Upload folder using huggingface_hub
Browse files- data/model.safetensors +1 -1
- data/optimizer.pt +1 -1
- data/rng_state.pth +1 -1
- data/scheduler.pt +1 -1
- data/trainer_state.json +703 -3
- data/training_args.bin +1 -1
data/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 576008736
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:502786fd8a726e2156aa16f8cdb0a508eaeff2e6b7935f2a126e5c7fb3fe4875
|
| 3 |
size 576008736
|
data/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1152256984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31920983e209ab4ab448d37797fc73ec6f5a2c677beb1998c0c67cde4d40d85a
|
| 3 |
size 1152256984
|
data/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:c31ec632e18c6039ab923941c5b9a34579988b2eae98f203c7491579abcdc560
|
| 3 |
size 14244
|
data/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:cc76b89758f0a3c24f4601172ddc998440db62e30dfdfe43a93a78b9b974a013
|
| 3 |
size 1064
|
data/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -6307,6 +6307,706 @@
|
|
| 6307 |
"learning_rate": 2.9948870643768915e-05,
|
| 6308 |
"loss": 1.7891,
|
| 6309 |
"step": 13500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6310 |
}
|
| 6311 |
],
|
| 6312 |
"logging_steps": 15,
|
|
@@ -6326,7 +7026,7 @@
|
|
| 6326 |
"attributes": {}
|
| 6327 |
}
|
| 6328 |
},
|
| 6329 |
-
"total_flos": 2.
|
| 6330 |
"train_batch_size": 4,
|
| 6331 |
"trial_name": null,
|
| 6332 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.02921027111026293,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 15000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 6307 |
"learning_rate": 2.9948870643768915e-05,
|
| 6308 |
"loss": 1.7891,
|
| 6309 |
"step": 13500
|
| 6310 |
+
},
|
| 6311 |
+
{
|
| 6312 |
+
"epoch": 0.026318454270346903,
|
| 6313 |
+
"grad_norm": 5.705179214477539,
|
| 6314 |
+
"learning_rate": 2.9948757024626645e-05,
|
| 6315 |
+
"loss": 1.8502,
|
| 6316 |
+
"step": 13515
|
| 6317 |
+
},
|
| 6318 |
+
{
|
| 6319 |
+
"epoch": 0.026347664541457164,
|
| 6320 |
+
"grad_norm": 4.427610397338867,
|
| 6321 |
+
"learning_rate": 2.994864327959879e-05,
|
| 6322 |
+
"loss": 1.746,
|
| 6323 |
+
"step": 13530
|
| 6324 |
+
},
|
| 6325 |
+
{
|
| 6326 |
+
"epoch": 0.026376874812567428,
|
| 6327 |
+
"grad_norm": 2.9682793617248535,
|
| 6328 |
+
"learning_rate": 2.994852940868631e-05,
|
| 6329 |
+
"loss": 1.8766,
|
| 6330 |
+
"step": 13545
|
| 6331 |
+
},
|
| 6332 |
+
{
|
| 6333 |
+
"epoch": 0.02640608508367769,
|
| 6334 |
+
"grad_norm": 2.8406543731689453,
|
| 6335 |
+
"learning_rate": 2.9948415411890164e-05,
|
| 6336 |
+
"loss": 1.8637,
|
| 6337 |
+
"step": 13560
|
| 6338 |
+
},
|
| 6339 |
+
{
|
| 6340 |
+
"epoch": 0.026435295354787954,
|
| 6341 |
+
"grad_norm": 2.9661149978637695,
|
| 6342 |
+
"learning_rate": 2.9948301289211308e-05,
|
| 6343 |
+
"loss": 1.7703,
|
| 6344 |
+
"step": 13575
|
| 6345 |
+
},
|
| 6346 |
+
{
|
| 6347 |
+
"epoch": 0.026464505625898214,
|
| 6348 |
+
"grad_norm": 2.961155652999878,
|
| 6349 |
+
"learning_rate": 2.99481870406507e-05,
|
| 6350 |
+
"loss": 1.8123,
|
| 6351 |
+
"step": 13590
|
| 6352 |
+
},
|
| 6353 |
+
{
|
| 6354 |
+
"epoch": 0.02649371589700848,
|
| 6355 |
+
"grad_norm": 3.7241668701171875,
|
| 6356 |
+
"learning_rate": 2.9948072666209308e-05,
|
| 6357 |
+
"loss": 2.018,
|
| 6358 |
+
"step": 13605
|
| 6359 |
+
},
|
| 6360 |
+
{
|
| 6361 |
+
"epoch": 0.026522926168118743,
|
| 6362 |
+
"grad_norm": 2.8102498054504395,
|
| 6363 |
+
"learning_rate": 2.9947958165888096e-05,
|
| 6364 |
+
"loss": 1.8577,
|
| 6365 |
+
"step": 13620
|
| 6366 |
+
},
|
| 6367 |
+
{
|
| 6368 |
+
"epoch": 0.026552136439229004,
|
| 6369 |
+
"grad_norm": 2.061007022857666,
|
| 6370 |
+
"learning_rate": 2.9947843539688027e-05,
|
| 6371 |
+
"loss": 1.9684,
|
| 6372 |
+
"step": 13635
|
| 6373 |
+
},
|
| 6374 |
+
{
|
| 6375 |
+
"epoch": 0.02658134671033927,
|
| 6376 |
+
"grad_norm": 4.699859619140625,
|
| 6377 |
+
"learning_rate": 2.994772878761006e-05,
|
| 6378 |
+
"loss": 1.9438,
|
| 6379 |
+
"step": 13650
|
| 6380 |
+
},
|
| 6381 |
+
{
|
| 6382 |
+
"epoch": 0.02661055698144953,
|
| 6383 |
+
"grad_norm": 5.8805952072143555,
|
| 6384 |
+
"learning_rate": 2.994761390965517e-05,
|
| 6385 |
+
"loss": 1.8862,
|
| 6386 |
+
"step": 13665
|
| 6387 |
+
},
|
| 6388 |
+
{
|
| 6389 |
+
"epoch": 0.026639767252559794,
|
| 6390 |
+
"grad_norm": 3.6178531646728516,
|
| 6391 |
+
"learning_rate": 2.994749890582432e-05,
|
| 6392 |
+
"loss": 1.9754,
|
| 6393 |
+
"step": 13680
|
| 6394 |
+
},
|
| 6395 |
+
{
|
| 6396 |
+
"epoch": 0.026668977523670055,
|
| 6397 |
+
"grad_norm": 2.891448497772217,
|
| 6398 |
+
"learning_rate": 2.9947383776118482e-05,
|
| 6399 |
+
"loss": 1.8838,
|
| 6400 |
+
"step": 13695
|
| 6401 |
+
},
|
| 6402 |
+
{
|
| 6403 |
+
"epoch": 0.02669818779478032,
|
| 6404 |
+
"grad_norm": 2.5380797386169434,
|
| 6405 |
+
"learning_rate": 2.994726852053862e-05,
|
| 6406 |
+
"loss": 2.0006,
|
| 6407 |
+
"step": 13710
|
| 6408 |
+
},
|
| 6409 |
+
{
|
| 6410 |
+
"epoch": 0.02672739806589058,
|
| 6411 |
+
"grad_norm": 3.083801031112671,
|
| 6412 |
+
"learning_rate": 2.994715313908571e-05,
|
| 6413 |
+
"loss": 1.9287,
|
| 6414 |
+
"step": 13725
|
| 6415 |
+
},
|
| 6416 |
+
{
|
| 6417 |
+
"epoch": 0.026756608337000845,
|
| 6418 |
+
"grad_norm": 3.9220306873321533,
|
| 6419 |
+
"learning_rate": 2.9947037631760717e-05,
|
| 6420 |
+
"loss": 2.0063,
|
| 6421 |
+
"step": 13740
|
| 6422 |
+
},
|
| 6423 |
+
{
|
| 6424 |
+
"epoch": 0.02678581860811111,
|
| 6425 |
+
"grad_norm": 2.41329288482666,
|
| 6426 |
+
"learning_rate": 2.994692199856462e-05,
|
| 6427 |
+
"loss": 1.7779,
|
| 6428 |
+
"step": 13755
|
| 6429 |
+
},
|
| 6430 |
+
{
|
| 6431 |
+
"epoch": 0.02681502887922137,
|
| 6432 |
+
"grad_norm": 3.137281656265259,
|
| 6433 |
+
"learning_rate": 2.9946806239498392e-05,
|
| 6434 |
+
"loss": 1.7686,
|
| 6435 |
+
"step": 13770
|
| 6436 |
+
},
|
| 6437 |
+
{
|
| 6438 |
+
"epoch": 0.026844239150331635,
|
| 6439 |
+
"grad_norm": 3.8897507190704346,
|
| 6440 |
+
"learning_rate": 2.994669035456301e-05,
|
| 6441 |
+
"loss": 1.9879,
|
| 6442 |
+
"step": 13785
|
| 6443 |
+
},
|
| 6444 |
+
{
|
| 6445 |
+
"epoch": 0.026873449421441896,
|
| 6446 |
+
"grad_norm": 2.888145685195923,
|
| 6447 |
+
"learning_rate": 2.994657434375944e-05,
|
| 6448 |
+
"loss": 2.0012,
|
| 6449 |
+
"step": 13800
|
| 6450 |
+
},
|
| 6451 |
+
{
|
| 6452 |
+
"epoch": 0.02690265969255216,
|
| 6453 |
+
"grad_norm": 2.683145523071289,
|
| 6454 |
+
"learning_rate": 2.9946458207088667e-05,
|
| 6455 |
+
"loss": 1.8579,
|
| 6456 |
+
"step": 13815
|
| 6457 |
+
},
|
| 6458 |
+
{
|
| 6459 |
+
"epoch": 0.02693186996366242,
|
| 6460 |
+
"grad_norm": 2.5023186206817627,
|
| 6461 |
+
"learning_rate": 2.9946341944551668e-05,
|
| 6462 |
+
"loss": 1.8899,
|
| 6463 |
+
"step": 13830
|
| 6464 |
+
},
|
| 6465 |
+
{
|
| 6466 |
+
"epoch": 0.026961080234772686,
|
| 6467 |
+
"grad_norm": 4.522122383117676,
|
| 6468 |
+
"learning_rate": 2.994622555614942e-05,
|
| 6469 |
+
"loss": 1.8373,
|
| 6470 |
+
"step": 13845
|
| 6471 |
+
},
|
| 6472 |
+
{
|
| 6473 |
+
"epoch": 0.02699029050588295,
|
| 6474 |
+
"grad_norm": 1.9197810888290405,
|
| 6475 |
+
"learning_rate": 2.9946109041882902e-05,
|
| 6476 |
+
"loss": 1.874,
|
| 6477 |
+
"step": 13860
|
| 6478 |
+
},
|
| 6479 |
+
{
|
| 6480 |
+
"epoch": 0.02701950077699321,
|
| 6481 |
+
"grad_norm": 2.0907135009765625,
|
| 6482 |
+
"learning_rate": 2.9945992401753103e-05,
|
| 6483 |
+
"loss": 1.9878,
|
| 6484 |
+
"step": 13875
|
| 6485 |
+
},
|
| 6486 |
+
{
|
| 6487 |
+
"epoch": 0.027048711048103476,
|
| 6488 |
+
"grad_norm": 3.0691592693328857,
|
| 6489 |
+
"learning_rate": 2.9945875635761e-05,
|
| 6490 |
+
"loss": 1.8859,
|
| 6491 |
+
"step": 13890
|
| 6492 |
+
},
|
| 6493 |
+
{
|
| 6494 |
+
"epoch": 0.027077921319213737,
|
| 6495 |
+
"grad_norm": 2.0707552433013916,
|
| 6496 |
+
"learning_rate": 2.9945758743907573e-05,
|
| 6497 |
+
"loss": 1.7612,
|
| 6498 |
+
"step": 13905
|
| 6499 |
+
},
|
| 6500 |
+
{
|
| 6501 |
+
"epoch": 0.027107131590324,
|
| 6502 |
+
"grad_norm": 2.2770462036132812,
|
| 6503 |
+
"learning_rate": 2.994564172619381e-05,
|
| 6504 |
+
"loss": 1.8028,
|
| 6505 |
+
"step": 13920
|
| 6506 |
+
},
|
| 6507 |
+
{
|
| 6508 |
+
"epoch": 0.027136341861434262,
|
| 6509 |
+
"grad_norm": 2.681814193725586,
|
| 6510 |
+
"learning_rate": 2.9945524582620695e-05,
|
| 6511 |
+
"loss": 1.7967,
|
| 6512 |
+
"step": 13935
|
| 6513 |
+
},
|
| 6514 |
+
{
|
| 6515 |
+
"epoch": 0.027165552132544526,
|
| 6516 |
+
"grad_norm": 3.0529186725616455,
|
| 6517 |
+
"learning_rate": 2.994540731318922e-05,
|
| 6518 |
+
"loss": 1.7972,
|
| 6519 |
+
"step": 13950
|
| 6520 |
+
},
|
| 6521 |
+
{
|
| 6522 |
+
"epoch": 0.027194762403654787,
|
| 6523 |
+
"grad_norm": 3.369091033935547,
|
| 6524 |
+
"learning_rate": 2.9945289917900368e-05,
|
| 6525 |
+
"loss": 1.8092,
|
| 6526 |
+
"step": 13965
|
| 6527 |
+
},
|
| 6528 |
+
{
|
| 6529 |
+
"epoch": 0.027223972674765052,
|
| 6530 |
+
"grad_norm": 2.190134048461914,
|
| 6531 |
+
"learning_rate": 2.9945172396755124e-05,
|
| 6532 |
+
"loss": 2.0228,
|
| 6533 |
+
"step": 13980
|
| 6534 |
+
},
|
| 6535 |
+
{
|
| 6536 |
+
"epoch": 0.027253182945875316,
|
| 6537 |
+
"grad_norm": 2.805100202560425,
|
| 6538 |
+
"learning_rate": 2.9945054749754483e-05,
|
| 6539 |
+
"loss": 1.9312,
|
| 6540 |
+
"step": 13995
|
| 6541 |
+
},
|
| 6542 |
+
{
|
| 6543 |
+
"epoch": 0.027282393216985577,
|
| 6544 |
+
"grad_norm": 2.195697546005249,
|
| 6545 |
+
"learning_rate": 2.9944936976899433e-05,
|
| 6546 |
+
"loss": 1.9791,
|
| 6547 |
+
"step": 14010
|
| 6548 |
+
},
|
| 6549 |
+
{
|
| 6550 |
+
"epoch": 0.027311603488095842,
|
| 6551 |
+
"grad_norm": 1.723713755607605,
|
| 6552 |
+
"learning_rate": 2.9944819078190967e-05,
|
| 6553 |
+
"loss": 1.8542,
|
| 6554 |
+
"step": 14025
|
| 6555 |
+
},
|
| 6556 |
+
{
|
| 6557 |
+
"epoch": 0.027340813759206103,
|
| 6558 |
+
"grad_norm": 2.633101463317871,
|
| 6559 |
+
"learning_rate": 2.9944701053630075e-05,
|
| 6560 |
+
"loss": 1.8127,
|
| 6561 |
+
"step": 14040
|
| 6562 |
+
},
|
| 6563 |
+
{
|
| 6564 |
+
"epoch": 0.027370024030316367,
|
| 6565 |
+
"grad_norm": 1.9390171766281128,
|
| 6566 |
+
"learning_rate": 2.9944582903217756e-05,
|
| 6567 |
+
"loss": 1.9183,
|
| 6568 |
+
"step": 14055
|
| 6569 |
+
},
|
| 6570 |
+
{
|
| 6571 |
+
"epoch": 0.027399234301426628,
|
| 6572 |
+
"grad_norm": 3.9491968154907227,
|
| 6573 |
+
"learning_rate": 2.9944464626955003e-05,
|
| 6574 |
+
"loss": 2.0849,
|
| 6575 |
+
"step": 14070
|
| 6576 |
+
},
|
| 6577 |
+
{
|
| 6578 |
+
"epoch": 0.027428444572536893,
|
| 6579 |
+
"grad_norm": 2.4679179191589355,
|
| 6580 |
+
"learning_rate": 2.9944346224842812e-05,
|
| 6581 |
+
"loss": 1.9285,
|
| 6582 |
+
"step": 14085
|
| 6583 |
+
},
|
| 6584 |
+
{
|
| 6585 |
+
"epoch": 0.027457654843647157,
|
| 6586 |
+
"grad_norm": 2.999509334564209,
|
| 6587 |
+
"learning_rate": 2.994422769688218e-05,
|
| 6588 |
+
"loss": 1.9523,
|
| 6589 |
+
"step": 14100
|
| 6590 |
+
},
|
| 6591 |
+
{
|
| 6592 |
+
"epoch": 0.027486865114757418,
|
| 6593 |
+
"grad_norm": 3.8798091411590576,
|
| 6594 |
+
"learning_rate": 2.9944109043074104e-05,
|
| 6595 |
+
"loss": 1.9014,
|
| 6596 |
+
"step": 14115
|
| 6597 |
+
},
|
| 6598 |
+
{
|
| 6599 |
+
"epoch": 0.027516075385867683,
|
| 6600 |
+
"grad_norm": 2.5288240909576416,
|
| 6601 |
+
"learning_rate": 2.9943990263419582e-05,
|
| 6602 |
+
"loss": 2.1135,
|
| 6603 |
+
"step": 14130
|
| 6604 |
+
},
|
| 6605 |
+
{
|
| 6606 |
+
"epoch": 0.027545285656977943,
|
| 6607 |
+
"grad_norm": 2.2120304107666016,
|
| 6608 |
+
"learning_rate": 2.994387135791962e-05,
|
| 6609 |
+
"loss": 1.7418,
|
| 6610 |
+
"step": 14145
|
| 6611 |
+
},
|
| 6612 |
+
{
|
| 6613 |
+
"epoch": 0.027574495928088208,
|
| 6614 |
+
"grad_norm": 2.805328607559204,
|
| 6615 |
+
"learning_rate": 2.994375232657521e-05,
|
| 6616 |
+
"loss": 1.8776,
|
| 6617 |
+
"step": 14160
|
| 6618 |
+
},
|
| 6619 |
+
{
|
| 6620 |
+
"epoch": 0.02760370619919847,
|
| 6621 |
+
"grad_norm": 2.8841097354888916,
|
| 6622 |
+
"learning_rate": 2.9943633169387365e-05,
|
| 6623 |
+
"loss": 1.9106,
|
| 6624 |
+
"step": 14175
|
| 6625 |
+
},
|
| 6626 |
+
{
|
| 6627 |
+
"epoch": 0.027632916470308733,
|
| 6628 |
+
"grad_norm": 1.8887025117874146,
|
| 6629 |
+
"learning_rate": 2.994351388635708e-05,
|
| 6630 |
+
"loss": 1.8916,
|
| 6631 |
+
"step": 14190
|
| 6632 |
+
},
|
| 6633 |
+
{
|
| 6634 |
+
"epoch": 0.027662126741418994,
|
| 6635 |
+
"grad_norm": 2.8623757362365723,
|
| 6636 |
+
"learning_rate": 2.9943394477485363e-05,
|
| 6637 |
+
"loss": 1.8735,
|
| 6638 |
+
"step": 14205
|
| 6639 |
+
},
|
| 6640 |
+
{
|
| 6641 |
+
"epoch": 0.02769133701252926,
|
| 6642 |
+
"grad_norm": 3.1046249866485596,
|
| 6643 |
+
"learning_rate": 2.994327494277322e-05,
|
| 6644 |
+
"loss": 1.9124,
|
| 6645 |
+
"step": 14220
|
| 6646 |
+
},
|
| 6647 |
+
{
|
| 6648 |
+
"epoch": 0.027720547283639523,
|
| 6649 |
+
"grad_norm": 2.653933525085449,
|
| 6650 |
+
"learning_rate": 2.9943155282221663e-05,
|
| 6651 |
+
"loss": 1.9387,
|
| 6652 |
+
"step": 14235
|
| 6653 |
+
},
|
| 6654 |
+
{
|
| 6655 |
+
"epoch": 0.027749757554749784,
|
| 6656 |
+
"grad_norm": 2.975820779800415,
|
| 6657 |
+
"learning_rate": 2.9943035495831688e-05,
|
| 6658 |
+
"loss": 1.8232,
|
| 6659 |
+
"step": 14250
|
| 6660 |
+
},
|
| 6661 |
+
{
|
| 6662 |
+
"epoch": 0.02777896782586005,
|
| 6663 |
+
"grad_norm": 5.906015396118164,
|
| 6664 |
+
"learning_rate": 2.9942915583604307e-05,
|
| 6665 |
+
"loss": 1.9167,
|
| 6666 |
+
"step": 14265
|
| 6667 |
+
},
|
| 6668 |
+
{
|
| 6669 |
+
"epoch": 0.02780817809697031,
|
| 6670 |
+
"grad_norm": 2.592456102371216,
|
| 6671 |
+
"learning_rate": 2.994279554554054e-05,
|
| 6672 |
+
"loss": 1.7433,
|
| 6673 |
+
"step": 14280
|
| 6674 |
+
},
|
| 6675 |
+
{
|
| 6676 |
+
"epoch": 0.027837388368080574,
|
| 6677 |
+
"grad_norm": 5.042680263519287,
|
| 6678 |
+
"learning_rate": 2.994267538164138e-05,
|
| 6679 |
+
"loss": 1.7878,
|
| 6680 |
+
"step": 14295
|
| 6681 |
+
},
|
| 6682 |
+
{
|
| 6683 |
+
"epoch": 0.027866598639190835,
|
| 6684 |
+
"grad_norm": 4.092184066772461,
|
| 6685 |
+
"learning_rate": 2.9942555091907853e-05,
|
| 6686 |
+
"loss": 1.6955,
|
| 6687 |
+
"step": 14310
|
| 6688 |
+
},
|
| 6689 |
+
{
|
| 6690 |
+
"epoch": 0.0278958089103011,
|
| 6691 |
+
"grad_norm": 4.623755931854248,
|
| 6692 |
+
"learning_rate": 2.994243467634097e-05,
|
| 6693 |
+
"loss": 1.7866,
|
| 6694 |
+
"step": 14325
|
| 6695 |
+
},
|
| 6696 |
+
{
|
| 6697 |
+
"epoch": 0.027925019181411364,
|
| 6698 |
+
"grad_norm": 3.042306661605835,
|
| 6699 |
+
"learning_rate": 2.994231413494174e-05,
|
| 6700 |
+
"loss": 1.8891,
|
| 6701 |
+
"step": 14340
|
| 6702 |
+
},
|
| 6703 |
+
{
|
| 6704 |
+
"epoch": 0.027954229452521625,
|
| 6705 |
+
"grad_norm": 2.784275531768799,
|
| 6706 |
+
"learning_rate": 2.9942193467711184e-05,
|
| 6707 |
+
"loss": 2.0112,
|
| 6708 |
+
"step": 14355
|
| 6709 |
+
},
|
| 6710 |
+
{
|
| 6711 |
+
"epoch": 0.02798343972363189,
|
| 6712 |
+
"grad_norm": 1.9308695793151855,
|
| 6713 |
+
"learning_rate": 2.9942072674650317e-05,
|
| 6714 |
+
"loss": 1.7964,
|
| 6715 |
+
"step": 14370
|
| 6716 |
+
},
|
| 6717 |
+
{
|
| 6718 |
+
"epoch": 0.02801264999474215,
|
| 6719 |
+
"grad_norm": 3.7377004623413086,
|
| 6720 |
+
"learning_rate": 2.994195175576015e-05,
|
| 6721 |
+
"loss": 1.8661,
|
| 6722 |
+
"step": 14385
|
| 6723 |
+
},
|
| 6724 |
+
{
|
| 6725 |
+
"epoch": 0.028041860265852415,
|
| 6726 |
+
"grad_norm": 2.484870195388794,
|
| 6727 |
+
"learning_rate": 2.994183071104171e-05,
|
| 6728 |
+
"loss": 1.7358,
|
| 6729 |
+
"step": 14400
|
| 6730 |
+
},
|
| 6731 |
+
{
|
| 6732 |
+
"epoch": 0.028071070536962676,
|
| 6733 |
+
"grad_norm": 2.6344974040985107,
|
| 6734 |
+
"learning_rate": 2.9941709540496013e-05,
|
| 6735 |
+
"loss": 1.9183,
|
| 6736 |
+
"step": 14415
|
| 6737 |
+
},
|
| 6738 |
+
{
|
| 6739 |
+
"epoch": 0.02810028080807294,
|
| 6740 |
+
"grad_norm": 2.168701410293579,
|
| 6741 |
+
"learning_rate": 2.9941588244124072e-05,
|
| 6742 |
+
"loss": 1.999,
|
| 6743 |
+
"step": 14430
|
| 6744 |
+
},
|
| 6745 |
+
{
|
| 6746 |
+
"epoch": 0.028129491079183205,
|
| 6747 |
+
"grad_norm": 2.986727476119995,
|
| 6748 |
+
"learning_rate": 2.994146682192692e-05,
|
| 6749 |
+
"loss": 1.8344,
|
| 6750 |
+
"step": 14445
|
| 6751 |
+
},
|
| 6752 |
+
{
|
| 6753 |
+
"epoch": 0.028158701350293466,
|
| 6754 |
+
"grad_norm": 3.3715713024139404,
|
| 6755 |
+
"learning_rate": 2.9941345273905573e-05,
|
| 6756 |
+
"loss": 2.0468,
|
| 6757 |
+
"step": 14460
|
| 6758 |
+
},
|
| 6759 |
+
{
|
| 6760 |
+
"epoch": 0.02818791162140373,
|
| 6761 |
+
"grad_norm": 2.2077038288116455,
|
| 6762 |
+
"learning_rate": 2.9941223600061054e-05,
|
| 6763 |
+
"loss": 2.0255,
|
| 6764 |
+
"step": 14475
|
| 6765 |
+
},
|
| 6766 |
+
{
|
| 6767 |
+
"epoch": 0.02821712189251399,
|
| 6768 |
+
"grad_norm": 3.4651224613189697,
|
| 6769 |
+
"learning_rate": 2.994110180039439e-05,
|
| 6770 |
+
"loss": 1.8604,
|
| 6771 |
+
"step": 14490
|
| 6772 |
+
},
|
| 6773 |
+
{
|
| 6774 |
+
"epoch": 0.028246332163624256,
|
| 6775 |
+
"grad_norm": 2.0584287643432617,
|
| 6776 |
+
"learning_rate": 2.994097987490661e-05,
|
| 6777 |
+
"loss": 2.0636,
|
| 6778 |
+
"step": 14505
|
| 6779 |
+
},
|
| 6780 |
+
{
|
| 6781 |
+
"epoch": 0.028275542434734516,
|
| 6782 |
+
"grad_norm": 3.1285014152526855,
|
| 6783 |
+
"learning_rate": 2.9940857823598736e-05,
|
| 6784 |
+
"loss": 1.7656,
|
| 6785 |
+
"step": 14520
|
| 6786 |
+
},
|
| 6787 |
+
{
|
| 6788 |
+
"epoch": 0.02830475270584478,
|
| 6789 |
+
"grad_norm": 2.541280746459961,
|
| 6790 |
+
"learning_rate": 2.9940735646471793e-05,
|
| 6791 |
+
"loss": 1.8682,
|
| 6792 |
+
"step": 14535
|
| 6793 |
+
},
|
| 6794 |
+
{
|
| 6795 |
+
"epoch": 0.028333962976955042,
|
| 6796 |
+
"grad_norm": 2.4748847484588623,
|
| 6797 |
+
"learning_rate": 2.9940613343526817e-05,
|
| 6798 |
+
"loss": 2.0047,
|
| 6799 |
+
"step": 14550
|
| 6800 |
+
},
|
| 6801 |
+
{
|
| 6802 |
+
"epoch": 0.028363173248065306,
|
| 6803 |
+
"grad_norm": 3.2105560302734375,
|
| 6804 |
+
"learning_rate": 2.9940490914764834e-05,
|
| 6805 |
+
"loss": 1.7709,
|
| 6806 |
+
"step": 14565
|
| 6807 |
+
},
|
| 6808 |
+
{
|
| 6809 |
+
"epoch": 0.02839238351917557,
|
| 6810 |
+
"grad_norm": 3.491591215133667,
|
| 6811 |
+
"learning_rate": 2.9940368360186878e-05,
|
| 6812 |
+
"loss": 1.786,
|
| 6813 |
+
"step": 14580
|
| 6814 |
+
},
|
| 6815 |
+
{
|
| 6816 |
+
"epoch": 0.028421593790285832,
|
| 6817 |
+
"grad_norm": 3.315342903137207,
|
| 6818 |
+
"learning_rate": 2.9940245679793978e-05,
|
| 6819 |
+
"loss": 1.917,
|
| 6820 |
+
"step": 14595
|
| 6821 |
+
},
|
| 6822 |
+
{
|
| 6823 |
+
"epoch": 0.028450804061396096,
|
| 6824 |
+
"grad_norm": 1.7594997882843018,
|
| 6825 |
+
"learning_rate": 2.9940122873587164e-05,
|
| 6826 |
+
"loss": 1.877,
|
| 6827 |
+
"step": 14610
|
| 6828 |
+
},
|
| 6829 |
+
{
|
| 6830 |
+
"epoch": 0.028480014332506357,
|
| 6831 |
+
"grad_norm": 2.442725896835327,
|
| 6832 |
+
"learning_rate": 2.9939999941567474e-05,
|
| 6833 |
+
"loss": 1.9577,
|
| 6834 |
+
"step": 14625
|
| 6835 |
+
},
|
| 6836 |
+
{
|
| 6837 |
+
"epoch": 0.02850922460361662,
|
| 6838 |
+
"grad_norm": 3.146977663040161,
|
| 6839 |
+
"learning_rate": 2.993987688373595e-05,
|
| 6840 |
+
"loss": 1.8722,
|
| 6841 |
+
"step": 14640
|
| 6842 |
+
},
|
| 6843 |
+
{
|
| 6844 |
+
"epoch": 0.028538434874726883,
|
| 6845 |
+
"grad_norm": 2.6678929328918457,
|
| 6846 |
+
"learning_rate": 2.9939753700093618e-05,
|
| 6847 |
+
"loss": 1.659,
|
| 6848 |
+
"step": 14655
|
| 6849 |
+
},
|
| 6850 |
+
{
|
| 6851 |
+
"epoch": 0.028567645145837147,
|
| 6852 |
+
"grad_norm": 2.4769906997680664,
|
| 6853 |
+
"learning_rate": 2.9939630390641518e-05,
|
| 6854 |
+
"loss": 1.8257,
|
| 6855 |
+
"step": 14670
|
| 6856 |
+
},
|
| 6857 |
+
{
|
| 6858 |
+
"epoch": 0.02859685541694741,
|
| 6859 |
+
"grad_norm": 2.9314770698547363,
|
| 6860 |
+
"learning_rate": 2.993950695538069e-05,
|
| 6861 |
+
"loss": 1.8004,
|
| 6862 |
+
"step": 14685
|
| 6863 |
+
},
|
| 6864 |
+
{
|
| 6865 |
+
"epoch": 0.028626065688057672,
|
| 6866 |
+
"grad_norm": 3.2279980182647705,
|
| 6867 |
+
"learning_rate": 2.993938339431217e-05,
|
| 6868 |
+
"loss": 1.9438,
|
| 6869 |
+
"step": 14700
|
| 6870 |
+
},
|
| 6871 |
+
{
|
| 6872 |
+
"epoch": 0.028655275959167937,
|
| 6873 |
+
"grad_norm": 2.8929495811462402,
|
| 6874 |
+
"learning_rate": 2.9939259707437002e-05,
|
| 6875 |
+
"loss": 1.7995,
|
| 6876 |
+
"step": 14715
|
| 6877 |
+
},
|
| 6878 |
+
{
|
| 6879 |
+
"epoch": 0.028684486230278198,
|
| 6880 |
+
"grad_norm": 4.861998558044434,
|
| 6881 |
+
"learning_rate": 2.9939135894756232e-05,
|
| 6882 |
+
"loss": 1.8188,
|
| 6883 |
+
"step": 14730
|
| 6884 |
+
},
|
| 6885 |
+
{
|
| 6886 |
+
"epoch": 0.028713696501388462,
|
| 6887 |
+
"grad_norm": 5.37394905090332,
|
| 6888 |
+
"learning_rate": 2.9939011956270893e-05,
|
| 6889 |
+
"loss": 2.061,
|
| 6890 |
+
"step": 14745
|
| 6891 |
+
},
|
| 6892 |
+
{
|
| 6893 |
+
"epoch": 0.028742906772498723,
|
| 6894 |
+
"grad_norm": 2.2253520488739014,
|
| 6895 |
+
"learning_rate": 2.9938887891982035e-05,
|
| 6896 |
+
"loss": 1.963,
|
| 6897 |
+
"step": 14760
|
| 6898 |
+
},
|
| 6899 |
+
{
|
| 6900 |
+
"epoch": 0.028772117043608988,
|
| 6901 |
+
"grad_norm": 3.424954414367676,
|
| 6902 |
+
"learning_rate": 2.99387637018907e-05,
|
| 6903 |
+
"loss": 2.0077,
|
| 6904 |
+
"step": 14775
|
| 6905 |
+
},
|
| 6906 |
+
{
|
| 6907 |
+
"epoch": 0.02880132731471925,
|
| 6908 |
+
"grad_norm": 2.8398706912994385,
|
| 6909 |
+
"learning_rate": 2.9938639385997934e-05,
|
| 6910 |
+
"loss": 2.0516,
|
| 6911 |
+
"step": 14790
|
| 6912 |
+
},
|
| 6913 |
+
{
|
| 6914 |
+
"epoch": 0.028830537585829513,
|
| 6915 |
+
"grad_norm": 2.371492385864258,
|
| 6916 |
+
"learning_rate": 2.9938514944304788e-05,
|
| 6917 |
+
"loss": 1.8057,
|
| 6918 |
+
"step": 14805
|
| 6919 |
+
},
|
| 6920 |
+
{
|
| 6921 |
+
"epoch": 0.028859747856939778,
|
| 6922 |
+
"grad_norm": 1.870301365852356,
|
| 6923 |
+
"learning_rate": 2.9938390376812304e-05,
|
| 6924 |
+
"loss": 1.8335,
|
| 6925 |
+
"step": 14820
|
| 6926 |
+
},
|
| 6927 |
+
{
|
| 6928 |
+
"epoch": 0.02888895812805004,
|
| 6929 |
+
"grad_norm": 3.1508800983428955,
|
| 6930 |
+
"learning_rate": 2.9938265683521533e-05,
|
| 6931 |
+
"loss": 2.0272,
|
| 6932 |
+
"step": 14835
|
| 6933 |
+
},
|
| 6934 |
+
{
|
| 6935 |
+
"epoch": 0.028918168399160303,
|
| 6936 |
+
"grad_norm": 2.8456640243530273,
|
| 6937 |
+
"learning_rate": 2.9938140864433528e-05,
|
| 6938 |
+
"loss": 1.889,
|
| 6939 |
+
"step": 14850
|
| 6940 |
+
},
|
| 6941 |
+
{
|
| 6942 |
+
"epoch": 0.028947378670270564,
|
| 6943 |
+
"grad_norm": 2.3040804862976074,
|
| 6944 |
+
"learning_rate": 2.9938015919549337e-05,
|
| 6945 |
+
"loss": 1.9274,
|
| 6946 |
+
"step": 14865
|
| 6947 |
+
},
|
| 6948 |
+
{
|
| 6949 |
+
"epoch": 0.02897658894138083,
|
| 6950 |
+
"grad_norm": 3.075559139251709,
|
| 6951 |
+
"learning_rate": 2.9937890848870012e-05,
|
| 6952 |
+
"loss": 1.9239,
|
| 6953 |
+
"step": 14880
|
| 6954 |
+
},
|
| 6955 |
+
{
|
| 6956 |
+
"epoch": 0.02900579921249109,
|
| 6957 |
+
"grad_norm": 2.4840190410614014,
|
| 6958 |
+
"learning_rate": 2.9937765652396608e-05,
|
| 6959 |
+
"loss": 1.9836,
|
| 6960 |
+
"step": 14895
|
| 6961 |
+
},
|
| 6962 |
+
{
|
| 6963 |
+
"epoch": 0.029035009483601354,
|
| 6964 |
+
"grad_norm": 2.915515422821045,
|
| 6965 |
+
"learning_rate": 2.9937640330130182e-05,
|
| 6966 |
+
"loss": 2.0196,
|
| 6967 |
+
"step": 14910
|
| 6968 |
+
},
|
| 6969 |
+
{
|
| 6970 |
+
"epoch": 0.02906421975471162,
|
| 6971 |
+
"grad_norm": 2.684401035308838,
|
| 6972 |
+
"learning_rate": 2.993751488207178e-05,
|
| 6973 |
+
"loss": 1.9699,
|
| 6974 |
+
"step": 14925
|
| 6975 |
+
},
|
| 6976 |
+
{
|
| 6977 |
+
"epoch": 0.02909343002582188,
|
| 6978 |
+
"grad_norm": 1.7906841039657593,
|
| 6979 |
+
"learning_rate": 2.9937389308222468e-05,
|
| 6980 |
+
"loss": 1.9435,
|
| 6981 |
+
"step": 14940
|
| 6982 |
+
},
|
| 6983 |
+
{
|
| 6984 |
+
"epoch": 0.029122640296932144,
|
| 6985 |
+
"grad_norm": 2.7629384994506836,
|
| 6986 |
+
"learning_rate": 2.9937263608583297e-05,
|
| 6987 |
+
"loss": 1.9266,
|
| 6988 |
+
"step": 14955
|
| 6989 |
+
},
|
| 6990 |
+
{
|
| 6991 |
+
"epoch": 0.029151850568042405,
|
| 6992 |
+
"grad_norm": 3.65447735786438,
|
| 6993 |
+
"learning_rate": 2.9937137783155326e-05,
|
| 6994 |
+
"loss": 1.8818,
|
| 6995 |
+
"step": 14970
|
| 6996 |
+
},
|
| 6997 |
+
{
|
| 6998 |
+
"epoch": 0.02918106083915267,
|
| 6999 |
+
"grad_norm": 2.684885025024414,
|
| 7000 |
+
"learning_rate": 2.993701183193962e-05,
|
| 7001 |
+
"loss": 1.744,
|
| 7002 |
+
"step": 14985
|
| 7003 |
+
},
|
| 7004 |
+
{
|
| 7005 |
+
"epoch": 0.02921027111026293,
|
| 7006 |
+
"grad_norm": 2.995678424835205,
|
| 7007 |
+
"learning_rate": 2.9936885754937237e-05,
|
| 7008 |
+
"loss": 1.868,
|
| 7009 |
+
"step": 15000
|
| 7010 |
}
|
| 7011 |
],
|
| 7012 |
"logging_steps": 15,
|
|
|
|
| 7026 |
"attributes": {}
|
| 7027 |
}
|
| 7028 |
},
|
| 7029 |
+
"total_flos": 2.384291701225267e+16,
|
| 7030 |
"train_batch_size": 4,
|
| 7031 |
"trial_name": null,
|
| 7032 |
"trial_params": null
|
data/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5304
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ec2ab14e2ce9ef4d500efa223a38f8e5ca01386342399ff2338f06b1fa66a7a
|
| 3 |
size 5304
|