Commit
•
ebd4b29
1
Parent(s):
71d7d67
huggingartists
Browse files- README.md +3 -3
- evaluation.txt +1 -1
- flax_model.msgpack +1 -1
- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +220 -1904
- 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/w7y07ejq/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/1j6e92ir) for full transparency and reproducibility.
|
| 55 |
|
| 56 |
+
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1j6e92ir/artifacts) is logged and versioned.
|
| 57 |
|
| 58 |
## How to use
|
| 59 |
|
evaluation.txt
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"eval_loss":
|
|
|
|
| 1 |
+
{"eval_loss": 2.026026487350464, "eval_runtime": 32.1272, "eval_samples_per_second": 20.699, "eval_steps_per_second": 2.615, "epoch": 5.0}
|
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 sha256:464d4e7e6e4c6382358d88fef714b72f5deae64e28bfb5b9bf00f9ff6030918b
|
| 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 sha256:9f578f796fe425b739c57c93d4870307639bdede01ade17192cbfd7dade1337c
|
| 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 sha256:5744d9da42094b4502ba2c2c52176dfb839e8656df36cf0783b4832a5444d42b
|
| 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 14503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1bd5d70a0d7170f2240ff3effc84cd6446ab98ec4b12f17cec590991fb7f5f94
|
| 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 sha256:5606edb00c1d85c79a7e6e65c60470a6b50e72dc1246597e70873019f5c126ed
|
| 3 |
size 623
|
trainer_state.json
CHANGED
|
@@ -1,2248 +1,564 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 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": 2.
|
| 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": 2.
|
| 26 |
"step": 15
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.04,
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": 2.
|
| 32 |
"step": 20
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 0.
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 2.
|
| 38 |
"step": 25
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.07,
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 30
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.08,
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss": 2.
|
| 50 |
"step": 35
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.09,
|
| 54 |
-
"learning_rate":
|
| 55 |
-
"loss": 2.
|
| 56 |
"step": 40
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.1,
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss": 2.
|
| 62 |
"step": 45
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.11,
|
| 66 |
-
"learning_rate":
|
| 67 |
-
"loss": 2.
|
| 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": 2.
|
| 80 |
"step": 60
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.14,
|
| 84 |
-
"learning_rate":
|
| 85 |
-
"loss": 2.
|
| 86 |
"step": 65
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.15,
|
| 90 |
-
"learning_rate":
|
| 91 |
-
"loss": 2.
|
| 92 |
"step": 70
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 0.
|
| 96 |
-
"learning_rate":
|
| 97 |
-
"loss": 2.
|
| 98 |
"step": 75
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.18,
|
| 102 |
-
"learning_rate":
|
| 103 |
-
"loss": 2.
|
| 104 |
"step": 80
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.19,
|
| 108 |
-
"learning_rate":
|
| 109 |
-
"loss": 2.
|
| 110 |
"step": 85
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.2,
|
| 114 |
-
"learning_rate":
|
| 115 |
-
"loss": 2.
|
| 116 |
"step": 90
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.21,
|
| 120 |
-
"learning_rate":
|
| 121 |
-
"loss": 2.
|
| 122 |
"step": 95
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.22,
|
| 126 |
-
"learning_rate":
|
| 127 |
-
"loss": 2.
|
| 128 |
"step": 100
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.23,
|
| 132 |
-
"learning_rate":
|
| 133 |
-
"loss": 2.
|
| 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": 2.
|
| 146 |
"step": 115
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 0.
|
| 150 |
-
"learning_rate":
|
| 151 |
-
"loss": 2.
|
| 152 |
"step": 120
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate":
|
| 157 |
-
"loss": 2.
|
| 158 |
"step": 125
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.29,
|
| 162 |
-
"learning_rate":
|
| 163 |
-
"loss": 2.
|
| 164 |
"step": 130
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.3,
|
| 168 |
-
"learning_rate":
|
| 169 |
-
"loss": 2.
|
| 170 |
"step": 135
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.31,
|
| 174 |
-
"learning_rate":
|
| 175 |
-
"loss": 2.
|
| 176 |
"step": 140
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.32,
|
| 180 |
-
"learning_rate":
|
| 181 |
-
"loss": 2.
|
| 182 |
"step": 145
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.33,
|
| 186 |
-
"learning_rate":
|
| 187 |
-
"loss": 2.
|
| 188 |
"step": 150
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.34,
|
| 192 |
-
"learning_rate":
|
| 193 |
-
"loss": 2.
|
| 194 |
"step": 155
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.35,
|
| 198 |
-
"learning_rate":
|
| 199 |
-
"loss":
|
| 200 |
"step": 160
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch": 0.
|
| 204 |
-
"learning_rate":
|
| 205 |
-
"loss": 2.
|
| 206 |
"step": 165
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 0.
|
| 210 |
-
"learning_rate":
|
| 211 |
-
"loss": 2.
|
| 212 |
"step": 170
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"learning_rate":
|
| 217 |
-
"loss": 2.
|
| 218 |
"step": 175
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.4,
|
| 222 |
-
"learning_rate":
|
| 223 |
-
"loss": 2.
|
| 224 |
"step": 180
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.41,
|
| 228 |
-
"learning_rate":
|
| 229 |
-
"loss": 2.
|
| 230 |
"step": 185
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.42,
|
| 234 |
-
"learning_rate":
|
| 235 |
-
"loss": 2.
|
| 236 |
"step": 190
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.43,
|
| 240 |
-
"learning_rate":
|
| 241 |
-
"loss": 2.
|
| 242 |
"step": 195
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.44,
|
| 246 |
-
"learning_rate": 6.
|
| 247 |
-
"loss": 2.
|
| 248 |
"step": 200
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.45,
|
| 252 |
-
"learning_rate": 6.
|
| 253 |
-
"loss": 2.
|
| 254 |
"step": 205
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.46,
|
| 258 |
-
"learning_rate":
|
| 259 |
-
"loss": 2.
|
| 260 |
"step": 210
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch": 0.
|
| 264 |
-
"learning_rate":
|
| 265 |
-
"loss": 2.
|
| 266 |
"step": 215
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"learning_rate":
|
| 271 |
-
"loss": 2.
|
| 272 |
"step": 220
|
| 273 |
},
|
| 274 |
{
|
| 275 |
-
"epoch": 0.
|
| 276 |
-
"learning_rate":
|
| 277 |
-
"loss": 2.
|
| 278 |
"step": 225
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.51,
|
| 282 |
-
"learning_rate":
|
| 283 |
-
"loss": 2.
|
| 284 |
"step": 230
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 0.52,
|
| 288 |
-
"learning_rate":
|
| 289 |
-
"loss": 2.
|
| 290 |
"step": 235
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 0.53,
|
| 294 |
-
"learning_rate":
|
| 295 |
-
"loss": 2.
|
| 296 |
"step": 240
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.54,
|
| 300 |
-
"learning_rate":
|
| 301 |
-
"loss": 2.
|
| 302 |
"step": 245
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.55,
|
| 306 |
-
"learning_rate":
|
| 307 |
-
"loss": 2.
|
| 308 |
"step": 250
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.56,
|
| 312 |
-
"learning_rate":
|
| 313 |
-
"loss": 2.
|
| 314 |
"step": 255
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch": 0.
|
| 318 |
-
"learning_rate":
|
| 319 |
-
"loss": 2.
|
| 320 |
"step": 260
|
| 321 |
},
|
| 322 |
{
|
| 323 |
-
"epoch": 0.
|
| 324 |
-
"learning_rate":
|
| 325 |
-
"loss": 2.
|
| 326 |
"step": 265
|
| 327 |
},
|
| 328 |
{
|
| 329 |
-
"epoch": 0.
|
| 330 |
-
"learning_rate":
|
| 331 |
-
"loss": 2.
|
| 332 |
"step": 270
|
| 333 |
},
|
| 334 |
{
|
| 335 |
-
"epoch": 0.
|
| 336 |
-
"learning_rate":
|
| 337 |
-
"loss": 2.
|
| 338 |
"step": 275
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.62,
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss": 2.
|
| 344 |
"step": 280
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.63,
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss": 2.
|
| 350 |
"step": 285
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.64,
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss": 2.
|
| 356 |
"step": 290
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 0.65,
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss": 2.
|
| 362 |
"step": 295
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 0.66,
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss": 2.
|
| 368 |
"step": 300
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 0.67,
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss": 2.
|
| 374 |
"step": 305
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch": 0.
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss": 2.
|
| 380 |
"step": 310
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch": 0.
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss": 2.
|
| 386 |
"step": 315
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch": 0.
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss": 2.
|
| 392 |
"step": 320
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch": 0.
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss": 2.
|
| 398 |
"step": 325
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 0.73,
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss": 2.
|
| 404 |
"step": 330
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 0.74,
|
| 408 |
-
"learning_rate":
|
| 409 |
-
"loss": 2.
|
| 410 |
"step": 335
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 0.75,
|
| 414 |
-
"learning_rate":
|
| 415 |
-
"loss":
|
| 416 |
"step": 340
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 0.76,
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"loss": 2.
|
| 422 |
"step": 345
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.77,
|
| 426 |
-
"learning_rate":
|
| 427 |
-
"loss": 2.
|
| 428 |
"step": 350
|
| 429 |
},
|
| 430 |
{
|
| 431 |
-
"epoch": 0.
|
| 432 |
-
"learning_rate":
|
| 433 |
-
"loss": 2.
|
| 434 |
"step": 355
|
| 435 |
},
|
| 436 |
{
|
| 437 |
-
"epoch": 0.
|
| 438 |
-
"learning_rate":
|
| 439 |
-
"loss": 2.
|
| 440 |
"step": 360
|
| 441 |
},
|
| 442 |
{
|
| 443 |
-
"epoch": 0.
|
| 444 |
-
"learning_rate":
|
| 445 |
-
"loss": 2.
|
| 446 |
"step": 365
|
| 447 |
},
|
| 448 |
{
|
| 449 |
-
"epoch": 0.
|
| 450 |
-
"learning_rate":
|
| 451 |
-
"loss": 2.
|
| 452 |
"step": 370
|
| 453 |
},
|
| 454 |
{
|
| 455 |
-
"epoch": 0.
|
| 456 |
-
"learning_rate":
|
| 457 |
-
"loss": 2.
|
| 458 |
"step": 375
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 0.84,
|
| 462 |
-
"learning_rate":
|
| 463 |
-
"loss":
|
| 464 |
"step": 380
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.85,
|
| 468 |
-
"learning_rate":
|
| 469 |
-
"loss": 2.
|
| 470 |
"step": 385
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 0.86,
|
| 474 |
-
"learning_rate":
|
| 475 |
-
"loss": 2.
|
| 476 |
"step": 390
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 0.87,
|
| 480 |
-
"learning_rate":
|
| 481 |
-
"loss": 2.
|
| 482 |
"step": 395
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 0.88,
|
| 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.
|
| 498 |
-
"learning_rate":
|
| 499 |
-
"loss": 2.
|
| 500 |
"step": 410
|
| 501 |
},
|
| 502 |
{
|
| 503 |
-
"epoch": 0.
|
| 504 |
-
"learning_rate":
|
| 505 |
-
"loss": 2.
|
| 506 |
"step": 415
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch": 0.
|
| 510 |
-
"learning_rate":
|
| 511 |
-
"loss": 2.
|
| 512 |
"step": 420
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch": 0.
|
| 516 |
-
"learning_rate":
|
| 517 |
-
"loss": 2.
|
| 518 |
"step": 425
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 0.95,
|
| 522 |
-
"learning_rate":
|
| 523 |
-
"loss": 2.
|
| 524 |
"step": 430
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 0.96,
|
| 528 |
-
"learning_rate":
|
| 529 |
-
"loss": 2.
|
| 530 |
"step": 435
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 0.97,
|
| 534 |
-
"learning_rate":
|
| 535 |
-
"loss": 2.
|
| 536 |
"step": 440
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 0.98,
|
| 540 |
-
"learning_rate":
|
| 541 |
-
"loss": 2.
|
| 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":
|
| 553 |
-
"loss": 2.
|
| 554 |
-
"step":
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 1.0,
|
| 558 |
-
"eval_loss": 2.
|
| 559 |
-
"eval_runtime":
|
| 560 |
-
"eval_samples_per_second": 20.
|
| 561 |
-
"eval_steps_per_second": 2.
|
| 562 |
-
"step":
|
| 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 |
-
"epoch": 2.04,
|
| 1120 |
-
"learning_rate": 1.1246491964533778e-05,
|
| 1121 |
-
"loss": 2.424,
|
| 1122 |
-
"step": 915
|
| 1123 |
-
},
|
| 1124 |
-
{
|
| 1125 |
-
"epoch": 2.05,
|
| 1126 |
-
"learning_rate": 1.2598018517843792e-05,
|
| 1127 |
-
"loss": 2.5084,
|
| 1128 |
-
"step": 920
|
| 1129 |
-
},
|
| 1130 |
-
{
|
| 1131 |
-
"epoch": 2.06,
|
| 1132 |
-
"learning_rate": 1.4018079085430647e-05,
|
| 1133 |
-
"loss": 2.3701,
|
| 1134 |
-
"step": 925
|
| 1135 |
-
},
|
| 1136 |
-
{
|
| 1137 |
-
"epoch": 2.07,
|
| 1138 |
-
"learning_rate": 1.5504935827871155e-05,
|
| 1139 |
-
"loss": 2.4517,
|
| 1140 |
-
"step": 930
|
| 1141 |
-
},
|
| 1142 |
-
{
|
| 1143 |
-
"epoch": 2.08,
|
| 1144 |
-
"learning_rate": 1.7056769162026474e-05,
|
| 1145 |
-
"loss": 2.6206,
|
| 1146 |
-
"step": 935
|
| 1147 |
-
},
|
| 1148 |
-
{
|
| 1149 |
-
"epoch": 2.09,
|
| 1150 |
-
"learning_rate": 1.867167998781036e-05,
|
| 1151 |
-
"loss": 2.5704,
|
| 1152 |
-
"step": 940
|
| 1153 |
-
},
|
| 1154 |
-
{
|
| 1155 |
-
"epoch": 2.1,
|
| 1156 |
-
"learning_rate": 2.0347692012266013e-05,
|
| 1157 |
-
"loss": 2.4849,
|
| 1158 |
-
"step": 945
|
| 1159 |
-
},
|
| 1160 |
-
{
|
| 1161 |
-
"epoch": 2.12,
|
| 1162 |
-
"learning_rate": 2.208275416810954e-05,
|
| 1163 |
-
"loss": 2.5616,
|
| 1164 |
-
"step": 950
|
| 1165 |
-
},
|
| 1166 |
-
{
|
| 1167 |
-
"epoch": 2.13,
|
| 1168 |
-
"learning_rate": 2.3874743123783855e-05,
|
| 1169 |
-
"loss": 2.3109,
|
| 1170 |
-
"step": 955
|
| 1171 |
-
},
|
| 1172 |
-
{
|
| 1173 |
-
"epoch": 2.14,
|
| 1174 |
-
"learning_rate": 2.572146588194454e-05,
|
| 1175 |
-
"loss": 2.4892,
|
| 1176 |
-
"step": 960
|
| 1177 |
-
},
|
| 1178 |
-
{
|
| 1179 |
-
"epoch": 2.15,
|
| 1180 |
-
"learning_rate": 2.7620662463202956e-05,
|
| 1181 |
-
"loss": 2.443,
|
| 1182 |
-
"step": 965
|
| 1183 |
-
},
|
| 1184 |
-
{
|
| 1185 |
-
"epoch": 2.16,
|
| 1186 |
-
"learning_rate": 2.957000867184013e-05,
|
| 1187 |
-
"loss": 2.4545,
|
| 1188 |
-
"step": 970
|
| 1189 |
-
},
|
| 1190 |
-
{
|
| 1191 |
-
"epoch": 2.17,
|
| 1192 |
-
"learning_rate": 3.156711894010725e-05,
|
| 1193 |
-
"loss": 2.4657,
|
| 1194 |
-
"step": 975
|
| 1195 |
-
},
|
| 1196 |
-
{
|
| 1197 |
-
"epoch": 2.18,
|
| 1198 |
-
"learning_rate": 3.360954924763191e-05,
|
| 1199 |
-
"loss": 2.1432,
|
| 1200 |
-
"step": 980
|
| 1201 |
-
},
|
| 1202 |
-
{
|
| 1203 |
-
"epoch": 2.19,
|
| 1204 |
-
"learning_rate": 3.569480011235739e-05,
|
| 1205 |
-
"loss": 2.5793,
|
| 1206 |
-
"step": 985
|
| 1207 |
-
},
|
| 1208 |
-
{
|
| 1209 |
-
"epoch": 2.2,
|
| 1210 |
-
"learning_rate": 3.782031964935521e-05,
|
| 1211 |
-
"loss": 2.3062,
|
| 1212 |
-
"step": 990
|
| 1213 |
-
},
|
| 1214 |
-
{
|
| 1215 |
-
"epoch": 2.22,
|
| 1216 |
-
"learning_rate": 3.998350669376499e-05,
|
| 1217 |
-
"loss": 2.3345,
|
| 1218 |
-
"step": 995
|
| 1219 |
-
},
|
| 1220 |
-
{
|
| 1221 |
-
"epoch": 2.23,
|
| 1222 |
-
"learning_rate": 4.218171398404682e-05,
|
| 1223 |
-
"loss": 2.5329,
|
| 1224 |
-
"step": 1000
|
| 1225 |
-
},
|
| 1226 |
-
{
|
| 1227 |
-
"epoch": 2.24,
|
| 1228 |
-
"learning_rate": 4.4412251401639226e-05,
|
| 1229 |
-
"loss": 2.201,
|
| 1230 |
-
"step": 1005
|
| 1231 |
-
},
|
| 1232 |
-
{
|
| 1233 |
-
"epoch": 2.25,
|
| 1234 |
-
"learning_rate": 4.667238926307116e-05,
|
| 1235 |
-
"loss": 2.5198,
|
| 1236 |
-
"step": 1010
|
| 1237 |
-
},
|
| 1238 |
-
{
|
| 1239 |
-
"epoch": 2.26,
|
| 1240 |
-
"learning_rate": 4.895936166048787e-05,
|
| 1241 |
-
"loss": 2.5241,
|
| 1242 |
-
"step": 1015
|
| 1243 |
-
},
|
| 1244 |
-
{
|
| 1245 |
-
"epoch": 2.27,
|
| 1246 |
-
"learning_rate": 5.1270369846509574e-05,
|
| 1247 |
-
"loss": 2.4079,
|
| 1248 |
-
"step": 1020
|
| 1249 |
-
},
|
| 1250 |
-
{
|
| 1251 |
-
"epoch": 2.28,
|
| 1252 |
-
"learning_rate": 5.360258565927803e-05,
|
| 1253 |
-
"loss": 2.5147,
|
| 1254 |
-
"step": 1025
|
| 1255 |
-
},
|
| 1256 |
-
{
|
| 1257 |
-
"epoch": 2.29,
|
| 1258 |
-
"learning_rate": 5.595315498349983e-05,
|
| 1259 |
-
"loss": 2.5063,
|
| 1260 |
-
"step": 1030
|
| 1261 |
-
},
|
| 1262 |
-
{
|
| 1263 |
-
"epoch": 2.31,
|
| 1264 |
-
"learning_rate": 5.83192012432515e-05,
|
| 1265 |
-
"loss": 2.3742,
|
| 1266 |
-
"step": 1035
|
| 1267 |
-
},
|
| 1268 |
-
{
|
| 1269 |
-
"epoch": 2.32,
|
| 1270 |
-
"learning_rate": 6.069782892226903e-05,
|
| 1271 |
-
"loss": 2.625,
|
| 1272 |
-
"step": 1040
|
| 1273 |
-
},
|
| 1274 |
-
{
|
| 1275 |
-
"epoch": 2.33,
|
| 1276 |
-
"learning_rate": 6.308612710742091e-05,
|
| 1277 |
-
"loss": 2.4382,
|
| 1278 |
-
"step": 1045
|
| 1279 |
-
},
|
| 1280 |
-
{
|
| 1281 |
-
"epoch": 2.34,
|
| 1282 |
-
"learning_rate": 6.548117305101883e-05,
|
| 1283 |
-
"loss": 2.4238,
|
| 1284 |
-
"step": 1050
|
| 1285 |
-
},
|
| 1286 |
-
{
|
| 1287 |
-
"epoch": 2.35,
|
| 1288 |
-
"learning_rate": 6.788003574761404e-05,
|
| 1289 |
-
"loss": 2.4157,
|
| 1290 |
-
"step": 1055
|
| 1291 |
-
},
|
| 1292 |
-
{
|
| 1293 |
-
"epoch": 2.36,
|
| 1294 |
-
"learning_rate": 7.027977952089649e-05,
|
| 1295 |
-
"loss": 2.5458,
|
| 1296 |
-
"step": 1060
|
| 1297 |
-
},
|
| 1298 |
-
{
|
| 1299 |
-
"epoch": 2.37,
|
| 1300 |
-
"learning_rate": 7.267746761631506e-05,
|
| 1301 |
-
"loss": 2.5048,
|
| 1302 |
-
"step": 1065
|
| 1303 |
-
},
|
| 1304 |
-
{
|
| 1305 |
-
"epoch": 2.38,
|
| 1306 |
-
"learning_rate": 7.507016579501221e-05,
|
| 1307 |
-
"loss": 2.5779,
|
| 1308 |
-
"step": 1070
|
| 1309 |
-
},
|
| 1310 |
-
{
|
| 1311 |
-
"epoch": 2.39,
|
| 1312 |
-
"learning_rate": 7.745494592468267e-05,
|
| 1313 |
-
"loss": 2.5149,
|
| 1314 |
-
"step": 1075
|
| 1315 |
-
},
|
| 1316 |
-
{
|
| 1317 |
-
"epoch": 2.41,
|
| 1318 |
-
"learning_rate": 7.98288895629588e-05,
|
| 1319 |
-
"loss": 2.4972,
|
| 1320 |
-
"step": 1080
|
| 1321 |
-
},
|
| 1322 |
-
{
|
| 1323 |
-
"epoch": 2.42,
|
| 1324 |
-
"learning_rate": 8.218909152893547e-05,
|
| 1325 |
-
"loss": 2.2248,
|
| 1326 |
-
"step": 1085
|
| 1327 |
-
},
|
| 1328 |
-
{
|
| 1329 |
-
"epoch": 2.43,
|
| 1330 |
-
"learning_rate": 8.453266345847049e-05,
|
| 1331 |
-
"loss": 2.3945,
|
| 1332 |
-
"step": 1090
|
| 1333 |
-
},
|
| 1334 |
-
{
|
| 1335 |
-
"epoch": 2.44,
|
| 1336 |
-
"learning_rate": 8.685673733890025e-05,
|
| 1337 |
-
"loss": 2.4223,
|
| 1338 |
-
"step": 1095
|
| 1339 |
-
},
|
| 1340 |
-
{
|
| 1341 |
-
"epoch": 2.45,
|
| 1342 |
-
"learning_rate": 8.915846901885212e-05,
|
| 1343 |
-
"loss": 2.5513,
|
| 1344 |
-
"step": 1100
|
| 1345 |
-
},
|
| 1346 |
-
{
|
| 1347 |
-
"epoch": 2.46,
|
| 1348 |
-
"learning_rate": 9.143504168885545e-05,
|
| 1349 |
-
"loss": 2.527,
|
| 1350 |
-
"step": 1105
|
| 1351 |
-
},
|
| 1352 |
-
{
|
| 1353 |
-
"epoch": 2.47,
|
| 1354 |
-
"learning_rate": 9.368366932849205e-05,
|
| 1355 |
-
"loss": 2.6526,
|
| 1356 |
-
"step": 1110
|
| 1357 |
-
},
|
| 1358 |
-
{
|
| 1359 |
-
"epoch": 2.48,
|
| 1360 |
-
"learning_rate": 9.590160011586822e-05,
|
| 1361 |
-
"loss": 2.6926,
|
| 1362 |
-
"step": 1115
|
| 1363 |
-
},
|
| 1364 |
-
{
|
| 1365 |
-
"epoch": 2.49,
|
| 1366 |
-
"learning_rate": 9.808611979523307e-05,
|
| 1367 |
-
"loss": 2.4965,
|
| 1368 |
-
"step": 1120
|
| 1369 |
-
},
|
| 1370 |
-
{
|
| 1371 |
-
"epoch": 2.51,
|
| 1372 |
-
"learning_rate": 0.00010023455499862926,
|
| 1373 |
-
"loss": 2.4985,
|
| 1374 |
-
"step": 1125
|
| 1375 |
-
},
|
| 1376 |
-
{
|
| 1377 |
-
"epoch": 2.52,
|
| 1378 |
-
"learning_rate": 0.00010234427651749925,
|
| 1379 |
-
"loss": 2.6502,
|
| 1380 |
-
"step": 1130
|
| 1381 |
-
},
|
| 1382 |
-
{
|
| 1383 |
-
"epoch": 2.53,
|
| 1384 |
-
"learning_rate": 0.00010441270252025704,
|
| 1385 |
-
"loss": 2.5715,
|
| 1386 |
-
"step": 1135
|
| 1387 |
-
},
|
| 1388 |
-
{
|
| 1389 |
-
"epoch": 2.54,
|
| 1390 |
-
"learning_rate": 0.00010643730171187608,
|
| 1391 |
-
"loss": 2.1817,
|
| 1392 |
-
"step": 1140
|
| 1393 |
-
},
|
| 1394 |
-
{
|
| 1395 |
-
"epoch": 2.55,
|
| 1396 |
-
"learning_rate": 0.00010841559643163376,
|
| 1397 |
-
"loss": 2.3754,
|
| 1398 |
-
"step": 1145
|
| 1399 |
-
},
|
| 1400 |
-
{
|
| 1401 |
-
"epoch": 2.56,
|
| 1402 |
-
"learning_rate": 0.00011034516568521922,
|
| 1403 |
-
"loss": 2.8706,
|
| 1404 |
-
"step": 1150
|
| 1405 |
-
},
|
| 1406 |
-
{
|
| 1407 |
-
"epoch": 2.57,
|
| 1408 |
-
"learning_rate": 0.0001122236481074937,
|
| 1409 |
-
"loss": 2.6822,
|
| 1410 |
-
"step": 1155
|
| 1411 |
-
},
|
| 1412 |
-
{
|
| 1413 |
-
"epoch": 2.58,
|
| 1414 |
-
"learning_rate": 0.0001140487448522782,
|
| 1415 |
-
"loss": 2.5523,
|
| 1416 |
-
"step": 1160
|
| 1417 |
-
},
|
| 1418 |
-
{
|
| 1419 |
-
"epoch": 2.59,
|
| 1420 |
-
"learning_rate": 0.0001158182224056317,
|
| 1421 |
-
"loss": 2.4356,
|
| 1422 |
-
"step": 1165
|
| 1423 |
-
},
|
| 1424 |
-
{
|
| 1425 |
-
"epoch": 2.61,
|
| 1426 |
-
"learning_rate": 0.00011752991531917766,
|
| 1427 |
-
"loss": 2.5516,
|
| 1428 |
-
"step": 1170
|
| 1429 |
-
},
|
| 1430 |
-
{
|
| 1431 |
-
"epoch": 2.62,
|
| 1432 |
-
"learning_rate": 0.00011918172886013147,
|
| 1433 |
-
"loss": 2.4273,
|
| 1434 |
-
"step": 1175
|
| 1435 |
-
},
|
| 1436 |
-
{
|
| 1437 |
-
"epoch": 2.63,
|
| 1438 |
-
"learning_rate": 0.00012077164157479157,
|
| 1439 |
-
"loss": 2.2673,
|
| 1440 |
-
"step": 1180
|
| 1441 |
-
},
|
| 1442 |
-
{
|
| 1443 |
-
"epoch": 2.64,
|
| 1444 |
-
"learning_rate": 0.00012229770776234816,
|
| 1445 |
-
"loss": 2.6163,
|
| 1446 |
-
"step": 1185
|
| 1447 |
-
},
|
| 1448 |
-
{
|
| 1449 |
-
"epoch": 2.65,
|
| 1450 |
-
"learning_rate": 0.000123758059855992,
|
| 1451 |
-
"loss": 2.5445,
|
| 1452 |
-
"step": 1190
|
| 1453 |
-
},
|
| 1454 |
-
{
|
| 1455 |
-
"epoch": 2.66,
|
| 1456 |
-
"learning_rate": 0.00012515091070840103,
|
| 1457 |
-
"loss": 2.3948,
|
| 1458 |
-
"step": 1195
|
| 1459 |
-
},
|
| 1460 |
-
{
|
| 1461 |
-
"epoch": 2.67,
|
| 1462 |
-
"learning_rate": 0.00012647455577881304,
|
| 1463 |
-
"loss": 2.5891,
|
| 1464 |
-
"step": 1200
|
| 1465 |
-
},
|
| 1466 |
-
{
|
| 1467 |
-
"epoch": 2.68,
|
| 1468 |
-
"learning_rate": 0.00012772737521900558,
|
| 1469 |
-
"loss": 2.5894,
|
| 1470 |
-
"step": 1205
|
| 1471 |
-
},
|
| 1472 |
-
{
|
| 1473 |
-
"epoch": 2.69,
|
| 1474 |
-
"learning_rate": 0.00012890783585563144,
|
| 1475 |
-
"loss": 2.5051,
|
| 1476 |
-
"step": 1210
|
| 1477 |
-
},
|
| 1478 |
-
{
|
| 1479 |
-
"epoch": 2.71,
|
| 1480 |
-
"learning_rate": 0.0001300144930664832,
|
| 1481 |
-
"loss": 2.4751,
|
| 1482 |
-
"step": 1215
|
| 1483 |
-
},
|
| 1484 |
-
{
|
| 1485 |
-
"epoch": 2.72,
|
| 1486 |
-
"learning_rate": 0.00013104599254838963,
|
| 1487 |
-
"loss": 2.4493,
|
| 1488 |
-
"step": 1220
|
| 1489 |
-
},
|
| 1490 |
-
{
|
| 1491 |
-
"epoch": 2.73,
|
| 1492 |
-
"learning_rate": 0.00013200107197458417,
|
| 1493 |
-
"loss": 2.4946,
|
| 1494 |
-
"step": 1225
|
| 1495 |
-
},
|
| 1496 |
-
{
|
| 1497 |
-
"epoch": 2.74,
|
| 1498 |
-
"learning_rate": 0.0001328785625395121,
|
| 1499 |
-
"loss": 2.4618,
|
| 1500 |
-
"step": 1230
|
| 1501 |
-
},
|
| 1502 |
-
{
|
| 1503 |
-
"epoch": 2.75,
|
| 1504 |
-
"learning_rate": 0.00013367739038918988,
|
| 1505 |
-
"loss": 2.5223,
|
| 1506 |
-
"step": 1235
|
| 1507 |
-
},
|
| 1508 |
-
{
|
| 1509 |
-
"epoch": 2.76,
|
| 1510 |
-
"learning_rate": 0.0001343965779353643,
|
| 1511 |
-
"loss": 2.4409,
|
| 1512 |
-
"step": 1240
|
| 1513 |
-
},
|
| 1514 |
-
{
|
| 1515 |
-
"epoch": 2.77,
|
| 1516 |
-
"learning_rate": 0.0001350352450518637,
|
| 1517 |
-
"loss": 2.3083,
|
| 1518 |
-
"step": 1245
|
| 1519 |
-
},
|
| 1520 |
-
{
|
| 1521 |
-
"epoch": 2.78,
|
| 1522 |
-
"learning_rate": 0.00013559261015167785,
|
| 1523 |
-
"loss": 2.4411,
|
| 1524 |
-
"step": 1250
|
| 1525 |
-
},
|
| 1526 |
-
{
|
| 1527 |
-
"epoch": 2.8,
|
| 1528 |
-
"learning_rate": 0.0001360679911434468,
|
| 1529 |
-
"loss": 2.3657,
|
| 1530 |
-
"step": 1255
|
| 1531 |
-
},
|
| 1532 |
-
{
|
| 1533 |
-
"epoch": 2.81,
|
| 1534 |
-
"learning_rate": 0.00013646080626618978,
|
| 1535 |
-
"loss": 2.4467,
|
| 1536 |
-
"step": 1260
|
| 1537 |
-
},
|
| 1538 |
-
{
|
| 1539 |
-
"epoch": 2.82,
|
| 1540 |
-
"learning_rate": 0.0001367705748012514,
|
| 1541 |
-
"loss": 2.3811,
|
| 1542 |
-
"step": 1265
|
| 1543 |
-
},
|
| 1544 |
-
{
|
| 1545 |
-
"epoch": 2.83,
|
| 1546 |
-
"learning_rate": 0.0001369969176605951,
|
| 1547 |
-
"loss": 2.5053,
|
| 1548 |
-
"step": 1270
|
| 1549 |
-
},
|
| 1550 |
-
{
|
| 1551 |
-
"epoch": 2.84,
|
| 1552 |
-
"learning_rate": 0.00013713955785072274,
|
| 1553 |
-
"loss": 2.7194,
|
| 1554 |
-
"step": 1275
|
| 1555 |
-
},
|
| 1556 |
-
{
|
| 1557 |
-
"epoch": 2.85,
|
| 1558 |
-
"learning_rate": 0.0001371983208116533,
|
| 1559 |
-
"loss": 2.5204,
|
| 1560 |
-
"step": 1280
|
| 1561 |
-
},
|
| 1562 |
-
{
|
| 1563 |
-
"epoch": 2.86,
|
| 1564 |
-
"learning_rate": 0.0001371731346305456,
|
| 1565 |
-
"loss": 2.5025,
|
| 1566 |
-
"step": 1285
|
| 1567 |
-
},
|
| 1568 |
-
{
|
| 1569 |
-
"epoch": 2.87,
|
| 1570 |
-
"learning_rate": 0.00013706403012970347,
|
| 1571 |
-
"loss": 2.4982,
|
| 1572 |
-
"step": 1290
|
| 1573 |
-
},
|
| 1574 |
-
{
|
| 1575 |
-
"epoch": 2.88,
|
| 1576 |
-
"learning_rate": 0.00013687114082885652,
|
| 1577 |
-
"loss": 2.7208,
|
| 1578 |
-
"step": 1295
|
| 1579 |
-
},
|
| 1580 |
-
{
|
| 1581 |
-
"epoch": 2.9,
|
| 1582 |
-
"learning_rate": 0.00013659470278176106,
|
| 1583 |
-
"loss": 2.5839,
|
| 1584 |
-
"step": 1300
|
| 1585 |
-
},
|
| 1586 |
-
{
|
| 1587 |
-
"epoch": 2.91,
|
| 1588 |
-
"learning_rate": 0.00013623505428732318,
|
| 1589 |
-
"loss": 2.5333,
|
| 1590 |
-
"step": 1305
|
| 1591 |
-
},
|
| 1592 |
-
{
|
| 1593 |
-
"epoch": 2.92,
|
| 1594 |
-
"learning_rate": 0.0001357926354755953,
|
| 1595 |
-
"loss": 2.5095,
|
| 1596 |
-
"step": 1310
|
| 1597 |
-
},
|
| 1598 |
-
{
|
| 1599 |
-
"epoch": 2.93,
|
| 1600 |
-
"learning_rate": 0.0001352679877691551,
|
| 1601 |
-
"loss": 2.2663,
|
| 1602 |
-
"step": 1315
|
| 1603 |
-
},
|
| 1604 |
-
{
|
| 1605 |
-
"epoch": 2.94,
|
| 1606 |
-
"learning_rate": 0.00013466175322052366,
|
| 1607 |
-
"loss": 2.5816,
|
| 1608 |
-
"step": 1320
|
| 1609 |
-
},
|
| 1610 |
-
{
|
| 1611 |
-
"epoch": 2.95,
|
| 1612 |
-
"learning_rate": 0.00013397467372643594,
|
| 1613 |
-
"loss": 2.4828,
|
| 1614 |
-
"step": 1325
|
| 1615 |
-
},
|
| 1616 |
-
{
|
| 1617 |
-
"epoch": 2.96,
|
| 1618 |
-
"learning_rate": 0.00013320759011992343,
|
| 1619 |
-
"loss": 2.5478,
|
| 1620 |
-
"step": 1330
|
| 1621 |
-
},
|
| 1622 |
-
{
|
| 1623 |
-
"epoch": 2.97,
|
| 1624 |
-
"learning_rate": 0.00013236144114132077,
|
| 1625 |
-
"loss": 2.5125,
|
| 1626 |
-
"step": 1335
|
| 1627 |
-
},
|
| 1628 |
-
{
|
| 1629 |
-
"epoch": 2.98,
|
| 1630 |
-
"learning_rate": 0.0001314372622894558,
|
| 1631 |
-
"loss": 2.5845,
|
| 1632 |
-
"step": 1340
|
| 1633 |
-
},
|
| 1634 |
-
{
|
| 1635 |
-
"epoch": 3.0,
|
| 1636 |
-
"learning_rate": 0.00013043618455442838,
|
| 1637 |
-
"loss": 2.6289,
|
| 1638 |
-
"step": 1345
|
| 1639 |
-
},
|
| 1640 |
-
{
|
| 1641 |
-
"epoch": 3.0,
|
| 1642 |
-
"eval_loss": 2.2795188426971436,
|
| 1643 |
-
"eval_runtime": 30.9908,
|
| 1644 |
-
"eval_samples_per_second": 22.329,
|
| 1645 |
-
"eval_steps_per_second": 2.807,
|
| 1646 |
-
"step": 1347
|
| 1647 |
-
},
|
| 1648 |
-
{
|
| 1649 |
-
"epoch": 2.99,
|
| 1650 |
-
"learning_rate": 0.000135403446735711,
|
| 1651 |
-
"loss": 2.1728,
|
| 1652 |
-
"step": 1350
|
| 1653 |
-
},
|
| 1654 |
-
{
|
| 1655 |
-
"epoch": 3.0,
|
| 1656 |
-
"learning_rate": 0.00013482119897697306,
|
| 1657 |
-
"loss": 2.2896,
|
| 1658 |
-
"step": 1355
|
| 1659 |
-
},
|
| 1660 |
-
{
|
| 1661 |
-
"epoch": 3.0,
|
| 1662 |
-
"eval_loss": 2.07898211479187,
|
| 1663 |
-
"eval_runtime": 29.7025,
|
| 1664 |
-
"eval_samples_per_second": 22.523,
|
| 1665 |
-
"eval_steps_per_second": 2.828,
|
| 1666 |
-
"step": 1356
|
| 1667 |
-
},
|
| 1668 |
-
{
|
| 1669 |
-
"epoch": 3.01,
|
| 1670 |
-
"learning_rate": 0.0001341589832217866,
|
| 1671 |
-
"loss": 2.2492,
|
| 1672 |
-
"step": 1360
|
| 1673 |
-
},
|
| 1674 |
-
{
|
| 1675 |
-
"epoch": 3.02,
|
| 1676 |
-
"learning_rate": 0.00013341759915466326,
|
| 1677 |
-
"loss": 2.4062,
|
| 1678 |
-
"step": 1365
|
| 1679 |
-
},
|
| 1680 |
-
{
|
| 1681 |
-
"epoch": 3.03,
|
| 1682 |
-
"learning_rate": 0.00013259794206290322,
|
| 1683 |
-
"loss": 2.4505,
|
| 1684 |
-
"step": 1370
|
| 1685 |
-
},
|
| 1686 |
-
{
|
| 1687 |
-
"epoch": 3.04,
|
| 1688 |
-
"learning_rate": 0.00013170100175545615,
|
| 1689 |
-
"loss": 2.3168,
|
| 1690 |
-
"step": 1375
|
| 1691 |
-
},
|
| 1692 |
-
{
|
| 1693 |
-
"epoch": 3.05,
|
| 1694 |
-
"learning_rate": 0.00013072786136763862,
|
| 1695 |
-
"loss": 2.4331,
|
| 1696 |
-
"step": 1380
|
| 1697 |
-
},
|
| 1698 |
-
{
|
| 1699 |
-
"epoch": 3.06,
|
| 1700 |
-
"learning_rate": 0.0001296796960531519,
|
| 1701 |
-
"loss": 2.3448,
|
| 1702 |
-
"step": 1385
|
| 1703 |
-
},
|
| 1704 |
-
{
|
| 1705 |
-
"epoch": 3.08,
|
| 1706 |
-
"learning_rate": 0.00012855777156497896,
|
| 1707 |
-
"loss": 2.5707,
|
| 1708 |
-
"step": 1390
|
| 1709 |
-
},
|
| 1710 |
-
{
|
| 1711 |
-
"epoch": 3.09,
|
| 1712 |
-
"learning_rate": 0.00012736344272687528,
|
| 1713 |
-
"loss": 2.3018,
|
| 1714 |
-
"step": 1395
|
| 1715 |
-
},
|
| 1716 |
-
{
|
| 1717 |
-
"epoch": 3.1,
|
| 1718 |
-
"learning_rate": 0.00012609815179729852,
|
| 1719 |
-
"loss": 2.2497,
|
| 1720 |
-
"step": 1400
|
| 1721 |
-
},
|
| 1722 |
-
{
|
| 1723 |
-
"epoch": 3.11,
|
| 1724 |
-
"learning_rate": 0.00012476342672775167,
|
| 1725 |
-
"loss": 2.2297,
|
| 1726 |
-
"step": 1405
|
| 1727 |
-
},
|
| 1728 |
-
{
|
| 1729 |
-
"epoch": 3.12,
|
| 1730 |
-
"learning_rate": 0.00012336087931764827,
|
| 1731 |
-
"loss": 2.5308,
|
| 1732 |
-
"step": 1410
|
| 1733 |
-
},
|
| 1734 |
-
{
|
| 1735 |
-
"epoch": 3.13,
|
| 1736 |
-
"learning_rate": 0.00012189220326791867,
|
| 1737 |
-
"loss": 2.4956,
|
| 1738 |
-
"step": 1415
|
| 1739 |
-
},
|
| 1740 |
-
{
|
| 1741 |
-
"epoch": 3.14,
|
| 1742 |
-
"learning_rate": 0.00012035917213571493,
|
| 1743 |
-
"loss": 2.259,
|
| 1744 |
-
"step": 1420
|
| 1745 |
-
},
|
| 1746 |
-
{
|
| 1747 |
-
"epoch": 3.15,
|
| 1748 |
-
"learning_rate": 0.0001187636371926859,
|
| 1749 |
-
"loss": 2.5147,
|
| 1750 |
-
"step": 1425
|
| 1751 |
-
},
|
| 1752 |
-
{
|
| 1753 |
-
"epoch": 3.16,
|
| 1754 |
-
"learning_rate": 0.00011710752518939737,
|
| 1755 |
-
"loss": 2.5795,
|
| 1756 |
-
"step": 1430
|
| 1757 |
-
},
|
| 1758 |
-
{
|
| 1759 |
-
"epoch": 3.17,
|
| 1760 |
-
"learning_rate": 0.00011539283602861218,
|
| 1761 |
-
"loss": 2.4607,
|
| 1762 |
-
"step": 1435
|
| 1763 |
-
},
|
| 1764 |
-
{
|
| 1765 |
-
"epoch": 3.19,
|
| 1766 |
-
"learning_rate": 0.00011362164035022977,
|
| 1767 |
-
"loss": 2.4427,
|
| 1768 |
-
"step": 1440
|
| 1769 |
-
},
|
| 1770 |
-
{
|
| 1771 |
-
"epoch": 3.2,
|
| 1772 |
-
"learning_rate": 0.00011179607703080082,
|
| 1773 |
-
"loss": 2.3976,
|
| 1774 |
-
"step": 1445
|
| 1775 |
-
},
|
| 1776 |
-
{
|
| 1777 |
-
"epoch": 3.21,
|
| 1778 |
-
"learning_rate": 0.00010991835060064675,
|
| 1779 |
-
"loss": 2.3127,
|
| 1780 |
-
"step": 1450
|
| 1781 |
-
},
|
| 1782 |
-
{
|
| 1783 |
-
"epoch": 3.22,
|
| 1784 |
-
"learning_rate": 0.00010799072858169134,
|
| 1785 |
-
"loss": 2.1402,
|
| 1786 |
-
"step": 1455
|
| 1787 |
-
},
|
| 1788 |
-
{
|
| 1789 |
-
"epoch": 3.23,
|
| 1790 |
-
"learning_rate": 0.0001060155387492263,
|
| 1791 |
-
"loss": 2.319,
|
| 1792 |
-
"step": 1460
|
| 1793 |
-
},
|
| 1794 |
-
{
|
| 1795 |
-
"epoch": 3.24,
|
| 1796 |
-
"learning_rate": 0.00010399516632091525,
|
| 1797 |
-
"loss": 2.6207,
|
| 1798 |
-
"step": 1465
|
| 1799 |
-
},
|
| 1800 |
-
{
|
| 1801 |
-
"epoch": 3.25,
|
| 1802 |
-
"learning_rate": 0.000101932051076431,
|
| 1803 |
-
"loss": 2.5146,
|
| 1804 |
-
"step": 1470
|
| 1805 |
-
},
|
| 1806 |
-
{
|
| 1807 |
-
"epoch": 3.26,
|
| 1808 |
-
"learning_rate": 9.982868441120418e-05,
|
| 1809 |
-
"loss": 2.513,
|
| 1810 |
-
"step": 1475
|
| 1811 |
-
},
|
| 1812 |
-
{
|
| 1813 |
-
"epoch": 3.27,
|
| 1814 |
-
"learning_rate": 9.768760632784142e-05,
|
| 1815 |
-
"loss": 2.5331,
|
| 1816 |
-
"step": 1480
|
| 1817 |
-
},
|
| 1818 |
-
{
|
| 1819 |
-
"epoch": 3.29,
|
| 1820 |
-
"learning_rate": 9.551140236884636e-05,
|
| 1821 |
-
"loss": 2.283,
|
| 1822 |
-
"step": 1485
|
| 1823 |
-
},
|
| 1824 |
-
{
|
| 1825 |
-
"epoch": 3.3,
|
| 1826 |
-
"learning_rate": 9.330270049434448e-05,
|
| 1827 |
-
"loss": 2.5398,
|
| 1828 |
-
"step": 1490
|
| 1829 |
-
},
|
| 1830 |
-
{
|
| 1831 |
-
"epoch": 3.31,
|
| 1832 |
-
"learning_rate": 9.106416790859171e-05,
|
| 1833 |
-
"loss": 2.4427,
|
| 1834 |
-
"step": 1495
|
| 1835 |
-
},
|
| 1836 |
-
{
|
| 1837 |
-
"epoch": 3.32,
|
| 1838 |
-
"learning_rate": 8.879850783908373e-05,
|
| 1839 |
-
"loss": 2.4624,
|
| 1840 |
-
"step": 1500
|
| 1841 |
-
},
|
| 1842 |
-
{
|
| 1843 |
-
"epoch": 3.33,
|
| 1844 |
-
"learning_rate": 8.650845627216773e-05,
|
| 1845 |
-
"loss": 2.3922,
|
| 1846 |
-
"step": 1505
|
| 1847 |
-
},
|
| 1848 |
-
{
|
| 1849 |
-
"epoch": 3.34,
|
| 1850 |
-
"learning_rate": 8.419677864910008e-05,
|
| 1851 |
-
"loss": 2.3814,
|
| 1852 |
-
"step": 1510
|
| 1853 |
-
},
|
| 1854 |
-
{
|
| 1855 |
-
"epoch": 3.35,
|
| 1856 |
-
"learning_rate": 8.186626652652247e-05,
|
| 1857 |
-
"loss": 2.4956,
|
| 1858 |
-
"step": 1515
|
| 1859 |
-
},
|
| 1860 |
-
{
|
| 1861 |
-
"epoch": 3.36,
|
| 1862 |
-
"learning_rate": 7.951973420541575e-05,
|
| 1863 |
-
"loss": 2.4434,
|
| 1864 |
-
"step": 1520
|
| 1865 |
-
},
|
| 1866 |
-
{
|
| 1867 |
-
"epoch": 3.37,
|
| 1868 |
-
"learning_rate": 7.716001533257615e-05,
|
| 1869 |
-
"loss": 2.2408,
|
| 1870 |
-
"step": 1525
|
| 1871 |
-
},
|
| 1872 |
-
{
|
| 1873 |
-
"epoch": 3.38,
|
| 1874 |
-
"learning_rate": 7.478995947873418e-05,
|
| 1875 |
-
"loss": 2.333,
|
| 1876 |
-
"step": 1530
|
| 1877 |
-
},
|
| 1878 |
-
{
|
| 1879 |
-
"epoch": 3.4,
|
| 1880 |
-
"learning_rate": 7.241242869744915e-05,
|
| 1881 |
-
"loss": 2.412,
|
| 1882 |
-
"step": 1535
|
| 1883 |
-
},
|
| 1884 |
-
{
|
| 1885 |
-
"epoch": 3.41,
|
| 1886 |
-
"learning_rate": 7.00302940689176e-05,
|
| 1887 |
-
"loss": 2.1938,
|
| 1888 |
-
"step": 1540
|
| 1889 |
-
},
|
| 1890 |
-
{
|
| 1891 |
-
"epoch": 3.42,
|
| 1892 |
-
"learning_rate": 6.764643223289375e-05,
|
| 1893 |
-
"loss": 2.4952,
|
| 1894 |
-
"step": 1545
|
| 1895 |
-
},
|
| 1896 |
-
{
|
| 1897 |
-
"epoch": 3.43,
|
| 1898 |
-
"learning_rate": 6.526372191488815e-05,
|
| 1899 |
-
"loss": 2.346,
|
| 1900 |
-
"step": 1550
|
| 1901 |
-
},
|
| 1902 |
-
{
|
| 1903 |
-
"epoch": 3.44,
|
| 1904 |
-
"learning_rate": 6.288504044985084e-05,
|
| 1905 |
-
"loss": 2.4523,
|
| 1906 |
-
"step": 1555
|
| 1907 |
-
},
|
| 1908 |
-
{
|
| 1909 |
-
"epoch": 3.45,
|
| 1910 |
-
"learning_rate": 6.051326030753298e-05,
|
| 1911 |
-
"loss": 2.3246,
|
| 1912 |
-
"step": 1560
|
| 1913 |
-
},
|
| 1914 |
-
{
|
| 1915 |
-
"epoch": 3.46,
|
| 1916 |
-
"learning_rate": 5.815124562372384e-05,
|
| 1917 |
-
"loss": 2.3573,
|
| 1918 |
-
"step": 1565
|
| 1919 |
-
},
|
| 1920 |
-
{
|
| 1921 |
-
"epoch": 3.47,
|
| 1922 |
-
"learning_rate": 5.580184874155164e-05,
|
| 1923 |
-
"loss": 2.5265,
|
| 1924 |
-
"step": 1570
|
| 1925 |
-
},
|
| 1926 |
-
{
|
| 1927 |
-
"epoch": 3.48,
|
| 1928 |
-
"learning_rate": 5.346790676702522e-05,
|
| 1929 |
-
"loss": 2.4817,
|
| 1930 |
-
"step": 1575
|
| 1931 |
-
},
|
| 1932 |
-
{
|
| 1933 |
-
"epoch": 3.5,
|
| 1934 |
-
"learning_rate": 5.115223814297577e-05,
|
| 1935 |
-
"loss": 2.3745,
|
| 1936 |
-
"step": 1580
|
| 1937 |
-
},
|
| 1938 |
-
{
|
| 1939 |
-
"epoch": 3.51,
|
| 1940 |
-
"learning_rate": 4.885763924553609e-05,
|
| 1941 |
-
"loss": 2.3791,
|
| 1942 |
-
"step": 1585
|
| 1943 |
-
},
|
| 1944 |
-
{
|
| 1945 |
-
"epoch": 3.52,
|
| 1946 |
-
"learning_rate": 4.658688100726727e-05,
|
| 1947 |
-
"loss": 2.438,
|
| 1948 |
-
"step": 1590
|
| 1949 |
-
},
|
| 1950 |
-
{
|
| 1951 |
-
"epoch": 3.53,
|
| 1952 |
-
"learning_rate": 4.434270557101133e-05,
|
| 1953 |
-
"loss": 2.2344,
|
| 1954 |
-
"step": 1595
|
| 1955 |
-
},
|
| 1956 |
-
{
|
| 1957 |
-
"epoch": 3.54,
|
| 1958 |
-
"learning_rate": 4.212782297850713e-05,
|
| 1959 |
-
"loss": 2.3451,
|
| 1960 |
-
"step": 1600
|
| 1961 |
-
},
|
| 1962 |
-
{
|
| 1963 |
-
"epoch": 3.55,
|
| 1964 |
-
"learning_rate": 3.9944907897778607e-05,
|
| 1965 |
-
"loss": 2.4611,
|
| 1966 |
-
"step": 1605
|
| 1967 |
-
},
|
| 1968 |
-
{
|
| 1969 |
-
"epoch": 3.56,
|
| 1970 |
-
"learning_rate": 3.7796596393228573e-05,
|
| 1971 |
-
"loss": 2.4182,
|
| 1972 |
-
"step": 1610
|
| 1973 |
-
},
|
| 1974 |
-
{
|
| 1975 |
-
"epoch": 3.57,
|
| 1976 |
-
"learning_rate": 3.568548274236061e-05,
|
| 1977 |
-
"loss": 2.5135,
|
| 1978 |
-
"step": 1615
|
| 1979 |
-
},
|
| 1980 |
-
{
|
| 1981 |
-
"epoch": 3.58,
|
| 1982 |
-
"learning_rate": 3.361411630295785e-05,
|
| 1983 |
-
"loss": 2.5817,
|
| 1984 |
-
"step": 1620
|
| 1985 |
-
},
|
| 1986 |
-
{
|
| 1987 |
-
"epoch": 3.6,
|
| 1988 |
-
"learning_rate": 3.158499843450252e-05,
|
| 1989 |
-
"loss": 2.3341,
|
| 1990 |
-
"step": 1625
|
| 1991 |
-
},
|
| 1992 |
-
{
|
| 1993 |
-
"epoch": 3.61,
|
| 1994 |
-
"learning_rate": 2.9600579477565787e-05,
|
| 1995 |
-
"loss": 2.3179,
|
| 1996 |
-
"step": 1630
|
| 1997 |
-
},
|
| 1998 |
-
{
|
| 1999 |
-
"epoch": 3.62,
|
| 2000 |
-
"learning_rate": 2.7663255794801203e-05,
|
| 2001 |
-
"loss": 2.29,
|
| 2002 |
-
"step": 1635
|
| 2003 |
-
},
|
| 2004 |
-
{
|
| 2005 |
-
"epoch": 3.63,
|
| 2006 |
-
"learning_rate": 2.57753668771237e-05,
|
| 2007 |
-
"loss": 2.5001,
|
| 2008 |
-
"step": 1640
|
| 2009 |
-
},
|
| 2010 |
-
{
|
| 2011 |
-
"epoch": 3.64,
|
| 2012 |
-
"learning_rate": 2.3939192518565938e-05,
|
| 2013 |
-
"loss": 2.5561,
|
| 2014 |
-
"step": 1645
|
| 2015 |
-
},
|
| 2016 |
-
{
|
| 2017 |
-
"epoch": 3.65,
|
| 2018 |
-
"learning_rate": 2.215695006322396e-05,
|
| 2019 |
-
"loss": 2.4063,
|
| 2020 |
-
"step": 1650
|
| 2021 |
-
},
|
| 2022 |
-
{
|
| 2023 |
-
"epoch": 3.66,
|
| 2024 |
-
"learning_rate": 2.0430791727616646e-05,
|
| 2025 |
-
"loss": 2.377,
|
| 2026 |
-
"step": 1655
|
| 2027 |
-
},
|
| 2028 |
-
{
|
| 2029 |
-
"epoch": 3.67,
|
| 2030 |
-
"learning_rate": 1.8762802001692604e-05,
|
| 2031 |
-
"loss": 2.6203,
|
| 2032 |
-
"step": 1660
|
| 2033 |
-
},
|
| 2034 |
-
{
|
| 2035 |
-
"epoch": 3.68,
|
| 2036 |
-
"learning_rate": 1.715499513162315e-05,
|
| 2037 |
-
"loss": 2.434,
|
| 2038 |
-
"step": 1665
|
| 2039 |
-
},
|
| 2040 |
-
{
|
| 2041 |
-
"epoch": 3.69,
|
| 2042 |
-
"learning_rate": 1.5609312687419068e-05,
|
| 2043 |
-
"loss": 2.2921,
|
| 2044 |
-
"step": 1670
|
| 2045 |
-
},
|
| 2046 |
-
{
|
| 2047 |
-
"epoch": 3.71,
|
| 2048 |
-
"learning_rate": 1.4127621218315069e-05,
|
| 2049 |
-
"loss": 2.2474,
|
| 2050 |
-
"step": 1675
|
| 2051 |
-
},
|
| 2052 |
-
{
|
| 2053 |
-
"epoch": 3.72,
|
| 2054 |
-
"learning_rate": 1.2711709998742125e-05,
|
| 2055 |
-
"loss": 2.4056,
|
| 2056 |
-
"step": 1680
|
| 2057 |
-
},
|
| 2058 |
-
{
|
| 2059 |
-
"epoch": 3.73,
|
| 2060 |
-
"learning_rate": 1.1363288867620807e-05,
|
| 2061 |
-
"loss": 2.1388,
|
| 2062 |
-
"step": 1685
|
| 2063 |
-
},
|
| 2064 |
-
{
|
| 2065 |
-
"epoch": 3.74,
|
| 2066 |
-
"learning_rate": 1.0083986163578448e-05,
|
| 2067 |
-
"loss": 2.4683,
|
| 2068 |
-
"step": 1690
|
| 2069 |
-
},
|
| 2070 |
-
{
|
| 2071 |
-
"epoch": 3.75,
|
| 2072 |
-
"learning_rate": 8.875346758583873e-06,
|
| 2073 |
-
"loss": 2.2937,
|
| 2074 |
-
"step": 1695
|
| 2075 |
-
},
|
| 2076 |
-
{
|
| 2077 |
-
"epoch": 3.76,
|
| 2078 |
-
"learning_rate": 7.73883019237949e-06,
|
| 2079 |
-
"loss": 2.4202,
|
| 2080 |
-
"step": 1700
|
| 2081 |
-
},
|
| 2082 |
-
{
|
| 2083 |
-
"epoch": 3.77,
|
| 2084 |
-
"learning_rate": 6.6758089099546965e-06,
|
| 2085 |
-
"loss": 2.5003,
|
| 2086 |
-
"step": 1705
|
| 2087 |
-
},
|
| 2088 |
-
{
|
| 2089 |
-
"epoch": 3.78,
|
| 2090 |
-
"learning_rate": 5.6875666041964786e-06,
|
| 2091 |
-
"loss": 2.3068,
|
| 2092 |
-
"step": 1710
|
| 2093 |
-
},
|
| 2094 |
-
{
|
| 2095 |
-
"epoch": 3.79,
|
| 2096 |
-
"learning_rate": 4.775296665718002e-06,
|
| 2097 |
-
"loss": 2.2147,
|
| 2098 |
-
"step": 1715
|
| 2099 |
-
},
|
| 2100 |
-
{
|
| 2101 |
-
"epoch": 3.81,
|
| 2102 |
-
"learning_rate": 3.940100741730249e-06,
|
| 2103 |
-
"loss": 2.4371,
|
| 2104 |
-
"step": 1720
|
| 2105 |
-
},
|
| 2106 |
-
{
|
| 2107 |
-
"epoch": 3.82,
|
| 2108 |
-
"learning_rate": 3.1829874057071e-06,
|
| 2109 |
-
"loss": 2.3255,
|
| 2110 |
-
"step": 1725
|
| 2111 |
-
},
|
| 2112 |
-
{
|
| 2113 |
-
"epoch": 3.83,
|
| 2114 |
-
"learning_rate": 2.504870939442012e-06,
|
| 2115 |
-
"loss": 2.3511,
|
| 2116 |
-
"step": 1730
|
| 2117 |
-
},
|
| 2118 |
-
{
|
| 2119 |
-
"epoch": 3.84,
|
| 2120 |
-
"learning_rate": 1.9065702289715435e-06,
|
| 2121 |
-
"loss": 2.587,
|
| 2122 |
-
"step": 1735
|
| 2123 |
-
},
|
| 2124 |
-
{
|
| 2125 |
-
"epoch": 3.85,
|
| 2126 |
-
"learning_rate": 1.3888077756974934e-06,
|
| 2127 |
-
"loss": 2.4066,
|
| 2128 |
-
"step": 1740
|
| 2129 |
-
},
|
| 2130 |
-
{
|
| 2131 |
-
"epoch": 3.86,
|
| 2132 |
-
"learning_rate": 9.522088239021213e-07,
|
| 2133 |
-
"loss": 2.3034,
|
| 2134 |
-
"step": 1745
|
| 2135 |
-
},
|
| 2136 |
-
{
|
| 2137 |
-
"epoch": 3.87,
|
| 2138 |
-
"learning_rate": 5.973006057099719e-07,
|
| 2139 |
-
"loss": 2.2354,
|
| 2140 |
-
"step": 1750
|
| 2141 |
-
},
|
| 2142 |
-
{
|
| 2143 |
-
"epoch": 3.88,
|
| 2144 |
-
"learning_rate": 3.2451170440812575e-07,
|
| 2145 |
-
"loss": 2.4044,
|
| 2146 |
-
"step": 1755
|
| 2147 |
-
},
|
| 2148 |
-
{
|
| 2149 |
-
"epoch": 3.89,
|
| 2150 |
-
"learning_rate": 1.3417153689369494e-07,
|
| 2151 |
-
"loss": 2.2793,
|
| 2152 |
-
"step": 1760
|
| 2153 |
-
},
|
| 2154 |
-
{
|
| 2155 |
-
"epoch": 3.9,
|
| 2156 |
-
"learning_rate": 2.650995587359215e-08,
|
| 2157 |
-
"loss": 2.3649,
|
| 2158 |
-
"step": 1765
|
| 2159 |
-
},
|
| 2160 |
-
{
|
| 2161 |
-
"epoch": 3.92,
|
| 2162 |
-
"learning_rate": 1.6569722969137058e-09,
|
| 2163 |
-
"loss": 2.5027,
|
| 2164 |
-
"step": 1770
|
| 2165 |
-
},
|
| 2166 |
-
{
|
| 2167 |
-
"epoch": 3.93,
|
| 2168 |
-
"learning_rate": 5.964259835509249e-08,
|
| 2169 |
-
"loss": 2.3313,
|
| 2170 |
-
"step": 1775
|
| 2171 |
-
},
|
| 2172 |
-
{
|
| 2173 |
-
"epoch": 3.94,
|
| 2174 |
-
"learning_rate": 2.0039681123962964e-07,
|
| 2175 |
-
"loss": 2.3823,
|
| 2176 |
-
"step": 1780
|
| 2177 |
-
},
|
| 2178 |
-
{
|
| 2179 |
-
"epoch": 3.95,
|
| 2180 |
-
"learning_rate": 4.2374963770060374e-07,
|
| 2181 |
-
"loss": 2.3593,
|
| 2182 |
-
"step": 1785
|
| 2183 |
-
},
|
| 2184 |
-
{
|
| 2185 |
-
"epoch": 3.96,
|
| 2186 |
-
"learning_rate": 7.294313593049221e-07,
|
| 2187 |
-
"loss": 2.4781,
|
| 2188 |
-
"step": 1790
|
| 2189 |
-
},
|
| 2190 |
-
{
|
| 2191 |
-
"epoch": 3.97,
|
| 2192 |
-
"learning_rate": 1.1170728381450536e-06,
|
| 2193 |
-
"loss": 2.4511,
|
| 2194 |
-
"step": 1795
|
| 2195 |
-
},
|
| 2196 |
-
{
|
| 2197 |
-
"epoch": 3.98,
|
| 2198 |
-
"learning_rate": 1.5862059626056908e-06,
|
| 2199 |
-
"loss": 2.0358,
|
| 2200 |
-
"step": 1800
|
| 2201 |
-
},
|
| 2202 |
-
{
|
| 2203 |
-
"epoch": 3.99,
|
| 2204 |
-
"learning_rate": 2.1362642126510735e-06,
|
| 2205 |
-
"loss": 2.4982,
|
| 2206 |
-
"step": 1805
|
| 2207 |
-
},
|
| 2208 |
-
{
|
| 2209 |
-
"epoch": 4.0,
|
| 2210 |
-
"eval_loss": 2.0775227546691895,
|
| 2211 |
-
"eval_runtime": 29.8853,
|
| 2212 |
-
"eval_samples_per_second": 22.386,
|
| 2213 |
-
"eval_steps_per_second": 2.811,
|
| 2214 |
-
"step": 1808
|
| 2215 |
-
},
|
| 2216 |
-
{
|
| 2217 |
-
"epoch": 3.97,
|
| 2218 |
-
"learning_rate": 1.9689817301781333e-07,
|
| 2219 |
-
"loss": 2.3835,
|
| 2220 |
-
"step": 1810
|
| 2221 |
-
},
|
| 2222 |
-
{
|
| 2223 |
-
"epoch": 3.98,
|
| 2224 |
-
"learning_rate": 5.860097463115006e-08,
|
| 2225 |
-
"loss": 2.2235,
|
| 2226 |
-
"step": 1815
|
| 2227 |
-
},
|
| 2228 |
-
{
|
| 2229 |
-
"epoch": 3.99,
|
| 2230 |
-
"learning_rate": 1.628030220611354e-09,
|
| 2231 |
-
"loss": 2.1558,
|
| 2232 |
-
"step": 1820
|
| 2233 |
-
},
|
| 2234 |
-
{
|
| 2235 |
-
"epoch": 4.0,
|
| 2236 |
-
"eval_loss": 1.9903991222381592,
|
| 2237 |
-
"eval_runtime": 27.7097,
|
| 2238 |
-
"eval_samples_per_second": 22.844,
|
| 2239 |
-
"eval_steps_per_second": 2.887,
|
| 2240 |
-
"step": 1824
|
| 2241 |
}
|
| 2242 |
],
|
| 2243 |
-
"max_steps":
|
| 2244 |
"num_train_epochs": 5,
|
| 2245 |
-
"total_flos":
|
| 2246 |
"trial_name": null,
|
| 2247 |
"trial_params": null
|
| 2248 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 2.026026487350464,
|
| 3 |
+
"best_model_checkpoint": "output/eminem/checkpoint-452",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 452,
|
| 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": 0.00013401701259429288,
|
| 13 |
+
"loss": 2.1682,
|
| 14 |
"step": 5
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 0.0001332598992582699,
|
| 19 |
+
"loss": 2.1572,
|
| 20 |
"step": 10
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.03,
|
| 24 |
+
"learning_rate": 0.00013242470333428216,
|
| 25 |
+
"loss": 2.2158,
|
| 26 |
"step": 15
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.04,
|
| 30 |
+
"learning_rate": 0.0001315124333958035,
|
| 31 |
+
"loss": 2.1302,
|
| 32 |
"step": 20
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 0.06,
|
| 36 |
+
"learning_rate": 0.0001305241910900455,
|
| 37 |
+
"loss": 2.2621,
|
| 38 |
"step": 25
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 0.0001294611698076205,
|
| 43 |
+
"loss": 2.1793,
|
| 44 |
"step": 30
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.08,
|
| 48 |
+
"learning_rate": 0.00012832465324141612,
|
| 49 |
+
"loss": 2.2551,
|
| 50 |
"step": 35
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.09,
|
| 54 |
+
"learning_rate": 0.0001271160138364218,
|
| 55 |
+
"loss": 2.1327,
|
| 56 |
"step": 40
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.1,
|
| 60 |
+
"learning_rate": 0.00012583671113237919,
|
| 61 |
+
"loss": 2.189,
|
| 62 |
"step": 45
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.11,
|
| 66 |
+
"learning_rate": 0.00012448829000125786,
|
| 67 |
+
"loss": 2.3515,
|
| 68 |
"step": 50
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.12,
|
| 72 |
+
"learning_rate": 0.00012307237878168522,
|
| 73 |
+
"loss": 2.2902,
|
| 74 |
"step": 55
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.13,
|
| 78 |
+
"learning_rate": 0.00012159068731258093,
|
| 79 |
+
"loss": 2.1917,
|
| 80 |
"step": 60
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.14,
|
| 84 |
+
"learning_rate": 0.00012004500486837715,
|
| 85 |
+
"loss": 2.0402,
|
| 86 |
"step": 65
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.15,
|
| 90 |
+
"learning_rate": 0.00011843719799830738,
|
| 91 |
+
"loss": 2.3137,
|
| 92 |
"step": 70
|
| 93 |
},
|
| 94 |
{
|
| 95 |
+
"epoch": 0.17,
|
| 96 |
+
"learning_rate": 0.00011676920827238334,
|
| 97 |
+
"loss": 2.1433,
|
| 98 |
"step": 75
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.18,
|
| 102 |
+
"learning_rate": 0.00011504304993677638,
|
| 103 |
+
"loss": 2.1509,
|
| 104 |
"step": 80
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.19,
|
| 108 |
+
"learning_rate": 0.00011326080748143405,
|
| 109 |
+
"loss": 2.1987,
|
| 110 |
"step": 85
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.2,
|
| 114 |
+
"learning_rate": 0.00011142463312287667,
|
| 115 |
+
"loss": 2.1876,
|
| 116 |
"step": 90
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.21,
|
| 120 |
+
"learning_rate": 0.00010953674420519918,
|
| 121 |
+
"loss": 2.3293,
|
| 122 |
"step": 95
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.22,
|
| 126 |
+
"learning_rate": 0.00010759942052243419,
|
| 127 |
+
"loss": 2.307,
|
| 128 |
"step": 100
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.23,
|
| 132 |
+
"learning_rate": 0.00010561500156549789,
|
| 133 |
+
"loss": 2.3729,
|
| 134 |
"step": 105
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.24,
|
| 138 |
+
"learning_rate": 0.00010358588369704255,
|
| 139 |
+
"loss": 2.3181,
|
| 140 |
"step": 110
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.25,
|
| 144 |
+
"learning_rate": 0.00010151451725763938,
|
| 145 |
+
"loss": 2.2169,
|
| 146 |
"step": 115
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 0.27,
|
| 150 |
+
"learning_rate": 9.940340360677185e-05,
|
| 151 |
+
"loss": 2.1579,
|
| 152 |
"step": 120
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 0.28,
|
| 156 |
+
"learning_rate": 9.725509210222182e-05,
|
| 157 |
+
"loss": 2.0936,
|
| 158 |
"step": 125
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.29,
|
| 162 |
+
"learning_rate": 9.507217702149286e-05,
|
| 163 |
+
"loss": 2.2298,
|
| 164 |
"step": 130
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.3,
|
| 168 |
+
"learning_rate": 9.28572944289891e-05,
|
| 169 |
+
"loss": 2.1296,
|
| 170 |
"step": 135
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.31,
|
| 174 |
+
"learning_rate": 9.061311899273271e-05,
|
| 175 |
+
"loss": 2.146,
|
| 176 |
"step": 140
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.32,
|
| 180 |
+
"learning_rate": 8.83423607544639e-05,
|
| 181 |
+
"loss": 2.2279,
|
| 182 |
"step": 145
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.33,
|
| 186 |
+
"learning_rate": 8.604776185702469e-05,
|
| 187 |
+
"loss": 2.1469,
|
| 188 |
"step": 150
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.34,
|
| 192 |
+
"learning_rate": 8.373209323297477e-05,
|
| 193 |
+
"loss": 2.2793,
|
| 194 |
"step": 155
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.35,
|
| 198 |
+
"learning_rate": 8.139815125844835e-05,
|
| 199 |
+
"loss": 1.9848,
|
| 200 |
"step": 160
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 0.37,
|
| 204 |
+
"learning_rate": 7.904875437627664e-05,
|
| 205 |
+
"loss": 2.6294,
|
| 206 |
"step": 165
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 0.38,
|
| 210 |
+
"learning_rate": 7.6686739692467e-05,
|
| 211 |
+
"loss": 2.2562,
|
| 212 |
"step": 170
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.39,
|
| 216 |
+
"learning_rate": 7.431495955014962e-05,
|
| 217 |
+
"loss": 2.337,
|
| 218 |
"step": 175
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.4,
|
| 222 |
+
"learning_rate": 7.193627808511231e-05,
|
| 223 |
+
"loss": 2.3747,
|
| 224 |
"step": 180
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.41,
|
| 228 |
+
"learning_rate": 6.955356776710623e-05,
|
| 229 |
+
"loss": 2.2054,
|
| 230 |
"step": 185
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.42,
|
| 234 |
+
"learning_rate": 6.716970593108287e-05,
|
| 235 |
+
"loss": 2.021,
|
| 236 |
"step": 190
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.43,
|
| 240 |
+
"learning_rate": 6.478757130255133e-05,
|
| 241 |
+
"loss": 2.3831,
|
| 242 |
"step": 195
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.44,
|
| 246 |
+
"learning_rate": 6.24100405212658e-05,
|
| 247 |
+
"loss": 2.3325,
|
| 248 |
"step": 200
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.45,
|
| 252 |
+
"learning_rate": 6.003998466742432e-05,
|
| 253 |
+
"loss": 2.0344,
|
| 254 |
"step": 205
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.46,
|
| 258 |
+
"learning_rate": 5.768026579458423e-05,
|
| 259 |
+
"loss": 2.1013,
|
| 260 |
"step": 210
|
| 261 |
},
|
| 262 |
{
|
| 263 |
+
"epoch": 0.48,
|
| 264 |
+
"learning_rate": 5.5333733473477525e-05,
|
| 265 |
+
"loss": 2.1656,
|
| 266 |
"step": 215
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 0.49,
|
| 270 |
+
"learning_rate": 5.300322135090038e-05,
|
| 271 |
+
"loss": 2.2654,
|
| 272 |
"step": 220
|
| 273 |
},
|
| 274 |
{
|
| 275 |
+
"epoch": 0.5,
|
| 276 |
+
"learning_rate": 5.069154372783225e-05,
|
| 277 |
+
"loss": 2.0446,
|
| 278 |
"step": 225
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.51,
|
| 282 |
+
"learning_rate": 4.840149216091673e-05,
|
| 283 |
+
"loss": 2.347,
|
| 284 |
"step": 230
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 0.52,
|
| 288 |
+
"learning_rate": 4.613583209140827e-05,
|
| 289 |
+
"loss": 2.0384,
|
| 290 |
"step": 235
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 0.53,
|
| 294 |
+
"learning_rate": 4.3897299505655516e-05,
|
| 295 |
+
"loss": 2.4872,
|
| 296 |
"step": 240
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.54,
|
| 300 |
+
"learning_rate": 4.1688597631154076e-05,
|
| 301 |
+
"loss": 2.2689,
|
| 302 |
"step": 245
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.55,
|
| 306 |
+
"learning_rate": 3.9512393672158556e-05,
|
| 307 |
+
"loss": 2.2838,
|
| 308 |
"step": 250
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.56,
|
| 312 |
+
"learning_rate": 3.7371315588795804e-05,
|
| 313 |
+
"loss": 2.0626,
|
| 314 |
"step": 255
|
| 315 |
},
|
| 316 |
{
|
| 317 |
+
"epoch": 0.58,
|
| 318 |
+
"learning_rate": 3.5267948923569414e-05,
|
| 319 |
+
"loss": 2.1429,
|
| 320 |
"step": 260
|
| 321 |
},
|
| 322 |
{
|
| 323 |
+
"epoch": 0.59,
|
| 324 |
+
"learning_rate": 3.320483367908473e-05,
|
| 325 |
+
"loss": 2.3835,
|
| 326 |
"step": 265
|
| 327 |
},
|
| 328 |
{
|
| 329 |
+
"epoch": 0.6,
|
| 330 |
+
"learning_rate": 3.118446125077409e-05,
|
| 331 |
+
"loss": 2.2844,
|
| 332 |
"step": 270
|
| 333 |
},
|
| 334 |
{
|
| 335 |
+
"epoch": 0.61,
|
| 336 |
+
"learning_rate": 2.9209271418309033e-05,
|
| 337 |
+
"loss": 2.1332,
|
| 338 |
"step": 275
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.62,
|
| 342 |
+
"learning_rate": 2.7281649399353233e-05,
|
| 343 |
+
"loss": 2.0147,
|
| 344 |
"step": 280
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.63,
|
| 348 |
+
"learning_rate": 2.5403922969199553e-05,
|
| 349 |
+
"loss": 2.0417,
|
| 350 |
"step": 285
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.64,
|
| 354 |
+
"learning_rate": 2.357835964977058e-05,
|
| 355 |
+
"loss": 2.2009,
|
| 356 |
"step": 290
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 0.65,
|
| 360 |
+
"learning_rate": 2.180716397138782e-05,
|
| 361 |
+
"loss": 2.1734,
|
| 362 |
"step": 295
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 0.66,
|
| 366 |
+
"learning_rate": 2.0092474810602975e-05,
|
| 367 |
+
"loss": 2.2568,
|
| 368 |
"step": 300
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 0.67,
|
| 372 |
+
"learning_rate": 1.843636280731442e-05,
|
| 373 |
+
"loss": 2.1657,
|
| 374 |
"step": 305
|
| 375 |
},
|
| 376 |
{
|
| 377 |
+
"epoch": 0.69,
|
| 378 |
+
"learning_rate": 1.6840827864285067e-05,
|
| 379 |
+
"loss": 2.1509,
|
| 380 |
"step": 310
|
| 381 |
},
|
| 382 |
{
|
| 383 |
+
"epoch": 0.7,
|
| 384 |
+
"learning_rate": 1.5307796732081633e-05,
|
| 385 |
+
"loss": 2.2271,
|
| 386 |
"step": 315
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 0.71,
|
| 390 |
+
"learning_rate": 1.3839120682351706e-05,
|
| 391 |
+
"loss": 2.2315,
|
| 392 |
"step": 320
|
| 393 |
},
|
| 394 |
{
|
| 395 |
+
"epoch": 0.72,
|
| 396 |
+
"learning_rate": 1.2436573272248327e-05,
|
| 397 |
+
"loss": 2.2438,
|
| 398 |
"step": 325
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 0.73,
|
| 402 |
+
"learning_rate": 1.1101848202701753e-05,
|
| 403 |
+
"loss": 2.2333,
|
| 404 |
"step": 330
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 0.74,
|
| 408 |
+
"learning_rate": 9.83655727312471e-06,
|
| 409 |
+
"loss": 2.3438,
|
| 410 |
"step": 335
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 0.75,
|
| 414 |
+
"learning_rate": 8.642228435021026e-06,
|
| 415 |
+
"loss": 1.9256,
|
| 416 |
"step": 340
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 0.76,
|
| 420 |
+
"learning_rate": 7.520303946848332e-06,
|
| 421 |
+
"loss": 2.1889,
|
| 422 |
"step": 345
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.77,
|
| 426 |
+
"learning_rate": 6.472138632361373e-06,
|
| 427 |
+
"loss": 2.1674,
|
| 428 |
"step": 350
|
| 429 |
},
|
| 430 |
{
|
| 431 |
+
"epoch": 0.79,
|
| 432 |
+
"learning_rate": 5.49899824454404e-06,
|
| 433 |
+
"loss": 2.099,
|
| 434 |
"step": 355
|
| 435 |
},
|
| 436 |
{
|
| 437 |
+
"epoch": 0.8,
|
| 438 |
+
"learning_rate": 4.602057937096952e-06,
|
| 439 |
+
"loss": 2.1071,
|
| 440 |
"step": 360
|
| 441 |
},
|
| 442 |
{
|
| 443 |
+
"epoch": 0.81,
|
| 444 |
+
"learning_rate": 3.7824008453367434e-06,
|
| 445 |
+
"loss": 2.2894,
|
| 446 |
"step": 365
|
| 447 |
},
|
| 448 |
{
|
| 449 |
+
"epoch": 0.82,
|
| 450 |
+
"learning_rate": 3.041016778213552e-06,
|
| 451 |
+
"loss": 2.3941,
|
| 452 |
"step": 370
|
| 453 |
},
|
| 454 |
{
|
| 455 |
+
"epoch": 0.83,
|
| 456 |
+
"learning_rate": 2.3788010230270542e-06,
|
| 457 |
+
"loss": 2.2867,
|
| 458 |
"step": 375
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 0.84,
|
| 462 |
+
"learning_rate": 1.7965532642889826e-06,
|
| 463 |
+
"loss": 1.972,
|
| 464 |
"step": 380
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.85,
|
| 468 |
+
"learning_rate": 1.2949766180319986e-06,
|
| 469 |
+
"loss": 2.2289,
|
| 470 |
"step": 385
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 0.86,
|
| 474 |
+
"learning_rate": 8.746767827345464e-07,
|
| 475 |
+
"loss": 2.212,
|
| 476 |
"step": 390
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 0.87,
|
| 480 |
+
"learning_rate": 5.36161307886377e-07,
|
| 481 |
+
"loss": 2.1037,
|
| 482 |
"step": 395
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 0.88,
|
| 486 |
+
"learning_rate": 2.798389810765762e-07,
|
| 487 |
+
"loss": 2.0082,
|
| 488 |
"step": 400
|
| 489 |
},
|
| 490 |
{
|
| 491 |
+
"epoch": 0.9,
|
| 492 |
+
"learning_rate": 1.0601933434685087e-07,
|
| 493 |
+
"loss": 2.1227,
|
| 494 |
"step": 405
|
| 495 |
},
|
| 496 |
{
|
| 497 |
+
"epoch": 0.91,
|
| 498 |
+
"learning_rate": 1.4912270403714612e-08,
|
| 499 |
+
"loss": 2.279,
|
| 500 |
"step": 410
|
| 501 |
},
|
| 502 |
{
|
| 503 |
+
"epoch": 0.92,
|
| 504 |
+
"learning_rate": 6.627809142258534e-09,
|
| 505 |
+
"loss": 2.1128,
|
| 506 |
"step": 415
|
| 507 |
},
|
| 508 |
{
|
| 509 |
+
"epoch": 0.93,
|
| 510 |
+
"learning_rate": 8.11759547874179e-08,
|
| 511 |
+
"loss": 2.1338,
|
| 512 |
"step": 420
|
| 513 |
},
|
| 514 |
{
|
| 515 |
+
"epoch": 0.94,
|
| 516 |
+
"learning_rate": 2.384666838129088e-07,
|
| 517 |
+
"loss": 2.1859,
|
| 518 |
"step": 425
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 0.95,
|
| 522 |
+
"learning_rate": 4.783100536526866e-07,
|
| 523 |
+
"loss": 2.1442,
|
| 524 |
"step": 430
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 0.96,
|
| 528 |
+
"learning_rate": 8.004164320735097e-07,
|
| 529 |
+
"loss": 2.3821,
|
| 530 |
"step": 435
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 0.97,
|
| 534 |
+
"learning_rate": 1.2043968469315794e-06,
|
| 535 |
+
"loss": 2.1965,
|
| 536 |
"step": 440
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 0.98,
|
| 540 |
+
"learning_rate": 1.6897634558914746e-06,
|
| 541 |
+
"loss": 2.2908,
|
| 542 |
"step": 445
|
| 543 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 544 |
{
|
| 545 |
"epoch": 1.0,
|
| 546 |
+
"learning_rate": 2.2559301355383585e-06,
|
| 547 |
+
"loss": 2.1072,
|
| 548 |
+
"step": 450
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 1.0,
|
| 552 |
+
"eval_loss": 2.026026487350464,
|
| 553 |
+
"eval_runtime": 32.0027,
|
| 554 |
+
"eval_samples_per_second": 20.779,
|
| 555 |
+
"eval_steps_per_second": 2.625,
|
| 556 |
+
"step": 452
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 557 |
}
|
| 558 |
],
|
| 559 |
+
"max_steps": 2260,
|
| 560 |
"num_train_epochs": 5,
|
| 561 |
+
"total_flos": 472024055808000.0,
|
| 562 |
"trial_name": null,
|
| 563 |
"trial_params": null
|
| 564 |
}
|
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 sha256:42cd89b75d7c0f3fe6da69fe6e477354496b002c63ce374ba6ee57264a045ed7
|
| 3 |
size 2671
|