Training in progress, step 20500, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -29,13 +29,13 @@
|
|
| 29 |
"rank_pattern": {},
|
| 30 |
"revision": null,
|
| 31 |
"target_modules": [
|
| 32 |
-
"
|
|
|
|
|
|
|
| 33 |
"v_proj",
|
|
|
|
| 34 |
"gate_proj",
|
| 35 |
-
"
|
| 36 |
-
"up_proj",
|
| 37 |
-
"q_proj",
|
| 38 |
-
"k_proj"
|
| 39 |
],
|
| 40 |
"target_parameters": null,
|
| 41 |
"task_type": "CAUSAL_LM",
|
|
|
|
| 29 |
"rank_pattern": {},
|
| 30 |
"revision": null,
|
| 31 |
"target_modules": [
|
| 32 |
+
"up_proj",
|
| 33 |
+
"down_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
"v_proj",
|
| 36 |
+
"o_proj",
|
| 37 |
"gate_proj",
|
| 38 |
+
"q_proj"
|
|
|
|
|
|
|
|
|
|
| 39 |
],
|
| 40 |
"target_parameters": null,
|
| 41 |
"task_type": "CAUSAL_LM",
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 295488936
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b92923f94a27b22dd17376a2e9ae2f92c375d23bce5ed25d4d822a15e2401b19
|
| 3 |
size 295488936
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 591209083
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd414f15e0e92404216b0e8591aefdbe0332698de9ef688d83b64d6dc057cf51
|
| 3 |
size 591209083
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a5c11622183d87222d1ac6c441b6bba34fdcd36747c102761fbee73401f8dd3
|
| 3 |
size 14645
|
last-checkpoint/scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1383
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6d08a9f614a259424e39bc87627e28a7ef5fee362682f7692f690647c56f2d52
|
| 3 |
size 1383
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1761a05344bf799fe63e3a0f79c26244624274ca3d4ba6b5ca782efc799fcb0a
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 2.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -14328,6 +14328,364 @@
|
|
| 14328 |
"eval_samples_per_second": 2.718,
|
| 14329 |
"eval_steps_per_second": 2.718,
|
| 14330 |
"step": 20000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14331 |
}
|
| 14332 |
],
|
| 14333 |
"logging_steps": 10,
|
|
@@ -14347,7 +14705,7 @@
|
|
| 14347 |
"attributes": {}
|
| 14348 |
}
|
| 14349 |
},
|
| 14350 |
-
"total_flos": 1.
|
| 14351 |
"train_batch_size": 1,
|
| 14352 |
"trial_name": null,
|
| 14353 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 2.1663474113901477,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 20500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 14328 |
"eval_samples_per_second": 2.718,
|
| 14329 |
"eval_steps_per_second": 2.718,
|
| 14330 |
"step": 20000
|
| 14331 |
+
},
|
| 14332 |
+
{
|
| 14333 |
+
"epoch": 2.114562003778221,
|
| 14334 |
+
"grad_norm": 0.9563535451889038,
|
| 14335 |
+
"learning_rate": 1.0014180084488345e-05,
|
| 14336 |
+
"loss": 0.7777707576751709,
|
| 14337 |
+
"step": 20010
|
| 14338 |
+
},
|
| 14339 |
+
{
|
| 14340 |
+
"epoch": 2.1156188488315255,
|
| 14341 |
+
"grad_norm": 1.2591712474822998,
|
| 14342 |
+
"learning_rate": 9.992029441642728e-06,
|
| 14343 |
+
"loss": 0.8797025680541992,
|
| 14344 |
+
"step": 20020
|
| 14345 |
+
},
|
| 14346 |
+
{
|
| 14347 |
+
"epoch": 2.1166756938848303,
|
| 14348 |
+
"grad_norm": 1.1730784177780151,
|
| 14349 |
+
"learning_rate": 9.969897203703602e-06,
|
| 14350 |
+
"loss": 0.9072016716003418,
|
| 14351 |
+
"step": 20030
|
| 14352 |
+
},
|
| 14353 |
+
{
|
| 14354 |
+
"epoch": 2.1177325389381347,
|
| 14355 |
+
"grad_norm": 1.8978923559188843,
|
| 14356 |
+
"learning_rate": 9.947783397812685e-06,
|
| 14357 |
+
"loss": 0.8863043785095215,
|
| 14358 |
+
"step": 20040
|
| 14359 |
+
},
|
| 14360 |
+
{
|
| 14361 |
+
"epoch": 2.1187893839914396,
|
| 14362 |
+
"grad_norm": 1.2955552339553833,
|
| 14363 |
+
"learning_rate": 9.92568805108904e-06,
|
| 14364 |
+
"loss": 0.9036584854125976,
|
| 14365 |
+
"step": 20050
|
| 14366 |
+
},
|
| 14367 |
+
{
|
| 14368 |
+
"epoch": 2.119846229044744,
|
| 14369 |
+
"grad_norm": 1.629235863685608,
|
| 14370 |
+
"learning_rate": 9.903611190629145e-06,
|
| 14371 |
+
"loss": 0.7058406352996827,
|
| 14372 |
+
"step": 20060
|
| 14373 |
+
},
|
| 14374 |
+
{
|
| 14375 |
+
"epoch": 2.120903074098049,
|
| 14376 |
+
"grad_norm": 1.4561424255371094,
|
| 14377 |
+
"learning_rate": 9.881552843506772e-06,
|
| 14378 |
+
"loss": 0.8088406562805176,
|
| 14379 |
+
"step": 20070
|
| 14380 |
+
},
|
| 14381 |
+
{
|
| 14382 |
+
"epoch": 2.1219599191513536,
|
| 14383 |
+
"grad_norm": 1.0040875673294067,
|
| 14384 |
+
"learning_rate": 9.859513036773007e-06,
|
| 14385 |
+
"loss": 0.8048751831054688,
|
| 14386 |
+
"step": 20080
|
| 14387 |
+
},
|
| 14388 |
+
{
|
| 14389 |
+
"epoch": 2.123016764204658,
|
| 14390 |
+
"grad_norm": 1.5896157026290894,
|
| 14391 |
+
"learning_rate": 9.837491797456189e-06,
|
| 14392 |
+
"loss": 0.7703395366668702,
|
| 14393 |
+
"step": 20090
|
| 14394 |
+
},
|
| 14395 |
+
{
|
| 14396 |
+
"epoch": 2.124073609257963,
|
| 14397 |
+
"grad_norm": 1.8215241432189941,
|
| 14398 |
+
"learning_rate": 9.815489152561885e-06,
|
| 14399 |
+
"loss": 0.9717734336853028,
|
| 14400 |
+
"step": 20100
|
| 14401 |
+
},
|
| 14402 |
+
{
|
| 14403 |
+
"epoch": 2.1251304543112672,
|
| 14404 |
+
"grad_norm": 0.9666393995285034,
|
| 14405 |
+
"learning_rate": 9.793505129072884e-06,
|
| 14406 |
+
"loss": 0.8387363433837891,
|
| 14407 |
+
"step": 20110
|
| 14408 |
+
},
|
| 14409 |
+
{
|
| 14410 |
+
"epoch": 2.126187299364572,
|
| 14411 |
+
"grad_norm": 1.0454455614089966,
|
| 14412 |
+
"learning_rate": 9.771539753949105e-06,
|
| 14413 |
+
"loss": 0.8208915710449218,
|
| 14414 |
+
"step": 20120
|
| 14415 |
+
},
|
| 14416 |
+
{
|
| 14417 |
+
"epoch": 2.1272441444178765,
|
| 14418 |
+
"grad_norm": 1.314156413078308,
|
| 14419 |
+
"learning_rate": 9.749593054127621e-06,
|
| 14420 |
+
"loss": 0.7010626792907715,
|
| 14421 |
+
"step": 20130
|
| 14422 |
+
},
|
| 14423 |
+
{
|
| 14424 |
+
"epoch": 2.1283009894711813,
|
| 14425 |
+
"grad_norm": 1.2660032510757446,
|
| 14426 |
+
"learning_rate": 9.727665056522577e-06,
|
| 14427 |
+
"loss": 0.8174416542053222,
|
| 14428 |
+
"step": 20140
|
| 14429 |
+
},
|
| 14430 |
+
{
|
| 14431 |
+
"epoch": 2.1293578345244857,
|
| 14432 |
+
"grad_norm": 1.3227639198303223,
|
| 14433 |
+
"learning_rate": 9.705755788025228e-06,
|
| 14434 |
+
"loss": 0.8532349586486816,
|
| 14435 |
+
"step": 20150
|
| 14436 |
+
},
|
| 14437 |
+
{
|
| 14438 |
+
"epoch": 2.1304146795777905,
|
| 14439 |
+
"grad_norm": 0.8495364189147949,
|
| 14440 |
+
"learning_rate": 9.683865275503817e-06,
|
| 14441 |
+
"loss": 0.7597885131835938,
|
| 14442 |
+
"step": 20160
|
| 14443 |
+
},
|
| 14444 |
+
{
|
| 14445 |
+
"epoch": 2.131471524631095,
|
| 14446 |
+
"grad_norm": 1.1927849054336548,
|
| 14447 |
+
"learning_rate": 9.661993545803602e-06,
|
| 14448 |
+
"loss": 0.9093083381652832,
|
| 14449 |
+
"step": 20170
|
| 14450 |
+
},
|
| 14451 |
+
{
|
| 14452 |
+
"epoch": 2.1325283696843997,
|
| 14453 |
+
"grad_norm": 1.396560549736023,
|
| 14454 |
+
"learning_rate": 9.6401406257468e-06,
|
| 14455 |
+
"loss": 0.9485726356506348,
|
| 14456 |
+
"step": 20180
|
| 14457 |
+
},
|
| 14458 |
+
{
|
| 14459 |
+
"epoch": 2.133585214737704,
|
| 14460 |
+
"grad_norm": 1.5068135261535645,
|
| 14461 |
+
"learning_rate": 9.618306542132583e-06,
|
| 14462 |
+
"loss": 0.7994839668273925,
|
| 14463 |
+
"step": 20190
|
| 14464 |
+
},
|
| 14465 |
+
{
|
| 14466 |
+
"epoch": 2.134642059791009,
|
| 14467 |
+
"grad_norm": 1.2415924072265625,
|
| 14468 |
+
"learning_rate": 9.596491321737e-06,
|
| 14469 |
+
"loss": 0.901005744934082,
|
| 14470 |
+
"step": 20200
|
| 14471 |
+
},
|
| 14472 |
+
{
|
| 14473 |
+
"epoch": 2.1356989048443134,
|
| 14474 |
+
"grad_norm": 1.360911250114441,
|
| 14475 |
+
"learning_rate": 9.574694991312977e-06,
|
| 14476 |
+
"loss": 0.9780127525329589,
|
| 14477 |
+
"step": 20210
|
| 14478 |
+
},
|
| 14479 |
+
{
|
| 14480 |
+
"epoch": 2.136755749897618,
|
| 14481 |
+
"grad_norm": 1.5826635360717773,
|
| 14482 |
+
"learning_rate": 9.552917577590257e-06,
|
| 14483 |
+
"loss": 0.8111052513122559,
|
| 14484 |
+
"step": 20220
|
| 14485 |
+
},
|
| 14486 |
+
{
|
| 14487 |
+
"epoch": 2.1378125949509226,
|
| 14488 |
+
"grad_norm": 1.7501533031463623,
|
| 14489 |
+
"learning_rate": 9.531159107275422e-06,
|
| 14490 |
+
"loss": 0.7468459129333496,
|
| 14491 |
+
"step": 20230
|
| 14492 |
+
},
|
| 14493 |
+
{
|
| 14494 |
+
"epoch": 2.1388694400042274,
|
| 14495 |
+
"grad_norm": 1.2226203680038452,
|
| 14496 |
+
"learning_rate": 9.509419607051792e-06,
|
| 14497 |
+
"loss": 0.8548343658447266,
|
| 14498 |
+
"step": 20240
|
| 14499 |
+
},
|
| 14500 |
+
{
|
| 14501 |
+
"epoch": 2.139926285057532,
|
| 14502 |
+
"grad_norm": 1.2593287229537964,
|
| 14503 |
+
"learning_rate": 9.48769910357943e-06,
|
| 14504 |
+
"loss": 0.8602652549743652,
|
| 14505 |
+
"step": 20250
|
| 14506 |
+
},
|
| 14507 |
+
{
|
| 14508 |
+
"epoch": 2.1409831301108366,
|
| 14509 |
+
"grad_norm": 1.5624585151672363,
|
| 14510 |
+
"learning_rate": 9.465997623495098e-06,
|
| 14511 |
+
"loss": 0.9549854278564454,
|
| 14512 |
+
"step": 20260
|
| 14513 |
+
},
|
| 14514 |
+
{
|
| 14515 |
+
"epoch": 2.142039975164141,
|
| 14516 |
+
"grad_norm": 1.1964505910873413,
|
| 14517 |
+
"learning_rate": 9.444315193412252e-06,
|
| 14518 |
+
"loss": 0.9200518608093262,
|
| 14519 |
+
"step": 20270
|
| 14520 |
+
},
|
| 14521 |
+
{
|
| 14522 |
+
"epoch": 2.143096820217446,
|
| 14523 |
+
"grad_norm": 1.00190007686615,
|
| 14524 |
+
"learning_rate": 9.42265183992096e-06,
|
| 14525 |
+
"loss": 0.9042006492614746,
|
| 14526 |
+
"step": 20280
|
| 14527 |
+
},
|
| 14528 |
+
{
|
| 14529 |
+
"epoch": 2.1441536652707507,
|
| 14530 |
+
"grad_norm": 1.6136075258255005,
|
| 14531 |
+
"learning_rate": 9.4010075895879e-06,
|
| 14532 |
+
"loss": 0.741997766494751,
|
| 14533 |
+
"step": 20290
|
| 14534 |
+
},
|
| 14535 |
+
{
|
| 14536 |
+
"epoch": 2.145210510324055,
|
| 14537 |
+
"grad_norm": 1.4245508909225464,
|
| 14538 |
+
"learning_rate": 9.37938246895633e-06,
|
| 14539 |
+
"loss": 0.8433687210083007,
|
| 14540 |
+
"step": 20300
|
| 14541 |
+
},
|
| 14542 |
+
{
|
| 14543 |
+
"epoch": 2.14626735537736,
|
| 14544 |
+
"grad_norm": 1.1085989475250244,
|
| 14545 |
+
"learning_rate": 9.357776504546054e-06,
|
| 14546 |
+
"loss": 0.8139483451843261,
|
| 14547 |
+
"step": 20310
|
| 14548 |
+
},
|
| 14549 |
+
{
|
| 14550 |
+
"epoch": 2.1473242004306643,
|
| 14551 |
+
"grad_norm": 1.2043304443359375,
|
| 14552 |
+
"learning_rate": 9.336189722853376e-06,
|
| 14553 |
+
"loss": 0.9100760459899903,
|
| 14554 |
+
"step": 20320
|
| 14555 |
+
},
|
| 14556 |
+
{
|
| 14557 |
+
"epoch": 2.148381045483969,
|
| 14558 |
+
"grad_norm": 1.3986866474151611,
|
| 14559 |
+
"learning_rate": 9.314622150351066e-06,
|
| 14560 |
+
"loss": 0.8731118202209472,
|
| 14561 |
+
"step": 20330
|
| 14562 |
+
},
|
| 14563 |
+
{
|
| 14564 |
+
"epoch": 2.1494378905372735,
|
| 14565 |
+
"grad_norm": 1.2441811561584473,
|
| 14566 |
+
"learning_rate": 9.293073813488346e-06,
|
| 14567 |
+
"loss": 0.7809085845947266,
|
| 14568 |
+
"step": 20340
|
| 14569 |
+
},
|
| 14570 |
+
{
|
| 14571 |
+
"epoch": 2.1504947355905784,
|
| 14572 |
+
"grad_norm": 1.355765461921692,
|
| 14573 |
+
"learning_rate": 9.271544738690863e-06,
|
| 14574 |
+
"loss": 0.8749888420104981,
|
| 14575 |
+
"step": 20350
|
| 14576 |
+
},
|
| 14577 |
+
{
|
| 14578 |
+
"epoch": 2.1515515806438827,
|
| 14579 |
+
"grad_norm": 1.072855830192566,
|
| 14580 |
+
"learning_rate": 9.25003495236062e-06,
|
| 14581 |
+
"loss": 0.8202900886535645,
|
| 14582 |
+
"step": 20360
|
| 14583 |
+
},
|
| 14584 |
+
{
|
| 14585 |
+
"epoch": 2.1526084256971876,
|
| 14586 |
+
"grad_norm": 1.4892768859863281,
|
| 14587 |
+
"learning_rate": 9.228544480875975e-06,
|
| 14588 |
+
"loss": 0.8801441192626953,
|
| 14589 |
+
"step": 20370
|
| 14590 |
+
},
|
| 14591 |
+
{
|
| 14592 |
+
"epoch": 2.153665270750492,
|
| 14593 |
+
"grad_norm": 1.2602416276931763,
|
| 14594 |
+
"learning_rate": 9.207073350591589e-06,
|
| 14595 |
+
"loss": 0.8223505973815918,
|
| 14596 |
+
"step": 20380
|
| 14597 |
+
},
|
| 14598 |
+
{
|
| 14599 |
+
"epoch": 2.154722115803797,
|
| 14600 |
+
"grad_norm": 1.2211744785308838,
|
| 14601 |
+
"learning_rate": 9.18562158783843e-06,
|
| 14602 |
+
"loss": 0.9198463439941407,
|
| 14603 |
+
"step": 20390
|
| 14604 |
+
},
|
| 14605 |
+
{
|
| 14606 |
+
"epoch": 2.155778960857101,
|
| 14607 |
+
"grad_norm": 1.5020973682403564,
|
| 14608 |
+
"learning_rate": 9.164189218923694e-06,
|
| 14609 |
+
"loss": 0.8651206970214844,
|
| 14610 |
+
"step": 20400
|
| 14611 |
+
},
|
| 14612 |
+
{
|
| 14613 |
+
"epoch": 2.156835805910406,
|
| 14614 |
+
"grad_norm": 1.3377459049224854,
|
| 14615 |
+
"learning_rate": 9.142776270130793e-06,
|
| 14616 |
+
"loss": 0.8239879608154297,
|
| 14617 |
+
"step": 20410
|
| 14618 |
+
},
|
| 14619 |
+
{
|
| 14620 |
+
"epoch": 2.1578926509637104,
|
| 14621 |
+
"grad_norm": 1.5689959526062012,
|
| 14622 |
+
"learning_rate": 9.12138276771932e-06,
|
| 14623 |
+
"loss": 0.8948896408081055,
|
| 14624 |
+
"step": 20420
|
| 14625 |
+
},
|
| 14626 |
+
{
|
| 14627 |
+
"epoch": 2.1589494960170152,
|
| 14628 |
+
"grad_norm": 1.1158559322357178,
|
| 14629 |
+
"learning_rate": 9.100008737925047e-06,
|
| 14630 |
+
"loss": 0.7634482383728027,
|
| 14631 |
+
"step": 20430
|
| 14632 |
+
},
|
| 14633 |
+
{
|
| 14634 |
+
"epoch": 2.1600063410703196,
|
| 14635 |
+
"grad_norm": 1.0613447427749634,
|
| 14636 |
+
"learning_rate": 9.078654206959838e-06,
|
| 14637 |
+
"loss": 0.8469998359680175,
|
| 14638 |
+
"step": 20440
|
| 14639 |
+
},
|
| 14640 |
+
{
|
| 14641 |
+
"epoch": 2.1610631861236245,
|
| 14642 |
+
"grad_norm": 1.4286909103393555,
|
| 14643 |
+
"learning_rate": 9.057319201011655e-06,
|
| 14644 |
+
"loss": 0.7755132675170898,
|
| 14645 |
+
"step": 20450
|
| 14646 |
+
},
|
| 14647 |
+
{
|
| 14648 |
+
"epoch": 2.1621200311769293,
|
| 14649 |
+
"grad_norm": 1.38727867603302,
|
| 14650 |
+
"learning_rate": 9.036003746244503e-06,
|
| 14651 |
+
"loss": 0.8403165817260743,
|
| 14652 |
+
"step": 20460
|
| 14653 |
+
},
|
| 14654 |
+
{
|
| 14655 |
+
"epoch": 2.1631768762302337,
|
| 14656 |
+
"grad_norm": 1.9889625310897827,
|
| 14657 |
+
"learning_rate": 9.01470786879844e-06,
|
| 14658 |
+
"loss": 0.893074893951416,
|
| 14659 |
+
"step": 20470
|
| 14660 |
+
},
|
| 14661 |
+
{
|
| 14662 |
+
"epoch": 2.1642337212835385,
|
| 14663 |
+
"grad_norm": 1.331406593322754,
|
| 14664 |
+
"learning_rate": 8.993431594789494e-06,
|
| 14665 |
+
"loss": 0.8329508781433106,
|
| 14666 |
+
"step": 20480
|
| 14667 |
+
},
|
| 14668 |
+
{
|
| 14669 |
+
"epoch": 2.165290566336843,
|
| 14670 |
+
"grad_norm": 1.3607966899871826,
|
| 14671 |
+
"learning_rate": 8.972174950309651e-06,
|
| 14672 |
+
"loss": 0.8310437202453613,
|
| 14673 |
+
"step": 20490
|
| 14674 |
+
},
|
| 14675 |
+
{
|
| 14676 |
+
"epoch": 2.1663474113901477,
|
| 14677 |
+
"grad_norm": 1.157660722732544,
|
| 14678 |
+
"learning_rate": 8.950937961426819e-06,
|
| 14679 |
+
"loss": 0.9204286575317383,
|
| 14680 |
+
"step": 20500
|
| 14681 |
+
},
|
| 14682 |
+
{
|
| 14683 |
+
"epoch": 2.1663474113901477,
|
| 14684 |
+
"eval_loss": 0.4862632155418396,
|
| 14685 |
+
"eval_runtime": 3305.9378,
|
| 14686 |
+
"eval_samples_per_second": 2.694,
|
| 14687 |
+
"eval_steps_per_second": 2.694,
|
| 14688 |
+
"step": 20500
|
| 14689 |
}
|
| 14690 |
],
|
| 14691 |
"logging_steps": 10,
|
|
|
|
| 14705 |
"attributes": {}
|
| 14706 |
}
|
| 14707 |
},
|
| 14708 |
+
"total_flos": 1.1067942153787392e+18,
|
| 14709 |
"train_batch_size": 1,
|
| 14710 |
"trial_name": null,
|
| 14711 |
"trial_params": null
|