Training in progress, step 1956, 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 2503003904
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e0a562a914d2be2b4ee279fe187629b69bd04971e3acab4f1c60939e5ec5996
|
| 3 |
size 2503003904
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5006244836
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9234c66d2cf43b48b78afca3f04cf7c13b9d3436c6cfdb169398c3ecd80cfe02
|
| 3 |
size 5006244836
|
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 sha256:48c25ffa179744c0719c7b65566206a3ffbc025b1b73bf62d6945f9035c21dfa
|
| 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 sha256:98f465f8ef34d3200760108c9ddb9bd27e97b140bce5b4d84a91de037dadb420
|
| 3 |
size 1064
|
last-checkpoint/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,
|
|
@@ -13391,6 +13391,314 @@
|
|
| 13391 |
"learning_rate": 1.3860803461989146e-07,
|
| 13392 |
"loss": 0.8676,
|
| 13393 |
"step": 1912
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13394 |
}
|
| 13395 |
],
|
| 13396 |
"logging_steps": 1,
|
|
@@ -13405,12 +13713,12 @@
|
|
| 13405 |
"should_evaluate": false,
|
| 13406 |
"should_log": false,
|
| 13407 |
"should_save": true,
|
| 13408 |
-
"should_training_stop":
|
| 13409 |
},
|
| 13410 |
"attributes": {}
|
| 13411 |
}
|
| 13412 |
},
|
| 13413 |
-
"total_flos": 7.
|
| 13414 |
"train_batch_size": 8,
|
| 13415 |
"trial_name": null,
|
| 13416 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1956,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 13391 |
"learning_rate": 1.3860803461989146e-07,
|
| 13392 |
"loss": 0.8676,
|
| 13393 |
"step": 1912
|
| 13394 |
+
},
|
| 13395 |
+
{
|
| 13396 |
+
"epoch": 0.9780163599182005,
|
| 13397 |
+
"grad_norm": 2.835423469543457,
|
| 13398 |
+
"learning_rate": 1.3238201455040844e-07,
|
| 13399 |
+
"loss": 0.8435,
|
| 13400 |
+
"step": 1913
|
| 13401 |
+
},
|
| 13402 |
+
{
|
| 13403 |
+
"epoch": 0.9785276073619632,
|
| 13404 |
+
"grad_norm": 3.2276833057403564,
|
| 13405 |
+
"learning_rate": 1.2629887148061792e-07,
|
| 13406 |
+
"loss": 0.9043,
|
| 13407 |
+
"step": 1914
|
| 13408 |
+
},
|
| 13409 |
+
{
|
| 13410 |
+
"epoch": 0.9790388548057259,
|
| 13411 |
+
"grad_norm": 3.260972261428833,
|
| 13412 |
+
"learning_rate": 1.203586228395004e-07,
|
| 13413 |
+
"loss": 0.9502,
|
| 13414 |
+
"step": 1915
|
| 13415 |
+
},
|
| 13416 |
+
{
|
| 13417 |
+
"epoch": 0.9795501022494888,
|
| 13418 |
+
"grad_norm": 3.3064229488372803,
|
| 13419 |
+
"learning_rate": 1.1456128564660273e-07,
|
| 13420 |
+
"loss": 0.9968,
|
| 13421 |
+
"step": 1916
|
| 13422 |
+
},
|
| 13423 |
+
{
|
| 13424 |
+
"epoch": 0.9800613496932515,
|
| 13425 |
+
"grad_norm": 3.274178981781006,
|
| 13426 |
+
"learning_rate": 1.0890687651203823e-07,
|
| 13427 |
+
"loss": 0.8302,
|
| 13428 |
+
"step": 1917
|
| 13429 |
+
},
|
| 13430 |
+
{
|
| 13431 |
+
"epoch": 0.9805725971370143,
|
| 13432 |
+
"grad_norm": 3.076536178588867,
|
| 13433 |
+
"learning_rate": 1.0339541163639776e-07,
|
| 13434 |
+
"loss": 0.9421,
|
| 13435 |
+
"step": 1918
|
| 13436 |
+
},
|
| 13437 |
+
{
|
| 13438 |
+
"epoch": 0.9810838445807771,
|
| 13439 |
+
"grad_norm": 3.247903823852539,
|
| 13440 |
+
"learning_rate": 9.802690681071647e-08,
|
| 13441 |
+
"loss": 0.9819,
|
| 13442 |
+
"step": 1919
|
| 13443 |
+
},
|
| 13444 |
+
{
|
| 13445 |
+
"epoch": 0.9815950920245399,
|
| 13446 |
+
"grad_norm": 3.3538260459899902,
|
| 13447 |
+
"learning_rate": 9.280137741643491e-08,
|
| 13448 |
+
"loss": 0.8744,
|
| 13449 |
+
"step": 1920
|
| 13450 |
+
},
|
| 13451 |
+
{
|
| 13452 |
+
"epoch": 0.9821063394683026,
|
| 13453 |
+
"grad_norm": 3.515782356262207,
|
| 13454 |
+
"learning_rate": 8.771883842536021e-08,
|
| 13455 |
+
"loss": 0.9124,
|
| 13456 |
+
"step": 1921
|
| 13457 |
+
},
|
| 13458 |
+
{
|
| 13459 |
+
"epoch": 0.9826175869120655,
|
| 13460 |
+
"grad_norm": 3.6226806640625,
|
| 13461 |
+
"learning_rate": 8.277930439959946e-08,
|
| 13462 |
+
"loss": 0.9011,
|
| 13463 |
+
"step": 1922
|
| 13464 |
+
},
|
| 13465 |
+
{
|
| 13466 |
+
"epoch": 0.9831288343558282,
|
| 13467 |
+
"grad_norm": 3.3394203186035156,
|
| 13468 |
+
"learning_rate": 7.798278949154303e-08,
|
| 13469 |
+
"loss": 0.8316,
|
| 13470 |
+
"step": 1923
|
| 13471 |
+
},
|
| 13472 |
+
{
|
| 13473 |
+
"epoch": 0.983640081799591,
|
| 13474 |
+
"grad_norm": 3.246371030807495,
|
| 13475 |
+
"learning_rate": 7.332930744380906e-08,
|
| 13476 |
+
"loss": 0.8556,
|
| 13477 |
+
"step": 1924
|
| 13478 |
+
},
|
| 13479 |
+
{
|
| 13480 |
+
"epoch": 0.9841513292433538,
|
| 13481 |
+
"grad_norm": 3.402927875518799,
|
| 13482 |
+
"learning_rate": 6.881887158920464e-08,
|
| 13483 |
+
"loss": 0.7978,
|
| 13484 |
+
"step": 1925
|
| 13485 |
+
},
|
| 13486 |
+
{
|
| 13487 |
+
"epoch": 0.9846625766871165,
|
| 13488 |
+
"grad_norm": 3.8112809658050537,
|
| 13489 |
+
"learning_rate": 6.445149485070357e-08,
|
| 13490 |
+
"loss": 0.9133,
|
| 13491 |
+
"step": 1926
|
| 13492 |
+
},
|
| 13493 |
+
{
|
| 13494 |
+
"epoch": 0.9851738241308794,
|
| 13495 |
+
"grad_norm": 3.5460119247436523,
|
| 13496 |
+
"learning_rate": 6.022718974137975e-08,
|
| 13497 |
+
"loss": 0.8158,
|
| 13498 |
+
"step": 1927
|
| 13499 |
+
},
|
| 13500 |
+
{
|
| 13501 |
+
"epoch": 0.9856850715746421,
|
| 13502 |
+
"grad_norm": 3.341395854949951,
|
| 13503 |
+
"learning_rate": 5.614596836440722e-08,
|
| 13504 |
+
"loss": 0.8246,
|
| 13505 |
+
"step": 1928
|
| 13506 |
+
},
|
| 13507 |
+
{
|
| 13508 |
+
"epoch": 0.9861963190184049,
|
| 13509 |
+
"grad_norm": 3.6873090267181396,
|
| 13510 |
+
"learning_rate": 5.2207842412999034e-08,
|
| 13511 |
+
"loss": 0.8714,
|
| 13512 |
+
"step": 1929
|
| 13513 |
+
},
|
| 13514 |
+
{
|
| 13515 |
+
"epoch": 0.9867075664621677,
|
| 13516 |
+
"grad_norm": 3.4815688133239746,
|
| 13517 |
+
"learning_rate": 4.841282317037399e-08,
|
| 13518 |
+
"loss": 0.8948,
|
| 13519 |
+
"step": 1930
|
| 13520 |
+
},
|
| 13521 |
+
{
|
| 13522 |
+
"epoch": 0.9872188139059305,
|
| 13523 |
+
"grad_norm": 3.5316038131713867,
|
| 13524 |
+
"learning_rate": 4.476092150975109e-08,
|
| 13525 |
+
"loss": 0.8622,
|
| 13526 |
+
"step": 1931
|
| 13527 |
+
},
|
| 13528 |
+
{
|
| 13529 |
+
"epoch": 0.9877300613496932,
|
| 13530 |
+
"grad_norm": 3.5975794792175293,
|
| 13531 |
+
"learning_rate": 4.1252147894277336e-08,
|
| 13532 |
+
"loss": 0.881,
|
| 13533 |
+
"step": 1932
|
| 13534 |
+
},
|
| 13535 |
+
{
|
| 13536 |
+
"epoch": 0.9882413087934561,
|
| 13537 |
+
"grad_norm": 3.441171646118164,
|
| 13538 |
+
"learning_rate": 3.7886512377033334e-08,
|
| 13539 |
+
"loss": 0.8396,
|
| 13540 |
+
"step": 1933
|
| 13541 |
+
},
|
| 13542 |
+
{
|
| 13543 |
+
"epoch": 0.9887525562372188,
|
| 13544 |
+
"grad_norm": 3.8511383533477783,
|
| 13545 |
+
"learning_rate": 3.4664024600988835e-08,
|
| 13546 |
+
"loss": 0.9208,
|
| 13547 |
+
"step": 1934
|
| 13548 |
+
},
|
| 13549 |
+
{
|
| 13550 |
+
"epoch": 0.9892638036809815,
|
| 13551 |
+
"grad_norm": 3.8687822818756104,
|
| 13552 |
+
"learning_rate": 3.158469379898055e-08,
|
| 13553 |
+
"loss": 0.9135,
|
| 13554 |
+
"step": 1935
|
| 13555 |
+
},
|
| 13556 |
+
{
|
| 13557 |
+
"epoch": 0.9897750511247444,
|
| 13558 |
+
"grad_norm": 3.593276023864746,
|
| 13559 |
+
"learning_rate": 2.8648528793673302e-08,
|
| 13560 |
+
"loss": 0.8474,
|
| 13561 |
+
"step": 1936
|
| 13562 |
+
},
|
| 13563 |
+
{
|
| 13564 |
+
"epoch": 0.9902862985685071,
|
| 13565 |
+
"grad_norm": 4.0986127853393555,
|
| 13566 |
+
"learning_rate": 2.5855537997548917e-08,
|
| 13567 |
+
"loss": 0.8883,
|
| 13568 |
+
"step": 1937
|
| 13569 |
+
},
|
| 13570 |
+
{
|
| 13571 |
+
"epoch": 0.99079754601227,
|
| 13572 |
+
"grad_norm": 4.03285551071167,
|
| 13573 |
+
"learning_rate": 2.3205729412884016e-08,
|
| 13574 |
+
"loss": 0.7779,
|
| 13575 |
+
"step": 1938
|
| 13576 |
+
},
|
| 13577 |
+
{
|
| 13578 |
+
"epoch": 0.9913087934560327,
|
| 13579 |
+
"grad_norm": 4.346153736114502,
|
| 13580 |
+
"learning_rate": 2.0699110631711148e-08,
|
| 13581 |
+
"loss": 0.8757,
|
| 13582 |
+
"step": 1939
|
| 13583 |
+
},
|
| 13584 |
+
{
|
| 13585 |
+
"epoch": 0.9918200408997955,
|
| 13586 |
+
"grad_norm": 4.283609390258789,
|
| 13587 |
+
"learning_rate": 1.8335688835802167e-08,
|
| 13588 |
+
"loss": 0.8173,
|
| 13589 |
+
"step": 1940
|
| 13590 |
+
},
|
| 13591 |
+
{
|
| 13592 |
+
"epoch": 0.9923312883435583,
|
| 13593 |
+
"grad_norm": 4.301876068115234,
|
| 13594 |
+
"learning_rate": 1.6115470796662647e-08,
|
| 13595 |
+
"loss": 0.9134,
|
| 13596 |
+
"step": 1941
|
| 13597 |
+
},
|
| 13598 |
+
{
|
| 13599 |
+
"epoch": 0.9928425357873211,
|
| 13600 |
+
"grad_norm": 4.885223865509033,
|
| 13601 |
+
"learning_rate": 1.4038462875504143e-08,
|
| 13602 |
+
"loss": 0.8289,
|
| 13603 |
+
"step": 1942
|
| 13604 |
+
},
|
| 13605 |
+
{
|
| 13606 |
+
"epoch": 0.9933537832310838,
|
| 13607 |
+
"grad_norm": 4.63042688369751,
|
| 13608 |
+
"learning_rate": 1.2104671023199787e-08,
|
| 13609 |
+
"loss": 0.8625,
|
| 13610 |
+
"step": 1943
|
| 13611 |
+
},
|
| 13612 |
+
{
|
| 13613 |
+
"epoch": 0.9938650306748467,
|
| 13614 |
+
"grad_norm": 4.702084064483643,
|
| 13615 |
+
"learning_rate": 1.0314100780317581e-08,
|
| 13616 |
+
"loss": 0.9342,
|
| 13617 |
+
"step": 1944
|
| 13618 |
+
},
|
| 13619 |
+
{
|
| 13620 |
+
"epoch": 0.9943762781186094,
|
| 13621 |
+
"grad_norm": 4.4585771560668945,
|
| 13622 |
+
"learning_rate": 8.666757277064897e-09,
|
| 13623 |
+
"loss": 0.6828,
|
| 13624 |
+
"step": 1945
|
| 13625 |
+
},
|
| 13626 |
+
{
|
| 13627 |
+
"epoch": 0.9948875255623721,
|
| 13628 |
+
"grad_norm": 4.869369029998779,
|
| 13629 |
+
"learning_rate": 7.162645233282916e-09,
|
| 13630 |
+
"loss": 0.8505,
|
| 13631 |
+
"step": 1946
|
| 13632 |
+
},
|
| 13633 |
+
{
|
| 13634 |
+
"epoch": 0.995398773006135,
|
| 13635 |
+
"grad_norm": 4.623004913330078,
|
| 13636 |
+
"learning_rate": 5.8017689584521915e-09,
|
| 13637 |
+
"loss": 0.6772,
|
| 13638 |
+
"step": 1947
|
| 13639 |
+
},
|
| 13640 |
+
{
|
| 13641 |
+
"epoch": 0.9959100204498977,
|
| 13642 |
+
"grad_norm": 5.718740940093994,
|
| 13643 |
+
"learning_rate": 4.584132351642678e-09,
|
| 13644 |
+
"loss": 0.8251,
|
| 13645 |
+
"step": 1948
|
| 13646 |
+
},
|
| 13647 |
+
{
|
| 13648 |
+
"epoch": 0.9964212678936605,
|
| 13649 |
+
"grad_norm": 5.196649551391602,
|
| 13650 |
+
"learning_rate": 3.509738901547044e-09,
|
| 13651 |
+
"loss": 0.6039,
|
| 13652 |
+
"step": 1949
|
| 13653 |
+
},
|
| 13654 |
+
{
|
| 13655 |
+
"epoch": 0.9969325153374233,
|
| 13656 |
+
"grad_norm": 6.253082752227783,
|
| 13657 |
+
"learning_rate": 2.5785916864307092e-09,
|
| 13658 |
+
"loss": 0.5829,
|
| 13659 |
+
"step": 1950
|
| 13660 |
+
},
|
| 13661 |
+
{
|
| 13662 |
+
"epoch": 0.9974437627811861,
|
| 13663 |
+
"grad_norm": 2.393533229827881,
|
| 13664 |
+
"learning_rate": 1.7906933741484999e-09,
|
| 13665 |
+
"loss": 0.869,
|
| 13666 |
+
"step": 1951
|
| 13667 |
+
},
|
| 13668 |
+
{
|
| 13669 |
+
"epoch": 0.9979550102249489,
|
| 13670 |
+
"grad_norm": 2.994673728942871,
|
| 13671 |
+
"learning_rate": 1.1460462221279944e-09,
|
| 13672 |
+
"loss": 0.9103,
|
| 13673 |
+
"step": 1952
|
| 13674 |
+
},
|
| 13675 |
+
{
|
| 13676 |
+
"epoch": 0.9984662576687117,
|
| 13677 |
+
"grad_norm": 3.5846893787384033,
|
| 13678 |
+
"learning_rate": 6.446520773695231e-10,
|
| 13679 |
+
"loss": 0.8366,
|
| 13680 |
+
"step": 1953
|
| 13681 |
+
},
|
| 13682 |
+
{
|
| 13683 |
+
"epoch": 0.9989775051124744,
|
| 13684 |
+
"grad_norm": 3.5298497676849365,
|
| 13685 |
+
"learning_rate": 2.8651237642396414e-10,
|
| 13686 |
+
"loss": 0.7694,
|
| 13687 |
+
"step": 1954
|
| 13688 |
+
},
|
| 13689 |
+
{
|
| 13690 |
+
"epoch": 0.9994887525562373,
|
| 13691 |
+
"grad_norm": 3.686530828475952,
|
| 13692 |
+
"learning_rate": 7.162814541494811e-11,
|
| 13693 |
+
"loss": 0.67,
|
| 13694 |
+
"step": 1955
|
| 13695 |
+
},
|
| 13696 |
+
{
|
| 13697 |
+
"epoch": 1.0,
|
| 13698 |
+
"grad_norm": 5.693774223327637,
|
| 13699 |
+
"learning_rate": 0.0,
|
| 13700 |
+
"loss": 0.7087,
|
| 13701 |
+
"step": 1956
|
| 13702 |
}
|
| 13703 |
],
|
| 13704 |
"logging_steps": 1,
|
|
|
|
| 13713 |
"should_evaluate": false,
|
| 13714 |
"should_log": false,
|
| 13715 |
"should_save": true,
|
| 13716 |
+
"should_training_stop": true
|
| 13717 |
},
|
| 13718 |
"attributes": {}
|
| 13719 |
}
|
| 13720 |
},
|
| 13721 |
+
"total_flos": 7.478583213010452e+17,
|
| 13722 |
"train_batch_size": 8,
|
| 13723 |
"trial_name": null,
|
| 13724 |
"trial_params": null
|