
"loss": 2.5935,
|
| 682 |
"step": 555
|
| 683 |
},
|
| 684 |
{
|
| 685 |
"epoch": 1.23,
|
| 686 |
+
"learning_rate": 0.0001160068977204988,
|
| 687 |
+
"loss": 2.5179,
|
| 688 |
"step": 560
|
| 689 |
},
|
| 690 |
{
|
| 691 |
"epoch": 1.24,
|
| 692 |
+
"learning_rate": 0.00011426721103804855,
|
| 693 |
+
"loss": 2.7228,
|
| 694 |
"step": 565
|
| 695 |
},
|
| 696 |
{
|
| 697 |
"epoch": 1.25,
|
| 698 |
+
"learning_rate": 0.00011247310187719211,
|
| 699 |
+
"loss": 2.5971,
|
| 700 |
"step": 570
|
| 701 |
},
|
| 702 |
{
|
| 703 |
"epoch": 1.26,
|
| 704 |
+
"learning_rate": 0.00011062670831172054,
|
| 705 |
+
"loss": 2.5625,
|
| 706 |
"step": 575
|
| 707 |
},
|
| 708 |
{
|
| 709 |
"epoch": 1.27,
|
| 710 |
+
"learning_rate": 0.0001087302307237391,
|
| 711 |
+
"loss": 2.7913,
|
| 712 |
"step": 580
|
| 713 |
},
|
| 714 |
{
|
| 715 |
+
"epoch": 1.29,
|
| 716 |
+
"learning_rate": 0.00010678592918142871,
|
| 717 |
+
"loss": 2.4481,
|
| 718 |
"step": 585
|
| 719 |
},
|
| 720 |
{
|
| 721 |
+
"epoch": 1.3,
|
| 722 |
+
"learning_rate": 0.00010479612074568051,
|
| 723 |
+
"loss": 2.4637,
|
| 724 |
"step": 590
|
| 725 |
},
|
| 726 |
{
|
| 727 |
+
"epoch": 1.31,
|
| 728 |
+
"learning_rate": 0.00010276317670881092,
|
| 729 |
+
"loss": 2.7515,
|
| 730 |
"step": 595
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 1.32,
|
| 734 |
+
"learning_rate": 0.0001006895197686461,
|
| 735 |
+
"loss": 2.685,
|
| 736 |
"step": 600
|
| 737 |
},
|
| 738 |
{
|
| 739 |
"epoch": 1.33,
|
| 740 |
+
"learning_rate": 9.857762114134977e-05,
|
| 741 |
+
"loss": 2.6865,
|
| 742 |
"step": 605
|
| 743 |
},
|
| 744 |
{
|
| 745 |
"epoch": 1.34,
|
| 746 |
+
"learning_rate": 9.642999761642645e-05,
|
| 747 |
+
"loss": 2.586,
|
| 748 |
"step": 610
|
| 749 |
},
|
| 750 |
{
|
| 751 |
"epoch": 1.35,
|
| 752 |
+
"learning_rate": 9.424920855741624e-05,
|
| 753 |
+
"loss": 2.5638,
|
| 754 |
"step": 615
|
| 755 |
},
|
| 756 |
{
|
| 757 |
"epoch": 1.36,
|
| 758 |
+
"learning_rate": 9.203785285185348e-05,
|
| 759 |
+
"loss": 2.8532,
|
| 760 |
"step": 620
|
| 761 |
},
|
| 762 |
{
|
| 763 |
"epoch": 1.37,
|
| 764 |
+
"learning_rate": 8.979856581412148e-05,
|
| 765 |
+
"loss": 2.4697,
|
| 766 |
"step": 625
|
| 767 |
},
|
| 768 |
{
|
| 769 |
"epoch": 1.38,
|
| 770 |
+
"learning_rate": 8.753401604490128e-05,
|
| 771 |
+
"loss": 2.4653,
|
| 772 |
"step": 630
|
| 773 |
},
|
| 774 |
{
|
| 775 |
+
"epoch": 1.4,
|
| 776 |
+
"learning_rate": 8.5246902250948e-05,
|
| 777 |
+
"loss": 2.4832,
|
| 778 |
"step": 635
|
| 779 |
},
|
| 780 |
{
|
| 781 |
+
"epoch": 1.41,
|
| 782 |
+
"learning_rate": 8.293995002899144e-05,
|
| 783 |
+
"loss": 2.5514,
|
| 784 |
"step": 640
|
| 785 |
},
|
| 786 |
{
|
| 787 |
+
"epoch": 1.42,
|
| 788 |
+
"learning_rate": 8.061590861759147e-05,
|
| 789 |
+
"loss": 2.6216,
|
| 790 |
"step": 645
|
| 791 |
},
|
| 792 |
{
|
| 793 |
"epoch": 1.43,
|
| 794 |
+
"learning_rate": 7.827754762081649e-05,
|
| 795 |
+
"loss": 2.4314,
|
| 796 |
"step": 650
|
| 797 |
},
|
| 798 |
{
|
| 799 |
"epoch": 1.44,
|
| 800 |
+
"learning_rate": 7.592765370765685e-05,
|
| 801 |
+
"loss": 2.702,
|
| 802 |
"step": 655
|
| 803 |
},
|
| 804 |
{
|
| 805 |
"epoch": 1.45,
|
| 806 |
+
"learning_rate": 7.356902729109682e-05,
|
| 807 |
+
"loss": 2.3754,
|
| 808 |
"step": 660
|
| 809 |
},
|
| 810 |
{
|
| 811 |
"epoch": 1.46,
|
| 812 |
+
"learning_rate": 7.120447919081049e-05,
|
| 813 |
+
"loss": 2.4599,
|
| 814 |
"step": 665
|
| 815 |
},
|
| 816 |
{
|
| 817 |
"epoch": 1.47,
|
| 818 |
+
"learning_rate": 6.883682728345293e-05,
|
| 819 |
+
"loss": 2.7681,
|
| 820 |
"step": 670
|
| 821 |
},
|
| 822 |
{
|
| 823 |
"epoch": 1.48,
|
| 824 |
+
"learning_rate": 6.646889314454614e-05,
|
| 825 |
+
"loss": 2.4768,
|
| 826 |
"step": 675
|
| 827 |
},
|
| 828 |
{
|
| 829 |
"epoch": 1.49,
|
| 830 |
+
"learning_rate": 6.410349868595235e-05,
|
| 831 |
+
"loss": 2.6323,
|
| 832 |
"step": 680
|
| 833 |
},
|
| 834 |
{
|
| 835 |
+
"epoch": 1.51,
|
| 836 |
+
"learning_rate": 6.174346279294876e-05,
|
| 837 |
+
"loss": 2.4092,
|
| 838 |
"step": 685
|
| 839 |
},
|
| 840 |
{
|
| 841 |
+
"epoch": 1.52,
|
| 842 |
+
"learning_rate": 5.939159796490916e-05,
|
| 843 |
+
"loss": 2.6557,
|
| 844 |
"step": 690
|
| 845 |
},
|
| 846 |
{
|
| 847 |
+
"epoch": 1.53,
|
| 848 |
+
"learning_rate": 5.705070696359339e-05,
|
| 849 |
+
"loss": 2.3712,
|
| 850 |
"step": 695
|
| 851 |
},
|
| 852 |
{
|
| 853 |
"epoch": 1.54,
|
| 854 |
+
"learning_rate": 5.4723579473046015e-05,
|
| 855 |
+
"loss": 2.3837,
|
| 856 |
"step": 700
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 1.55,
|
| 860 |
+
"learning_rate": 5.241298877507526e-05,
|
| 861 |
+
"loss": 2.4259,
|
| 862 |
"step": 705
|
| 863 |
},
|
| 864 |
{
|
| 865 |
"epoch": 1.56,
|
| 866 |
+
"learning_rate": 5.01216884442811e-05,
|
| 867 |
+
"loss": 2.6717,
|
| 868 |
"step": 710
|
| 869 |
},
|
| 870 |
{
|
| 871 |
"epoch": 1.57,
|
| 872 |
+
"learning_rate": 4.785240906656888e-05,
|
| 873 |
+
"loss": 2.361,
|
| 874 |
"step": 715
|
| 875 |
},
|
| 876 |
{
|
| 877 |
"epoch": 1.58,
|
| 878 |
+
"learning_rate": 4.560785498505662e-05,
|
| 879 |
+
"loss": 2.4714,
|
| 880 |
"step": 720
|
| 881 |
},
|
| 882 |
{
|
| 883 |
"epoch": 1.59,
|
| 884 |
+
"learning_rate": 4.339070107726098e-05,
|
| 885 |
+
"loss": 2.6358,
|
| 886 |
"step": 725
|
| 887 |
},
|
| 888 |
{
|
| 889 |
"epoch": 1.6,
|
| 890 |
+
"learning_rate": 4.120358956739335e-05,
|
| 891 |
+
"loss": 2.5129,
|
| 892 |
"step": 730
|
| 893 |
},
|
| 894 |
{
|
| 895 |
+
"epoch": 1.62,
|
| 896 |
+
"learning_rate": 3.904912687757176e-05,
|
| 897 |
+
"loss": 2.5261,
|
| 898 |
"step": 735
|
| 899 |
},
|
| 900 |
{
|
| 901 |
+
"epoch": 1.63,
|
| 902 |
+
"learning_rate": 3.6929880521698846e-05,
|
| 903 |
+
"loss": 2.5826,
|
| 904 |
"step": 740
|
| 905 |
},
|
| 906 |
{
|
| 907 |
+
"epoch": 1.64,
|
| 908 |
+
"learning_rate": 3.484837604570524e-05,
|
| 909 |
+
"loss": 2.3181,
|
| 910 |
"step": 745
|
| 911 |
},
|
| 912 |
{
|
| 913 |
+
"epoch": 1.65,
|
| 914 |
+
"learning_rate": 3.2807094017811405e-05,
|
| 915 |
+
"loss": 2.2812,
|
| 916 |
"step": 750
|
| 917 |
},
|
| 918 |
{
|
| 919 |
"epoch": 1.66,
|
| 920 |
+
"learning_rate": 3.080846707238584e-05,
|
| 921 |
+
"loss": 2.479,
|
| 922 |
"step": 755
|
| 923 |
},
|
| 924 |
{
|
| 925 |
"epoch": 1.67,
|
| 926 |
+
"learning_rate": 2.885487701092958e-05,
|
| 927 |
+
"loss": 2.6417,
|
| 928 |
"step": 760
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 1.68,
|
| 932 |
+
"learning_rate": 2.694865196363672e-05,
|
| 933 |
+
"loss": 2.6846,
|
| 934 |
"step": 765
|
| 935 |
},
|
| 936 |
{
|
| 937 |
"epoch": 1.69,
|
| 938 |
+
"learning_rate": 2.509206361492011e-05,
|
| 939 |
+
"loss": 2.4952,
|
| 940 |
"step": 770
|
| 941 |
},
|
| 942 |
{
|
| 943 |
"epoch": 1.7,
|
| 944 |
+
"learning_rate": 2.3287324496200448e-05,
|
| 945 |
+
"loss": 2.5992,
|
| 946 |
"step": 775
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 1.71,
|
| 950 |
+
"learning_rate": 2.1536585349191225e-05,
|
| 951 |
+
"loss": 2.585,
|
| 952 |
"step": 780
|
| 953 |
},
|
| 954 |
{
|
| 955 |
+
"epoch": 1.73,
|
| 956 |
+
"learning_rate": 1.9841932562819154e-05,
|
| 957 |
+
"loss": 2.597,
|
| 958 |
"step": 785
|
| 959 |
},
|
| 960 |
{
|
| 961 |
+
"epoch": 1.74,
|
| 962 |
+
"learning_rate": 1.8205385686835072e-05,
|
| 963 |
+
"loss": 2.2072,
|
| 964 |
"step": 790
|
| 965 |
},
|
| 966 |
{
|
| 967 |
+
"epoch": 1.75,
|
| 968 |
+
"learning_rate": 1.6628895025078307e-05,
|
| 969 |
+
"loss": 2.5062,
|
| 970 |
"step": 795
|
| 971 |
},
|
| 972 |
{
|
| 973 |
+
"epoch": 1.76,
|
| 974 |
+
"learning_rate": 1.5114339311262777e-05,
|
| 975 |
+
"loss": 2.5392,
|
| 976 |
"step": 800
|
| 977 |
},
|
| 978 |
{
|
| 979 |
"epoch": 1.77,
|
| 980 |
+
"learning_rate": 1.3663523470054795e-05,
|
| 981 |
+
"loss": 2.5593,
|
| 982 |
"step": 805
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 1.78,
|
| 986 |
+
"learning_rate": 1.2278176466109037e-05,
|
| 987 |
+
"loss": 2.5837,
|
| 988 |
"step": 810
|
| 989 |
},
|
| 990 |
{
|
| 991 |
"epoch": 1.79,
|
| 992 |
+
"learning_rate": 1.0959949243630505e-05,
|
| 993 |
+
"loss": 2.3897,
|
| 994 |
"step": 815
|
| 995 |
},
|
| 996 |
{
|
| 997 |
"epoch": 1.8,
|
| 998 |
+
"learning_rate": 9.710412758911909e-06,
|
| 999 |
+
"loss": 2.4527,
|
| 1000 |
"step": 820
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 1.81,
|
| 1004 |
+
"learning_rate": 8.531056108195637e-06,
|
| 1005 |
+
"loss": 2.5424,
|
| 1006 |
"step": 825
|
| 1007 |
},
|
| 1008 |
{
|
| 1009 |
"epoch": 1.82,
|
| 1010 |
+
"learning_rate": 7.423284753089891e-06,
|
| 1011 |
+
"loss": 2.4882,
|
| 1012 |
"step": 830
|
| 1013 |
},
|
| 1014 |
{
|
| 1015 |
+
"epoch": 1.84,
|
| 1016 |
+
"learning_rate": 6.3884188456526086e-06,
|
| 1017 |
+
"loss": 2.4503,
|
| 1018 |
"step": 835
|
| 1019 |
},
|
| 1020 |
{
|
| 1021 |
+
"epoch": 1.85,
|
| 1022 |
+
"learning_rate": 5.427691655142655e-06,
|
| 1023 |
+
"loss": 2.3948,
|
| 1024 |
"step": 840
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
+
"epoch": 1.86,
|
| 1028 |
+
"learning_rate": 4.5422480983086555e-06,
|
| 1029 |
+
"loss": 2.6461,
|
| 1030 |
"step": 845
|
| 1031 |
},
|
| 1032 |
{
|
| 1033 |
+
"epoch": 1.87,
|
| 1034 |
+
"learning_rate": 3.7331433749702816e-06,
|
| 1035 |
+
"loss": 2.4818,
|
| 1036 |
"step": 850
|
| 1037 |
},
|
| 1038 |
{
|
| 1039 |
"epoch": 1.88,
|
| 1040 |
+
"learning_rate": 3.0013417105168102e-06,
|
| 1041 |
+
"loss": 2.6764,
|
| 1042 |
"step": 855
|
| 1043 |
},
|
| 1044 |
{
|
| 1045 |
"epoch": 1.89,
|
| 1046 |
+
"learning_rate": 2.347715206820977e-06,
|
| 1047 |
+
"loss": 2.4561,
|
| 1048 |
"step": 860
|
| 1049 |
},
|
| 1050 |
{
|
| 1051 |
"epoch": 1.9,
|
| 1052 |
+
"learning_rate": 1.7730428029395095e-06,
|
| 1053 |
+
"loss": 2.4363,
|
| 1054 |
"step": 865
|
| 1055 |
},
|
| 1056 |
{
|
| 1057 |
"epoch": 1.91,
|
| 1058 |
+
"learning_rate": 1.2780093468362156e-06,
|
| 1059 |
+
"loss": 2.5073,
|
| 1060 |
"step": 870
|
| 1061 |
},
|
| 1062 |
{
|
| 1063 |
"epoch": 1.92,
|
| 1064 |
+
"learning_rate": 8.632047792358677e-07,
|
| 1065 |
+
"loss": 2.3678,
|
| 1066 |
"step": 875
|
| 1067 |
},
|
| 1068 |
{
|
| 1069 |
"epoch": 1.93,
|
| 1070 |
+
"learning_rate": 5.291234305804284e-07,
|
| 1071 |
+
"loss": 2.5126,
|
| 1072 |
"step": 880
|
| 1073 |
},
|
| 1074 |
{
|
| 1075 |
+
"epoch": 1.95,
|
| 1076 |
+
"learning_rate": 2.761634319266075e-07,
|
| 1077 |
+
"loss": 2.5024,
|
| 1078 |
"step": 885
|
| 1079 |
},
|
| 1080 |
{
|
| 1081 |
+
"epoch": 1.96,
|
| 1082 |
+
"learning_rate": 1.0462624048520178e-07,
|
| 1083 |
+
"loss": 2.4416,
|
| 1084 |
"step": 890
|
| 1085 |
},
|
| 1086 |
{
|
| 1087 |
+
"epoch": 1.97,
|
| 1088 |
+
"learning_rate": 1.4716280368937374e-08,
|
| 1089 |
+
"loss": 2.3922,
|
| 1090 |
"step": 895
|
| 1091 |
},
|
| 1092 |
{
|
| 1093 |
+
"epoch": 1.98,
|
| 1094 |
+
"learning_rate": 6.540698976388004e-09,
|
| 1095 |
+
"loss": 2.7093,
|
| 1096 |
"step": 900
|
| 1097 |
},
|
| 1098 |
{
|
| 1099 |
+
"epoch": 1.99,
|
| 1100 |
+
"learning_rate": 8.010923930238477e-08,
|
| 1101 |
+
"loss": 2.4296,
|
| 1102 |
"step": 905
|
| 1103 |
},
|
| 1104 |
{
|
| 1105 |
"epoch": 2.0,
|
| 1106 |
+
"learning_rate": 2.3533422832715752e-07,
|
| 1107 |
+
"loss": 2.6948,
|
| 1108 |
"step": 910
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 2.0,
|
| 1112 |
+
"eval_loss": 2.4912426471710205,
|
| 1113 |
+
"eval_runtime": 30.6235,
|
| 1114 |
+
"eval_samples_per_second": 20.932,
|
| 1115 |
+
"eval_steps_per_second": 2.645,
|
| 1116 |
+
"step": 910
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1117 |
}
|
| 1118 |
],
|
| 1119 |
+
"max_steps": 2730,
|
| 1120 |
"num_train_epochs": 6,
|
| 1121 |
+
"total_flos": 950057828352000.0,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
| 1124 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2671
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:041b139d86c291eb088ecf942b80f42d53377dc136dbec1a58bef984db28f0fc
|
| 3 |
size 2671
|