Training in progress, step 150, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d9d2e3782fc6fe40bc3b1168bb0a1e565dab83c2e38644c8cca45ee7ae2e7e02
|
| 3 |
size 639691872
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 325339796
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9f7cca1f606b34df1171c2dbe2f2178cd69b1908ea392812b8e4381be79ffa2
|
| 3 |
size 325339796
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:46260221d6c51c5804e4a1725ef2c4b4a5bedc2ba8378b0e9f3ebf0455d9616f
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d8ce05761f46e7cf72fb17a02e3a0ca15c9d25ce3babf590eeb40568923b8bac
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -373,6 +373,722 @@
|
|
| 373 |
"eval_samples_per_second": 12.706,
|
| 374 |
"eval_steps_per_second": 3.177,
|
| 375 |
"step": 50
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 376 |
}
|
| 377 |
],
|
| 378 |
"logging_steps": 1,
|
|
@@ -401,7 +1117,7 @@
|
|
| 401 |
"attributes": {}
|
| 402 |
}
|
| 403 |
},
|
| 404 |
-
"total_flos":
|
| 405 |
"train_batch_size": 8,
|
| 406 |
"trial_name": null,
|
| 407 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.3852198123931885,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-150",
|
| 4 |
+
"epoch": 0.09776763891152028,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 150,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 373 |
"eval_samples_per_second": 12.706,
|
| 374 |
"eval_steps_per_second": 3.177,
|
| 375 |
"step": 50
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.0332409972299169,
|
| 379 |
+
"grad_norm": 0.5773328542709351,
|
| 380 |
+
"learning_rate": 8.894386393810563e-05,
|
| 381 |
+
"loss": 1.1741,
|
| 382 |
+
"step": 51
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.033892781489327034,
|
| 386 |
+
"grad_norm": 0.569870114326477,
|
| 387 |
+
"learning_rate": 8.842005554284296e-05,
|
| 388 |
+
"loss": 1.1699,
|
| 389 |
+
"step": 52
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.03454456574873717,
|
| 393 |
+
"grad_norm": 0.5505066514015198,
|
| 394 |
+
"learning_rate": 8.788574348801675e-05,
|
| 395 |
+
"loss": 1.3949,
|
| 396 |
+
"step": 53
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.035196350008147304,
|
| 400 |
+
"grad_norm": 0.5093855261802673,
|
| 401 |
+
"learning_rate": 8.73410738492077e-05,
|
| 402 |
+
"loss": 1.2573,
|
| 403 |
+
"step": 54
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.03584813426755744,
|
| 407 |
+
"grad_norm": 0.6201264262199402,
|
| 408 |
+
"learning_rate": 8.678619553365659e-05,
|
| 409 |
+
"loss": 1.2551,
|
| 410 |
+
"step": 55
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.03649991852696757,
|
| 414 |
+
"grad_norm": 0.6050621867179871,
|
| 415 |
+
"learning_rate": 8.622126023955446e-05,
|
| 416 |
+
"loss": 1.3761,
|
| 417 |
+
"step": 56
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.03715170278637771,
|
| 421 |
+
"grad_norm": 0.4428724944591522,
|
| 422 |
+
"learning_rate": 8.564642241456986e-05,
|
| 423 |
+
"loss": 1.1776,
|
| 424 |
+
"step": 57
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.03780348704578784,
|
| 428 |
+
"grad_norm": 0.5018085837364197,
|
| 429 |
+
"learning_rate": 8.506183921362443e-05,
|
| 430 |
+
"loss": 1.2878,
|
| 431 |
+
"step": 58
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.03845527130519798,
|
| 435 |
+
"grad_norm": 0.4629620313644409,
|
| 436 |
+
"learning_rate": 8.44676704559283e-05,
|
| 437 |
+
"loss": 1.4328,
|
| 438 |
+
"step": 59
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 0.03910705556460811,
|
| 442 |
+
"grad_norm": 0.47736626863479614,
|
| 443 |
+
"learning_rate": 8.386407858128706e-05,
|
| 444 |
+
"loss": 1.4597,
|
| 445 |
+
"step": 60
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.03975883982401825,
|
| 449 |
+
"grad_norm": 0.46020087599754333,
|
| 450 |
+
"learning_rate": 8.32512286056924e-05,
|
| 451 |
+
"loss": 1.496,
|
| 452 |
+
"step": 61
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.04041062408342839,
|
| 456 |
+
"grad_norm": 0.5311948657035828,
|
| 457 |
+
"learning_rate": 8.262928807620843e-05,
|
| 458 |
+
"loss": 1.5641,
|
| 459 |
+
"step": 62
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.04106240834283852,
|
| 463 |
+
"grad_norm": 0.5824150443077087,
|
| 464 |
+
"learning_rate": 8.199842702516583e-05,
|
| 465 |
+
"loss": 1.3796,
|
| 466 |
+
"step": 63
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.04171419260224866,
|
| 470 |
+
"grad_norm": 0.5060763955116272,
|
| 471 |
+
"learning_rate": 8.135881792367686e-05,
|
| 472 |
+
"loss": 1.5424,
|
| 473 |
+
"step": 64
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.04236597686165879,
|
| 477 |
+
"grad_norm": 0.5149311423301697,
|
| 478 |
+
"learning_rate": 8.07106356344834e-05,
|
| 479 |
+
"loss": 1.5314,
|
| 480 |
+
"step": 65
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.04301776112106893,
|
| 484 |
+
"grad_norm": 0.6181354522705078,
|
| 485 |
+
"learning_rate": 8.005405736415126e-05,
|
| 486 |
+
"loss": 1.666,
|
| 487 |
+
"step": 66
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.04366954538047906,
|
| 491 |
+
"grad_norm": 0.6071786880493164,
|
| 492 |
+
"learning_rate": 7.938926261462366e-05,
|
| 493 |
+
"loss": 1.6037,
|
| 494 |
+
"step": 67
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.0443213296398892,
|
| 498 |
+
"grad_norm": 0.5892857313156128,
|
| 499 |
+
"learning_rate": 7.871643313414718e-05,
|
| 500 |
+
"loss": 1.5234,
|
| 501 |
+
"step": 68
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.04497311389929933,
|
| 505 |
+
"grad_norm": 0.6549196839332581,
|
| 506 |
+
"learning_rate": 7.803575286758364e-05,
|
| 507 |
+
"loss": 1.7427,
|
| 508 |
+
"step": 69
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.045624898158709466,
|
| 512 |
+
"grad_norm": 0.6623611450195312,
|
| 513 |
+
"learning_rate": 7.734740790612136e-05,
|
| 514 |
+
"loss": 1.6478,
|
| 515 |
+
"step": 70
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.0462766824181196,
|
| 519 |
+
"grad_norm": 0.6412277817726135,
|
| 520 |
+
"learning_rate": 7.66515864363997e-05,
|
| 521 |
+
"loss": 1.654,
|
| 522 |
+
"step": 71
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.046928466677529736,
|
| 526 |
+
"grad_norm": 0.7149087190628052,
|
| 527 |
+
"learning_rate": 7.594847868906076e-05,
|
| 528 |
+
"loss": 1.5823,
|
| 529 |
+
"step": 72
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.04758025093693987,
|
| 533 |
+
"grad_norm": 0.6478785276412964,
|
| 534 |
+
"learning_rate": 7.52382768867422e-05,
|
| 535 |
+
"loss": 1.5092,
|
| 536 |
+
"step": 73
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.048232035196350005,
|
| 540 |
+
"grad_norm": 0.8339841365814209,
|
| 541 |
+
"learning_rate": 7.452117519152542e-05,
|
| 542 |
+
"loss": 1.7081,
|
| 543 |
+
"step": 74
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.04888381945576014,
|
| 547 |
+
"grad_norm": 0.6634137630462646,
|
| 548 |
+
"learning_rate": 7.379736965185368e-05,
|
| 549 |
+
"loss": 1.5685,
|
| 550 |
+
"step": 75
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.04953560371517028,
|
| 554 |
+
"grad_norm": 0.7319443821907043,
|
| 555 |
+
"learning_rate": 7.30670581489344e-05,
|
| 556 |
+
"loss": 1.3952,
|
| 557 |
+
"step": 76
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.050187387974580416,
|
| 561 |
+
"grad_norm": 0.7438411116600037,
|
| 562 |
+
"learning_rate": 7.233044034264034e-05,
|
| 563 |
+
"loss": 1.3501,
|
| 564 |
+
"step": 77
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.05083917223399055,
|
| 568 |
+
"grad_norm": 0.7561342120170593,
|
| 569 |
+
"learning_rate": 7.158771761692464e-05,
|
| 570 |
+
"loss": 1.5413,
|
| 571 |
+
"step": 78
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.051490956493400686,
|
| 575 |
+
"grad_norm": 0.6312757134437561,
|
| 576 |
+
"learning_rate": 7.083909302476453e-05,
|
| 577 |
+
"loss": 1.3898,
|
| 578 |
+
"step": 79
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.05214274075281082,
|
| 582 |
+
"grad_norm": 0.7573202252388,
|
| 583 |
+
"learning_rate": 7.008477123264848e-05,
|
| 584 |
+
"loss": 1.3744,
|
| 585 |
+
"step": 80
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.052794525012220955,
|
| 589 |
+
"grad_norm": 0.9921998381614685,
|
| 590 |
+
"learning_rate": 6.932495846462261e-05,
|
| 591 |
+
"loss": 1.5812,
|
| 592 |
+
"step": 81
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.05344630927163109,
|
| 596 |
+
"grad_norm": 0.9239053726196289,
|
| 597 |
+
"learning_rate": 6.855986244591104e-05,
|
| 598 |
+
"loss": 1.3269,
|
| 599 |
+
"step": 82
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.054098093531041225,
|
| 603 |
+
"grad_norm": 0.8641653060913086,
|
| 604 |
+
"learning_rate": 6.778969234612584e-05,
|
| 605 |
+
"loss": 1.4101,
|
| 606 |
+
"step": 83
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.05474987779045136,
|
| 610 |
+
"grad_norm": 0.9148449301719666,
|
| 611 |
+
"learning_rate": 6.701465872208216e-05,
|
| 612 |
+
"loss": 1.3028,
|
| 613 |
+
"step": 84
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.055401662049861494,
|
| 617 |
+
"grad_norm": 1.0409218072891235,
|
| 618 |
+
"learning_rate": 6.623497346023418e-05,
|
| 619 |
+
"loss": 1.3792,
|
| 620 |
+
"step": 85
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.05605344630927163,
|
| 624 |
+
"grad_norm": 0.978447437286377,
|
| 625 |
+
"learning_rate": 6.545084971874738e-05,
|
| 626 |
+
"loss": 1.3389,
|
| 627 |
+
"step": 86
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.056705230568681764,
|
| 631 |
+
"grad_norm": 1.385927677154541,
|
| 632 |
+
"learning_rate": 6.466250186922325e-05,
|
| 633 |
+
"loss": 1.4932,
|
| 634 |
+
"step": 87
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.0573570148280919,
|
| 638 |
+
"grad_norm": 0.9205519556999207,
|
| 639 |
+
"learning_rate": 6.387014543809223e-05,
|
| 640 |
+
"loss": 1.0915,
|
| 641 |
+
"step": 88
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.05800879908750204,
|
| 645 |
+
"grad_norm": 1.234601378440857,
|
| 646 |
+
"learning_rate": 6.307399704769099e-05,
|
| 647 |
+
"loss": 1.2098,
|
| 648 |
+
"step": 89
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 0.058660583346912175,
|
| 652 |
+
"grad_norm": 1.015903115272522,
|
| 653 |
+
"learning_rate": 6.227427435703997e-05,
|
| 654 |
+
"loss": 1.2165,
|
| 655 |
+
"step": 90
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.05931236760632231,
|
| 659 |
+
"grad_norm": 1.420213222503662,
|
| 660 |
+
"learning_rate": 6.147119600233758e-05,
|
| 661 |
+
"loss": 1.582,
|
| 662 |
+
"step": 91
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.059964151865732444,
|
| 666 |
+
"grad_norm": 1.2678552865982056,
|
| 667 |
+
"learning_rate": 6.066498153718735e-05,
|
| 668 |
+
"loss": 1.3321,
|
| 669 |
+
"step": 92
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.06061593612514258,
|
| 673 |
+
"grad_norm": 1.171578288078308,
|
| 674 |
+
"learning_rate": 5.985585137257401e-05,
|
| 675 |
+
"loss": 1.2771,
|
| 676 |
+
"step": 93
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.061267720384552714,
|
| 680 |
+
"grad_norm": 1.3661903142929077,
|
| 681 |
+
"learning_rate": 5.90440267166055e-05,
|
| 682 |
+
"loss": 1.4461,
|
| 683 |
+
"step": 94
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.06191950464396285,
|
| 687 |
+
"grad_norm": 1.2631621360778809,
|
| 688 |
+
"learning_rate": 5.8229729514036705e-05,
|
| 689 |
+
"loss": 1.11,
|
| 690 |
+
"step": 95
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 0.06257128890337299,
|
| 694 |
+
"grad_norm": 1.7056459188461304,
|
| 695 |
+
"learning_rate": 5.74131823855921e-05,
|
| 696 |
+
"loss": 1.3257,
|
| 697 |
+
"step": 96
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.06322307316278312,
|
| 701 |
+
"grad_norm": 1.7923287153244019,
|
| 702 |
+
"learning_rate": 5.6594608567103456e-05,
|
| 703 |
+
"loss": 1.3669,
|
| 704 |
+
"step": 97
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.06387485742219326,
|
| 708 |
+
"grad_norm": 1.509992003440857,
|
| 709 |
+
"learning_rate": 5.577423184847932e-05,
|
| 710 |
+
"loss": 1.3823,
|
| 711 |
+
"step": 98
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.0645266416816034,
|
| 715 |
+
"grad_norm": 2.4850306510925293,
|
| 716 |
+
"learning_rate": 5.495227651252315e-05,
|
| 717 |
+
"loss": 1.7633,
|
| 718 |
+
"step": 99
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.06517842594101353,
|
| 722 |
+
"grad_norm": 3.989614725112915,
|
| 723 |
+
"learning_rate": 5.4128967273616625e-05,
|
| 724 |
+
"loss": 1.3741,
|
| 725 |
+
"step": 100
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.06517842594101353,
|
| 729 |
+
"eval_loss": 1.439968466758728,
|
| 730 |
+
"eval_runtime": 203.9683,
|
| 731 |
+
"eval_samples_per_second": 12.669,
|
| 732 |
+
"eval_steps_per_second": 3.167,
|
| 733 |
+
"step": 100
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.06583021020042366,
|
| 737 |
+
"grad_norm": 0.31170693039894104,
|
| 738 |
+
"learning_rate": 5.330452921628497e-05,
|
| 739 |
+
"loss": 1.0834,
|
| 740 |
+
"step": 101
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.0664819944598338,
|
| 744 |
+
"grad_norm": 0.32115522027015686,
|
| 745 |
+
"learning_rate": 5.247918773366112e-05,
|
| 746 |
+
"loss": 1.1309,
|
| 747 |
+
"step": 102
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.06713377871924393,
|
| 751 |
+
"grad_norm": 0.3525278568267822,
|
| 752 |
+
"learning_rate": 5.165316846586541e-05,
|
| 753 |
+
"loss": 1.1598,
|
| 754 |
+
"step": 103
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.06778556297865407,
|
| 758 |
+
"grad_norm": 0.358556866645813,
|
| 759 |
+
"learning_rate": 5.0826697238317935e-05,
|
| 760 |
+
"loss": 1.3622,
|
| 761 |
+
"step": 104
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 0.0684373472380642,
|
| 765 |
+
"grad_norm": 0.32107800245285034,
|
| 766 |
+
"learning_rate": 5e-05,
|
| 767 |
+
"loss": 1.1696,
|
| 768 |
+
"step": 105
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 0.06908913149747434,
|
| 772 |
+
"grad_norm": 0.4025322496891022,
|
| 773 |
+
"learning_rate": 4.917330276168208e-05,
|
| 774 |
+
"loss": 1.2357,
|
| 775 |
+
"step": 106
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 0.06974091575688447,
|
| 779 |
+
"grad_norm": 0.36745789647102356,
|
| 780 |
+
"learning_rate": 4.834683153413459e-05,
|
| 781 |
+
"loss": 1.3919,
|
| 782 |
+
"step": 107
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.07039270001629461,
|
| 786 |
+
"grad_norm": 0.3489651083946228,
|
| 787 |
+
"learning_rate": 4.7520812266338885e-05,
|
| 788 |
+
"loss": 1.106,
|
| 789 |
+
"step": 108
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.07104448427570474,
|
| 793 |
+
"grad_norm": 0.37726274132728577,
|
| 794 |
+
"learning_rate": 4.669547078371504e-05,
|
| 795 |
+
"loss": 1.2702,
|
| 796 |
+
"step": 109
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.07169626853511488,
|
| 800 |
+
"grad_norm": 0.3918270766735077,
|
| 801 |
+
"learning_rate": 4.5871032726383386e-05,
|
| 802 |
+
"loss": 1.1876,
|
| 803 |
+
"step": 110
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.07234805279452501,
|
| 807 |
+
"grad_norm": 0.415508896112442,
|
| 808 |
+
"learning_rate": 4.504772348747687e-05,
|
| 809 |
+
"loss": 1.3956,
|
| 810 |
+
"step": 111
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.07299983705393515,
|
| 814 |
+
"grad_norm": 0.380808562040329,
|
| 815 |
+
"learning_rate": 4.4225768151520694e-05,
|
| 816 |
+
"loss": 1.3316,
|
| 817 |
+
"step": 112
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.07365162131334528,
|
| 821 |
+
"grad_norm": 0.46469783782958984,
|
| 822 |
+
"learning_rate": 4.3405391432896555e-05,
|
| 823 |
+
"loss": 1.4574,
|
| 824 |
+
"step": 113
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.07430340557275542,
|
| 828 |
+
"grad_norm": 0.4693603217601776,
|
| 829 |
+
"learning_rate": 4.2586817614407895e-05,
|
| 830 |
+
"loss": 1.3827,
|
| 831 |
+
"step": 114
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.07495518983216555,
|
| 835 |
+
"grad_norm": 0.5363829731941223,
|
| 836 |
+
"learning_rate": 4.17702704859633e-05,
|
| 837 |
+
"loss": 1.5779,
|
| 838 |
+
"step": 115
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.07560697409157568,
|
| 842 |
+
"grad_norm": 0.5474920868873596,
|
| 843 |
+
"learning_rate": 4.095597328339452e-05,
|
| 844 |
+
"loss": 1.4873,
|
| 845 |
+
"step": 116
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.07625875835098582,
|
| 849 |
+
"grad_norm": 0.5812809467315674,
|
| 850 |
+
"learning_rate": 4.0144148627425993e-05,
|
| 851 |
+
"loss": 1.5276,
|
| 852 |
+
"step": 117
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 0.07691054261039595,
|
| 856 |
+
"grad_norm": 0.6456193923950195,
|
| 857 |
+
"learning_rate": 3.933501846281267e-05,
|
| 858 |
+
"loss": 1.4315,
|
| 859 |
+
"step": 118
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.07756232686980609,
|
| 863 |
+
"grad_norm": 0.5870631337165833,
|
| 864 |
+
"learning_rate": 3.852880399766243e-05,
|
| 865 |
+
"loss": 1.6514,
|
| 866 |
+
"step": 119
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.07821411112921622,
|
| 870 |
+
"grad_norm": 0.6266175508499146,
|
| 871 |
+
"learning_rate": 3.772572564296005e-05,
|
| 872 |
+
"loss": 1.6129,
|
| 873 |
+
"step": 120
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.07886589538862636,
|
| 877 |
+
"grad_norm": 0.6647204160690308,
|
| 878 |
+
"learning_rate": 3.6926002952309016e-05,
|
| 879 |
+
"loss": 1.6369,
|
| 880 |
+
"step": 121
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.0795176796480365,
|
| 884 |
+
"grad_norm": 0.6961146593093872,
|
| 885 |
+
"learning_rate": 3.612985456190778e-05,
|
| 886 |
+
"loss": 1.6736,
|
| 887 |
+
"step": 122
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.08016946390744664,
|
| 891 |
+
"grad_norm": 0.7353977560997009,
|
| 892 |
+
"learning_rate": 3.533749813077677e-05,
|
| 893 |
+
"loss": 1.6961,
|
| 894 |
+
"step": 123
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.08082124816685678,
|
| 898 |
+
"grad_norm": 0.6045286059379578,
|
| 899 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 900 |
+
"loss": 1.4501,
|
| 901 |
+
"step": 124
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.08147303242626691,
|
| 905 |
+
"grad_norm": 0.6582840085029602,
|
| 906 |
+
"learning_rate": 3.3765026539765834e-05,
|
| 907 |
+
"loss": 1.4565,
|
| 908 |
+
"step": 125
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.08212481668567705,
|
| 912 |
+
"grad_norm": 0.6543602347373962,
|
| 913 |
+
"learning_rate": 3.298534127791785e-05,
|
| 914 |
+
"loss": 1.399,
|
| 915 |
+
"step": 126
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.08277660094508718,
|
| 919 |
+
"grad_norm": 0.7159081697463989,
|
| 920 |
+
"learning_rate": 3.221030765387417e-05,
|
| 921 |
+
"loss": 1.4648,
|
| 922 |
+
"step": 127
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.08342838520449732,
|
| 926 |
+
"grad_norm": 0.7286494374275208,
|
| 927 |
+
"learning_rate": 3.144013755408895e-05,
|
| 928 |
+
"loss": 1.4932,
|
| 929 |
+
"step": 128
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.08408016946390745,
|
| 933 |
+
"grad_norm": 0.6608577370643616,
|
| 934 |
+
"learning_rate": 3.0675041535377405e-05,
|
| 935 |
+
"loss": 1.2703,
|
| 936 |
+
"step": 129
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.08473195372331759,
|
| 940 |
+
"grad_norm": 0.8380109667778015,
|
| 941 |
+
"learning_rate": 2.991522876735154e-05,
|
| 942 |
+
"loss": 1.2489,
|
| 943 |
+
"step": 130
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.08538373798272772,
|
| 947 |
+
"grad_norm": 0.7555432319641113,
|
| 948 |
+
"learning_rate": 2.916090697523549e-05,
|
| 949 |
+
"loss": 1.2819,
|
| 950 |
+
"step": 131
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.08603552224213785,
|
| 954 |
+
"grad_norm": 0.7918082475662231,
|
| 955 |
+
"learning_rate": 2.8412282383075363e-05,
|
| 956 |
+
"loss": 1.2549,
|
| 957 |
+
"step": 132
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.08668730650154799,
|
| 961 |
+
"grad_norm": 0.9338359236717224,
|
| 962 |
+
"learning_rate": 2.766955965735968e-05,
|
| 963 |
+
"loss": 1.4434,
|
| 964 |
+
"step": 133
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.08733909076095812,
|
| 968 |
+
"grad_norm": 0.8965134024620056,
|
| 969 |
+
"learning_rate": 2.693294185106562e-05,
|
| 970 |
+
"loss": 1.5997,
|
| 971 |
+
"step": 134
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.08799087502036826,
|
| 975 |
+
"grad_norm": 0.864808976650238,
|
| 976 |
+
"learning_rate": 2.6202630348146324e-05,
|
| 977 |
+
"loss": 1.1591,
|
| 978 |
+
"step": 135
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.0886426592797784,
|
| 982 |
+
"grad_norm": 0.8886546492576599,
|
| 983 |
+
"learning_rate": 2.547882480847461e-05,
|
| 984 |
+
"loss": 1.1788,
|
| 985 |
+
"step": 136
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.08929444353918853,
|
| 989 |
+
"grad_norm": 0.9865848422050476,
|
| 990 |
+
"learning_rate": 2.476172311325783e-05,
|
| 991 |
+
"loss": 1.1863,
|
| 992 |
+
"step": 137
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.08994622779859866,
|
| 996 |
+
"grad_norm": 1.2445539236068726,
|
| 997 |
+
"learning_rate": 2.405152131093926e-05,
|
| 998 |
+
"loss": 1.4217,
|
| 999 |
+
"step": 138
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.0905980120580088,
|
| 1003 |
+
"grad_norm": 1.043266773223877,
|
| 1004 |
+
"learning_rate": 2.3348413563600325e-05,
|
| 1005 |
+
"loss": 1.1928,
|
| 1006 |
+
"step": 139
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.09124979631741893,
|
| 1010 |
+
"grad_norm": 1.2551209926605225,
|
| 1011 |
+
"learning_rate": 2.2652592093878666e-05,
|
| 1012 |
+
"loss": 1.2908,
|
| 1013 |
+
"step": 140
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.09190158057682907,
|
| 1017 |
+
"grad_norm": 1.1211011409759521,
|
| 1018 |
+
"learning_rate": 2.196424713241637e-05,
|
| 1019 |
+
"loss": 1.0974,
|
| 1020 |
+
"step": 141
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.0925533648362392,
|
| 1024 |
+
"grad_norm": 1.3813427686691284,
|
| 1025 |
+
"learning_rate": 2.128356686585282e-05,
|
| 1026 |
+
"loss": 1.516,
|
| 1027 |
+
"step": 142
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.09320514909564934,
|
| 1031 |
+
"grad_norm": 1.0231101512908936,
|
| 1032 |
+
"learning_rate": 2.061073738537635e-05,
|
| 1033 |
+
"loss": 1.1618,
|
| 1034 |
+
"step": 143
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.09385693335505947,
|
| 1038 |
+
"grad_norm": 1.2551326751708984,
|
| 1039 |
+
"learning_rate": 1.9945942635848748e-05,
|
| 1040 |
+
"loss": 1.3938,
|
| 1041 |
+
"step": 144
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.0945087176144696,
|
| 1045 |
+
"grad_norm": 1.2844427824020386,
|
| 1046 |
+
"learning_rate": 1.928936436551661e-05,
|
| 1047 |
+
"loss": 1.3299,
|
| 1048 |
+
"step": 145
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.09516050187387974,
|
| 1052 |
+
"grad_norm": 1.5775208473205566,
|
| 1053 |
+
"learning_rate": 1.8641182076323148e-05,
|
| 1054 |
+
"loss": 1.4055,
|
| 1055 |
+
"step": 146
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.09581228613328988,
|
| 1059 |
+
"grad_norm": 1.5789850950241089,
|
| 1060 |
+
"learning_rate": 1.800157297483417e-05,
|
| 1061 |
+
"loss": 1.3381,
|
| 1062 |
+
"step": 147
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.09646407039270001,
|
| 1066 |
+
"grad_norm": 1.4319522380828857,
|
| 1067 |
+
"learning_rate": 1.7370711923791567e-05,
|
| 1068 |
+
"loss": 1.1653,
|
| 1069 |
+
"step": 148
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.09711585465211015,
|
| 1073 |
+
"grad_norm": 1.9949793815612793,
|
| 1074 |
+
"learning_rate": 1.6748771394307585e-05,
|
| 1075 |
+
"loss": 1.2896,
|
| 1076 |
+
"step": 149
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.09776763891152028,
|
| 1080 |
+
"grad_norm": 2.9262423515319824,
|
| 1081 |
+
"learning_rate": 1.6135921418712956e-05,
|
| 1082 |
+
"loss": 2.3794,
|
| 1083 |
+
"step": 150
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.09776763891152028,
|
| 1087 |
+
"eval_loss": 1.3852198123931885,
|
| 1088 |
+
"eval_runtime": 203.8229,
|
| 1089 |
+
"eval_samples_per_second": 12.678,
|
| 1090 |
+
"eval_steps_per_second": 3.169,
|
| 1091 |
+
"step": 150
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
|
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
+
"total_flos": 2.15995700477952e+17,
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|