Training in progress, step 2346, 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 56662456
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c143e49ec80b9efa7c889526d4266a7b197596fb39eb75eb8f0c60a959dfb5b3
|
| 3 |
size 56662456
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 29091284
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:229ef2852a3e0ca7253143b05e4b688888a7bf5fee9127a04663067e45d8cf7d
|
| 3 |
size 29091284
|
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 sha256:07d6aa011454a4748f5199ba02a257f0c3397f7d63ff5d7de731bdce7a2a6006
|
| 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 sha256:82f6fb6c04e83ad8c7e7774f50cfceed3bf90e6e42ded09b4deef26723be76bc
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.007140692323446274,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-2300",
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -16299,6 +16299,328 @@
|
|
| 16299 |
"eval_samples_per_second": 10.677,
|
| 16300 |
"eval_steps_per_second": 2.669,
|
| 16301 |
"step": 2300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16302 |
}
|
| 16303 |
],
|
| 16304 |
"logging_steps": 1,
|
|
@@ -16322,12 +16644,12 @@
|
|
| 16322 |
"should_evaluate": false,
|
| 16323 |
"should_log": false,
|
| 16324 |
"should_save": true,
|
| 16325 |
-
"should_training_stop":
|
| 16326 |
},
|
| 16327 |
"attributes": {}
|
| 16328 |
}
|
| 16329 |
},
|
| 16330 |
-
"total_flos": 1.
|
| 16331 |
"train_batch_size": 4,
|
| 16332 |
"trial_name": null,
|
| 16333 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.007140692323446274,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-2300",
|
| 4 |
+
"epoch": 0.6973063347575701,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 2346,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 16299 |
"eval_samples_per_second": 10.677,
|
| 16300 |
"eval_steps_per_second": 2.669,
|
| 16301 |
"step": 2300
|
| 16302 |
+
},
|
| 16303 |
+
{
|
| 16304 |
+
"epoch": 0.6839308935537805,
|
| 16305 |
+
"grad_norm": 0.11718721687793732,
|
| 16306 |
+
"learning_rate": 1.8306973193326084e-07,
|
| 16307 |
+
"loss": 0.0639,
|
| 16308 |
+
"step": 2301
|
| 16309 |
+
},
|
| 16310 |
+
{
|
| 16311 |
+
"epoch": 0.6842281255805313,
|
| 16312 |
+
"grad_norm": 0.11004548519849777,
|
| 16313 |
+
"learning_rate": 1.7502605215715672e-07,
|
| 16314 |
+
"loss": 0.0404,
|
| 16315 |
+
"step": 2302
|
| 16316 |
+
},
|
| 16317 |
+
{
|
| 16318 |
+
"epoch": 0.6845253576072822,
|
| 16319 |
+
"grad_norm": 0.10758353769779205,
|
| 16320 |
+
"learning_rate": 1.671629205999836e-07,
|
| 16321 |
+
"loss": 0.0538,
|
| 16322 |
+
"step": 2303
|
| 16323 |
+
},
|
| 16324 |
+
{
|
| 16325 |
+
"epoch": 0.6848225896340331,
|
| 16326 |
+
"grad_norm": 0.1499020904302597,
|
| 16327 |
+
"learning_rate": 1.5948035148338757e-07,
|
| 16328 |
+
"loss": 0.082,
|
| 16329 |
+
"step": 2304
|
| 16330 |
+
},
|
| 16331 |
+
{
|
| 16332 |
+
"epoch": 0.685119821660784,
|
| 16333 |
+
"grad_norm": 0.10471717268228531,
|
| 16334 |
+
"learning_rate": 1.5197835870242038e-07,
|
| 16335 |
+
"loss": 0.049,
|
| 16336 |
+
"step": 2305
|
| 16337 |
+
},
|
| 16338 |
+
{
|
| 16339 |
+
"epoch": 0.6854170536875348,
|
| 16340 |
+
"grad_norm": 0.09845124185085297,
|
| 16341 |
+
"learning_rate": 1.446569558255395e-07,
|
| 16342 |
+
"loss": 0.0395,
|
| 16343 |
+
"step": 2306
|
| 16344 |
+
},
|
| 16345 |
+
{
|
| 16346 |
+
"epoch": 0.6857142857142857,
|
| 16347 |
+
"grad_norm": 0.10133890062570572,
|
| 16348 |
+
"learning_rate": 1.375161560946081e-07,
|
| 16349 |
+
"loss": 0.044,
|
| 16350 |
+
"step": 2307
|
| 16351 |
+
},
|
| 16352 |
+
{
|
| 16353 |
+
"epoch": 0.6860115177410366,
|
| 16354 |
+
"grad_norm": 0.12090011686086655,
|
| 16355 |
+
"learning_rate": 1.305559724248062e-07,
|
| 16356 |
+
"loss": 0.0723,
|
| 16357 |
+
"step": 2308
|
| 16358 |
+
},
|
| 16359 |
+
{
|
| 16360 |
+
"epoch": 0.6863087497677874,
|
| 16361 |
+
"grad_norm": 0.10536789149045944,
|
| 16362 |
+
"learning_rate": 1.2377641740464187e-07,
|
| 16363 |
+
"loss": 0.0536,
|
| 16364 |
+
"step": 2309
|
| 16365 |
+
},
|
| 16366 |
+
{
|
| 16367 |
+
"epoch": 0.6866059817945384,
|
| 16368 |
+
"grad_norm": 0.13396191596984863,
|
| 16369 |
+
"learning_rate": 1.1717750329595101e-07,
|
| 16370 |
+
"loss": 0.0716,
|
| 16371 |
+
"step": 2310
|
| 16372 |
+
},
|
| 16373 |
+
{
|
| 16374 |
+
"epoch": 0.6869032138212893,
|
| 16375 |
+
"grad_norm": 0.11562539637088776,
|
| 16376 |
+
"learning_rate": 1.1075924203385324e-07,
|
| 16377 |
+
"loss": 0.061,
|
| 16378 |
+
"step": 2311
|
| 16379 |
+
},
|
| 16380 |
+
{
|
| 16381 |
+
"epoch": 0.6872004458480401,
|
| 16382 |
+
"grad_norm": 0.08731340616941452,
|
| 16383 |
+
"learning_rate": 1.0452164522671837e-07,
|
| 16384 |
+
"loss": 0.0424,
|
| 16385 |
+
"step": 2312
|
| 16386 |
+
},
|
| 16387 |
+
{
|
| 16388 |
+
"epoch": 0.687497677874791,
|
| 16389 |
+
"grad_norm": 0.10618704557418823,
|
| 16390 |
+
"learning_rate": 9.846472415615537e-08,
|
| 16391 |
+
"loss": 0.0484,
|
| 16392 |
+
"step": 2313
|
| 16393 |
+
},
|
| 16394 |
+
{
|
| 16395 |
+
"epoch": 0.6877949099015419,
|
| 16396 |
+
"grad_norm": 0.10936351865530014,
|
| 16397 |
+
"learning_rate": 9.258848977700129e-08,
|
| 16398 |
+
"loss": 0.0648,
|
| 16399 |
+
"step": 2314
|
| 16400 |
+
},
|
| 16401 |
+
{
|
| 16402 |
+
"epoch": 0.6880921419282928,
|
| 16403 |
+
"grad_norm": 0.09159952402114868,
|
| 16404 |
+
"learning_rate": 8.689295271729902e-08,
|
| 16405 |
+
"loss": 0.0359,
|
| 16406 |
+
"step": 2315
|
| 16407 |
+
},
|
| 16408 |
+
{
|
| 16409 |
+
"epoch": 0.6883893739550436,
|
| 16410 |
+
"grad_norm": 0.09538505971431732,
|
| 16411 |
+
"learning_rate": 8.13781232782751e-08,
|
| 16412 |
+
"loss": 0.0423,
|
| 16413 |
+
"step": 2316
|
| 16414 |
+
},
|
| 16415 |
+
{
|
| 16416 |
+
"epoch": 0.6886866059817945,
|
| 16417 |
+
"grad_norm": 0.1170530915260315,
|
| 16418 |
+
"learning_rate": 7.604401143430639e-08,
|
| 16419 |
+
"loss": 0.0589,
|
| 16420 |
+
"step": 2317
|
| 16421 |
+
},
|
| 16422 |
+
{
|
| 16423 |
+
"epoch": 0.6889838380085455,
|
| 16424 |
+
"grad_norm": 0.10562342405319214,
|
| 16425 |
+
"learning_rate": 7.089062683292014e-08,
|
| 16426 |
+
"loss": 0.0508,
|
| 16427 |
+
"step": 2318
|
| 16428 |
+
},
|
| 16429 |
+
{
|
| 16430 |
+
"epoch": 0.6892810700352963,
|
| 16431 |
+
"grad_norm": 0.11546720564365387,
|
| 16432 |
+
"learning_rate": 6.591797879478279e-08,
|
| 16433 |
+
"loss": 0.0588,
|
| 16434 |
+
"step": 2319
|
| 16435 |
+
},
|
| 16436 |
+
{
|
| 16437 |
+
"epoch": 0.6895783020620472,
|
| 16438 |
+
"grad_norm": 0.14187178015708923,
|
| 16439 |
+
"learning_rate": 6.112607631364453e-08,
|
| 16440 |
+
"loss": 0.0705,
|
| 16441 |
+
"step": 2320
|
| 16442 |
+
},
|
| 16443 |
+
{
|
| 16444 |
+
"epoch": 0.6898755340887981,
|
| 16445 |
+
"grad_norm": 0.10787046700716019,
|
| 16446 |
+
"learning_rate": 5.65149280563948e-08,
|
| 16447 |
+
"loss": 0.0587,
|
| 16448 |
+
"step": 2321
|
| 16449 |
+
},
|
| 16450 |
+
{
|
| 16451 |
+
"epoch": 0.6901727661155489,
|
| 16452 |
+
"grad_norm": 0.08616691827774048,
|
| 16453 |
+
"learning_rate": 5.208454236296234e-08,
|
| 16454 |
+
"loss": 0.0341,
|
| 16455 |
+
"step": 2322
|
| 16456 |
+
},
|
| 16457 |
+
{
|
| 16458 |
+
"epoch": 0.6904699981422998,
|
| 16459 |
+
"grad_norm": 0.13538287580013275,
|
| 16460 |
+
"learning_rate": 4.783492724635963e-08,
|
| 16461 |
+
"loss": 0.0772,
|
| 16462 |
+
"step": 2323
|
| 16463 |
+
},
|
| 16464 |
+
{
|
| 16465 |
+
"epoch": 0.6907672301690507,
|
| 16466 |
+
"grad_norm": 0.11333715915679932,
|
| 16467 |
+
"learning_rate": 4.376609039262736e-08,
|
| 16468 |
+
"loss": 0.0569,
|
| 16469 |
+
"step": 2324
|
| 16470 |
+
},
|
| 16471 |
+
{
|
| 16472 |
+
"epoch": 0.6910644621958016,
|
| 16473 |
+
"grad_norm": 0.12235751003026962,
|
| 16474 |
+
"learning_rate": 3.9878039160878844e-08,
|
| 16475 |
+
"loss": 0.0556,
|
| 16476 |
+
"step": 2325
|
| 16477 |
+
},
|
| 16478 |
+
{
|
| 16479 |
+
"epoch": 0.6913616942225524,
|
| 16480 |
+
"grad_norm": 0.10557149350643158,
|
| 16481 |
+
"learning_rate": 3.617078058322232e-08,
|
| 16482 |
+
"loss": 0.046,
|
| 16483 |
+
"step": 2326
|
| 16484 |
+
},
|
| 16485 |
+
{
|
| 16486 |
+
"epoch": 0.6916589262493034,
|
| 16487 |
+
"grad_norm": 0.12478914111852646,
|
| 16488 |
+
"learning_rate": 3.264432136478313e-08,
|
| 16489 |
+
"loss": 0.0733,
|
| 16490 |
+
"step": 2327
|
| 16491 |
+
},
|
| 16492 |
+
{
|
| 16493 |
+
"epoch": 0.6919561582760543,
|
| 16494 |
+
"grad_norm": 0.08897604048252106,
|
| 16495 |
+
"learning_rate": 2.9298667883692622e-08,
|
| 16496 |
+
"loss": 0.0388,
|
| 16497 |
+
"step": 2328
|
| 16498 |
+
},
|
| 16499 |
+
{
|
| 16500 |
+
"epoch": 0.6922533903028051,
|
| 16501 |
+
"grad_norm": 0.09747358411550522,
|
| 16502 |
+
"learning_rate": 2.6133826191032663e-08,
|
| 16503 |
+
"loss": 0.0401,
|
| 16504 |
+
"step": 2329
|
| 16505 |
+
},
|
| 16506 |
+
{
|
| 16507 |
+
"epoch": 0.692550622329556,
|
| 16508 |
+
"grad_norm": 0.09661777317523956,
|
| 16509 |
+
"learning_rate": 2.3149802010913323e-08,
|
| 16510 |
+
"loss": 0.0403,
|
| 16511 |
+
"step": 2330
|
| 16512 |
+
},
|
| 16513 |
+
{
|
| 16514 |
+
"epoch": 0.6928478543563069,
|
| 16515 |
+
"grad_norm": 0.12822963297367096,
|
| 16516 |
+
"learning_rate": 2.034660074037298e-08,
|
| 16517 |
+
"loss": 0.0739,
|
| 16518 |
+
"step": 2331
|
| 16519 |
+
},
|
| 16520 |
+
{
|
| 16521 |
+
"epoch": 0.6931450863830578,
|
| 16522 |
+
"grad_norm": 0.12262982130050659,
|
| 16523 |
+
"learning_rate": 1.7724227449422705e-08,
|
| 16524 |
+
"loss": 0.0491,
|
| 16525 |
+
"step": 2332
|
| 16526 |
+
},
|
| 16527 |
+
{
|
| 16528 |
+
"epoch": 0.6934423184098086,
|
| 16529 |
+
"grad_norm": 0.11021570861339569,
|
| 16530 |
+
"learning_rate": 1.5282686881001875e-08,
|
| 16531 |
+
"loss": 0.0444,
|
| 16532 |
+
"step": 2333
|
| 16533 |
+
},
|
| 16534 |
+
{
|
| 16535 |
+
"epoch": 0.6937395504365595,
|
| 16536 |
+
"grad_norm": 0.10723091661930084,
|
| 16537 |
+
"learning_rate": 1.3021983451000364e-08,
|
| 16538 |
+
"loss": 0.0525,
|
| 16539 |
+
"step": 2334
|
| 16540 |
+
},
|
| 16541 |
+
{
|
| 16542 |
+
"epoch": 0.6940367824633105,
|
| 16543 |
+
"grad_norm": 0.12233065813779831,
|
| 16544 |
+
"learning_rate": 1.094212124824745e-08,
|
| 16545 |
+
"loss": 0.0541,
|
| 16546 |
+
"step": 2335
|
| 16547 |
+
},
|
| 16548 |
+
{
|
| 16549 |
+
"epoch": 0.6943340144900613,
|
| 16550 |
+
"grad_norm": 0.12585288286209106,
|
| 16551 |
+
"learning_rate": 9.043104034456295e-09,
|
| 16552 |
+
"loss": 0.0507,
|
| 16553 |
+
"step": 2336
|
| 16554 |
+
},
|
| 16555 |
+
{
|
| 16556 |
+
"epoch": 0.6946312465168122,
|
| 16557 |
+
"grad_norm": 0.08642668277025223,
|
| 16558 |
+
"learning_rate": 7.324935244301667e-09,
|
| 16559 |
+
"loss": 0.0327,
|
| 16560 |
+
"step": 2337
|
| 16561 |
+
},
|
| 16562 |
+
{
|
| 16563 |
+
"epoch": 0.6949284785435631,
|
| 16564 |
+
"grad_norm": 0.10163281857967377,
|
| 16565 |
+
"learning_rate": 5.78761798534222e-09,
|
| 16566 |
+
"loss": 0.0378,
|
| 16567 |
+
"step": 2338
|
| 16568 |
+
},
|
| 16569 |
+
{
|
| 16570 |
+
"epoch": 0.6952257105703139,
|
| 16571 |
+
"grad_norm": 0.1360352784395218,
|
| 16572 |
+
"learning_rate": 4.431155038031598e-09,
|
| 16573 |
+
"loss": 0.0524,
|
| 16574 |
+
"step": 2339
|
| 16575 |
+
},
|
| 16576 |
+
{
|
| 16577 |
+
"epoch": 0.6955229425970648,
|
| 16578 |
+
"grad_norm": 0.12937673926353455,
|
| 16579 |
+
"learning_rate": 3.255548855740642e-09,
|
| 16580 |
+
"loss": 0.0615,
|
| 16581 |
+
"step": 2340
|
| 16582 |
+
},
|
| 16583 |
+
{
|
| 16584 |
+
"epoch": 0.6958201746238157,
|
| 16585 |
+
"grad_norm": 0.11729301512241364,
|
| 16586 |
+
"learning_rate": 2.260801564735182e-09,
|
| 16587 |
+
"loss": 0.0516,
|
| 16588 |
+
"step": 2341
|
| 16589 |
+
},
|
| 16590 |
+
{
|
| 16591 |
+
"epoch": 0.6961174066505667,
|
| 16592 |
+
"grad_norm": 0.1295451819896698,
|
| 16593 |
+
"learning_rate": 1.4469149641538337e-09,
|
| 16594 |
+
"loss": 0.0682,
|
| 16595 |
+
"step": 2342
|
| 16596 |
+
},
|
| 16597 |
+
{
|
| 16598 |
+
"epoch": 0.6964146386773175,
|
| 16599 |
+
"grad_norm": 0.13985326886177063,
|
| 16600 |
+
"learning_rate": 8.138905260302032e-10,
|
| 16601 |
+
"loss": 0.0777,
|
| 16602 |
+
"step": 2343
|
| 16603 |
+
},
|
| 16604 |
+
{
|
| 16605 |
+
"epoch": 0.6967118707040684,
|
| 16606 |
+
"grad_norm": 0.10082249343395233,
|
| 16607 |
+
"learning_rate": 3.617293952817846e-10,
|
| 16608 |
+
"loss": 0.0439,
|
| 16609 |
+
"step": 2344
|
| 16610 |
+
},
|
| 16611 |
+
{
|
| 16612 |
+
"epoch": 0.6970091027308193,
|
| 16613 |
+
"grad_norm": 0.11734326928853989,
|
| 16614 |
+
"learning_rate": 9.043238970996016e-11,
|
| 16615 |
+
"loss": 0.0561,
|
| 16616 |
+
"step": 2345
|
| 16617 |
+
},
|
| 16618 |
+
{
|
| 16619 |
+
"epoch": 0.6973063347575701,
|
| 16620 |
+
"grad_norm": 0.10924455523490906,
|
| 16621 |
+
"learning_rate": 0.0,
|
| 16622 |
+
"loss": 0.0507,
|
| 16623 |
+
"step": 2346
|
| 16624 |
}
|
| 16625 |
],
|
| 16626 |
"logging_steps": 1,
|
|
|
|
| 16644 |
"should_evaluate": false,
|
| 16645 |
"should_log": false,
|
| 16646 |
"should_save": true,
|
| 16647 |
+
"should_training_stop": true
|
| 16648 |
},
|
| 16649 |
"attributes": {}
|
| 16650 |
}
|
| 16651 |
},
|
| 16652 |
+
"total_flos": 1.0845724172605194e+18,
|
| 16653 |
"train_batch_size": 4,
|
| 16654 |
"trial_name": null,
|
| 16655 |
"trial_params": null
|