Training in progress, step 161000, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -23,13 +23,13 @@
|
|
| 23 |
"rank_pattern": {},
|
| 24 |
"revision": null,
|
| 25 |
"target_modules": [
|
|
|
|
|
|
|
| 26 |
"up_proj",
|
|
|
|
| 27 |
"gate_proj",
|
| 28 |
"k_proj",
|
| 29 |
-
"
|
| 30 |
-
"down_proj",
|
| 31 |
-
"o_proj",
|
| 32 |
-
"q_proj"
|
| 33 |
],
|
| 34 |
"task_type": "CAUSAL_LM",
|
| 35 |
"use_dora": false,
|
|
|
|
| 23 |
"rank_pattern": {},
|
| 24 |
"revision": null,
|
| 25 |
"target_modules": [
|
| 26 |
+
"down_proj",
|
| 27 |
+
"v_proj",
|
| 28 |
"up_proj",
|
| 29 |
+
"q_proj",
|
| 30 |
"gate_proj",
|
| 31 |
"k_proj",
|
| 32 |
+
"o_proj"
|
|
|
|
|
|
|
|
|
|
| 33 |
],
|
| 34 |
"task_type": "CAUSAL_LM",
|
| 35 |
"use_dora": false,
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1556140392
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:776b57e8cf86af8f45ec851c7f40dff293a2260f466d76c2635a529669ea7be1
|
| 3 |
size 1556140392
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 791683586
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e35e6b330681427d92536d5b866820fd3210e6f1045fb36af35a966e2802ccfa
|
| 3 |
size 791683586
|
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 sha256:1cd19186e1a1c47e5da80a5032267ec74bc9c4149fe528d2a46b67a9b8d64752
|
| 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,
|
|
@@ -5292,6 +5292,356 @@
|
|
| 5292 |
"learning_rate": 1.9904972598324345e-05,
|
| 5293 |
"loss": 1.8064,
|
| 5294 |
"step": 151000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5295 |
}
|
| 5296 |
],
|
| 5297 |
"logging_steps": 200,
|
|
@@ -5311,7 +5661,7 @@
|
|
| 5311 |
"attributes": {}
|
| 5312 |
}
|
| 5313 |
},
|
| 5314 |
-
"total_flos": 8.
|
| 5315 |
"train_batch_size": 1,
|
| 5316 |
"trial_name": null,
|
| 5317 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.11197265919764983,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 161000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 5292 |
"learning_rate": 1.9904972598324345e-05,
|
| 5293 |
"loss": 1.8064,
|
| 5294 |
"step": 151000
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.10515693211605374,
|
| 5298 |
+
"grad_norm": 0.6552620530128479,
|
| 5299 |
+
"learning_rate": 1.9904721276394122e-05,
|
| 5300 |
+
"loss": 1.7752,
|
| 5301 |
+
"step": 151200
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.10529602858710672,
|
| 5305 |
+
"grad_norm": 0.8357005715370178,
|
| 5306 |
+
"learning_rate": 1.990446962484043e-05,
|
| 5307 |
+
"loss": 1.7828,
|
| 5308 |
+
"step": 151400
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.10543512505815972,
|
| 5312 |
+
"grad_norm": 1.1673591136932373,
|
| 5313 |
+
"learning_rate": 1.9904217643675287e-05,
|
| 5314 |
+
"loss": 1.8316,
|
| 5315 |
+
"step": 151600
|
| 5316 |
+
},
|
| 5317 |
+
{
|
| 5318 |
+
"epoch": 0.1055742215292127,
|
| 5319 |
+
"grad_norm": 3.075096845626831,
|
| 5320 |
+
"learning_rate": 1.9903965332910706e-05,
|
| 5321 |
+
"loss": 1.7649,
|
| 5322 |
+
"step": 151800
|
| 5323 |
+
},
|
| 5324 |
+
{
|
| 5325 |
+
"epoch": 0.10571331800026568,
|
| 5326 |
+
"grad_norm": 1.600201964378357,
|
| 5327 |
+
"learning_rate": 1.990371269255875e-05,
|
| 5328 |
+
"loss": 1.815,
|
| 5329 |
+
"step": 152000
|
| 5330 |
+
},
|
| 5331 |
+
{
|
| 5332 |
+
"epoch": 0.10585241447131866,
|
| 5333 |
+
"grad_norm": 1.1825553178787231,
|
| 5334 |
+
"learning_rate": 1.9903459722631466e-05,
|
| 5335 |
+
"loss": 1.8219,
|
| 5336 |
+
"step": 152200
|
| 5337 |
+
},
|
| 5338 |
+
{
|
| 5339 |
+
"epoch": 0.10599151094237164,
|
| 5340 |
+
"grad_norm": 0.8359824419021606,
|
| 5341 |
+
"learning_rate": 1.9903206423140936e-05,
|
| 5342 |
+
"loss": 1.7919,
|
| 5343 |
+
"step": 152400
|
| 5344 |
+
},
|
| 5345 |
+
{
|
| 5346 |
+
"epoch": 0.10613060741342462,
|
| 5347 |
+
"grad_norm": 1.7081679105758667,
|
| 5348 |
+
"learning_rate": 1.9902952794099257e-05,
|
| 5349 |
+
"loss": 1.7764,
|
| 5350 |
+
"step": 152600
|
| 5351 |
+
},
|
| 5352 |
+
{
|
| 5353 |
+
"epoch": 0.1062697038844776,
|
| 5354 |
+
"grad_norm": 0.9274424314498901,
|
| 5355 |
+
"learning_rate": 1.9902698835518533e-05,
|
| 5356 |
+
"loss": 1.7792,
|
| 5357 |
+
"step": 152800
|
| 5358 |
+
},
|
| 5359 |
+
{
|
| 5360 |
+
"epoch": 0.10640880035553058,
|
| 5361 |
+
"grad_norm": 1.007084608078003,
|
| 5362 |
+
"learning_rate": 1.9902444547410883e-05,
|
| 5363 |
+
"loss": 1.7766,
|
| 5364 |
+
"step": 153000
|
| 5365 |
+
},
|
| 5366 |
+
{
|
| 5367 |
+
"epoch": 0.10654789682658355,
|
| 5368 |
+
"grad_norm": 0.8499106764793396,
|
| 5369 |
+
"learning_rate": 1.9902189929788453e-05,
|
| 5370 |
+
"loss": 1.7173,
|
| 5371 |
+
"step": 153200
|
| 5372 |
+
},
|
| 5373 |
+
{
|
| 5374 |
+
"epoch": 0.10668699329763655,
|
| 5375 |
+
"grad_norm": 0.8728858232498169,
|
| 5376 |
+
"learning_rate": 1.9901934982663393e-05,
|
| 5377 |
+
"loss": 1.7235,
|
| 5378 |
+
"step": 153400
|
| 5379 |
+
},
|
| 5380 |
+
{
|
| 5381 |
+
"epoch": 0.10682608976868953,
|
| 5382 |
+
"grad_norm": 1.5445681810379028,
|
| 5383 |
+
"learning_rate": 1.990167970604788e-05,
|
| 5384 |
+
"loss": 1.8182,
|
| 5385 |
+
"step": 153600
|
| 5386 |
+
},
|
| 5387 |
+
{
|
| 5388 |
+
"epoch": 0.10696518623974251,
|
| 5389 |
+
"grad_norm": 1.0307021141052246,
|
| 5390 |
+
"learning_rate": 1.9901424099954094e-05,
|
| 5391 |
+
"loss": 1.7935,
|
| 5392 |
+
"step": 153800
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 0.10710428271079549,
|
| 5396 |
+
"grad_norm": 0.4203742444515228,
|
| 5397 |
+
"learning_rate": 1.9901168164394242e-05,
|
| 5398 |
+
"loss": 1.7632,
|
| 5399 |
+
"step": 154000
|
| 5400 |
+
},
|
| 5401 |
+
{
|
| 5402 |
+
"epoch": 0.10724337918184847,
|
| 5403 |
+
"grad_norm": 1.2339669466018677,
|
| 5404 |
+
"learning_rate": 1.990091189938054e-05,
|
| 5405 |
+
"loss": 1.7986,
|
| 5406 |
+
"step": 154200
|
| 5407 |
+
},
|
| 5408 |
+
{
|
| 5409 |
+
"epoch": 0.10738247565290145,
|
| 5410 |
+
"grad_norm": 0.9977461099624634,
|
| 5411 |
+
"learning_rate": 1.9900655304925225e-05,
|
| 5412 |
+
"loss": 1.7751,
|
| 5413 |
+
"step": 154400
|
| 5414 |
+
},
|
| 5415 |
+
{
|
| 5416 |
+
"epoch": 0.10752157212395443,
|
| 5417 |
+
"grad_norm": 0.8145340085029602,
|
| 5418 |
+
"learning_rate": 1.990039838104054e-05,
|
| 5419 |
+
"loss": 1.7175,
|
| 5420 |
+
"step": 154600
|
| 5421 |
+
},
|
| 5422 |
+
{
|
| 5423 |
+
"epoch": 0.1076606685950074,
|
| 5424 |
+
"grad_norm": 0.8700305819511414,
|
| 5425 |
+
"learning_rate": 1.9900141127738757e-05,
|
| 5426 |
+
"loss": 1.8136,
|
| 5427 |
+
"step": 154800
|
| 5428 |
+
},
|
| 5429 |
+
{
|
| 5430 |
+
"epoch": 0.10779976506606039,
|
| 5431 |
+
"grad_norm": 1.0328384637832642,
|
| 5432 |
+
"learning_rate": 1.989988354503215e-05,
|
| 5433 |
+
"loss": 1.8334,
|
| 5434 |
+
"step": 155000
|
| 5435 |
+
},
|
| 5436 |
+
{
|
| 5437 |
+
"epoch": 0.10793886153711338,
|
| 5438 |
+
"grad_norm": 1.1774046421051025,
|
| 5439 |
+
"learning_rate": 1.9899625632933027e-05,
|
| 5440 |
+
"loss": 1.7683,
|
| 5441 |
+
"step": 155200
|
| 5442 |
+
},
|
| 5443 |
+
{
|
| 5444 |
+
"epoch": 0.10807795800816636,
|
| 5445 |
+
"grad_norm": 1.0147098302841187,
|
| 5446 |
+
"learning_rate": 1.989936739145369e-05,
|
| 5447 |
+
"loss": 1.8265,
|
| 5448 |
+
"step": 155400
|
| 5449 |
+
},
|
| 5450 |
+
{
|
| 5451 |
+
"epoch": 0.10821705447921934,
|
| 5452 |
+
"grad_norm": 1.3206279277801514,
|
| 5453 |
+
"learning_rate": 1.989910882060647e-05,
|
| 5454 |
+
"loss": 1.7738,
|
| 5455 |
+
"step": 155600
|
| 5456 |
+
},
|
| 5457 |
+
{
|
| 5458 |
+
"epoch": 0.10835615095027232,
|
| 5459 |
+
"grad_norm": 0.8004774451255798,
|
| 5460 |
+
"learning_rate": 1.9898849920403708e-05,
|
| 5461 |
+
"loss": 1.8271,
|
| 5462 |
+
"step": 155800
|
| 5463 |
+
},
|
| 5464 |
+
{
|
| 5465 |
+
"epoch": 0.1084952474213253,
|
| 5466 |
+
"grad_norm": 0.9754022359848022,
|
| 5467 |
+
"learning_rate": 1.9898590690857774e-05,
|
| 5468 |
+
"loss": 1.7882,
|
| 5469 |
+
"step": 156000
|
| 5470 |
+
},
|
| 5471 |
+
{
|
| 5472 |
+
"epoch": 0.10863434389237828,
|
| 5473 |
+
"grad_norm": 1.2540456056594849,
|
| 5474 |
+
"learning_rate": 1.989833113198103e-05,
|
| 5475 |
+
"loss": 1.8104,
|
| 5476 |
+
"step": 156200
|
| 5477 |
+
},
|
| 5478 |
+
{
|
| 5479 |
+
"epoch": 0.10877344036343126,
|
| 5480 |
+
"grad_norm": 0.7259592413902283,
|
| 5481 |
+
"learning_rate": 1.9898071243785876e-05,
|
| 5482 |
+
"loss": 1.7974,
|
| 5483 |
+
"step": 156400
|
| 5484 |
+
},
|
| 5485 |
+
{
|
| 5486 |
+
"epoch": 0.10891253683448424,
|
| 5487 |
+
"grad_norm": 1.9316837787628174,
|
| 5488 |
+
"learning_rate": 1.9897811026284718e-05,
|
| 5489 |
+
"loss": 1.7933,
|
| 5490 |
+
"step": 156600
|
| 5491 |
+
},
|
| 5492 |
+
{
|
| 5493 |
+
"epoch": 0.10905163330553722,
|
| 5494 |
+
"grad_norm": 0.8343173861503601,
|
| 5495 |
+
"learning_rate": 1.989755047948997e-05,
|
| 5496 |
+
"loss": 1.7924,
|
| 5497 |
+
"step": 156800
|
| 5498 |
+
},
|
| 5499 |
+
{
|
| 5500 |
+
"epoch": 0.10919072977659021,
|
| 5501 |
+
"grad_norm": 0.7665181159973145,
|
| 5502 |
+
"learning_rate": 1.989728960341408e-05,
|
| 5503 |
+
"loss": 1.7123,
|
| 5504 |
+
"step": 157000
|
| 5505 |
+
},
|
| 5506 |
+
{
|
| 5507 |
+
"epoch": 0.10932982624764319,
|
| 5508 |
+
"grad_norm": 1.001929759979248,
|
| 5509 |
+
"learning_rate": 1.9897028398069503e-05,
|
| 5510 |
+
"loss": 1.8386,
|
| 5511 |
+
"step": 157200
|
| 5512 |
+
},
|
| 5513 |
+
{
|
| 5514 |
+
"epoch": 0.10946892271869617,
|
| 5515 |
+
"grad_norm": 1.0837798118591309,
|
| 5516 |
+
"learning_rate": 1.98967668634687e-05,
|
| 5517 |
+
"loss": 1.7943,
|
| 5518 |
+
"step": 157400
|
| 5519 |
+
},
|
| 5520 |
+
{
|
| 5521 |
+
"epoch": 0.10960801918974915,
|
| 5522 |
+
"grad_norm": 0.9598066806793213,
|
| 5523 |
+
"learning_rate": 1.9896504999624163e-05,
|
| 5524 |
+
"loss": 1.7959,
|
| 5525 |
+
"step": 157600
|
| 5526 |
+
},
|
| 5527 |
+
{
|
| 5528 |
+
"epoch": 0.10974711566080213,
|
| 5529 |
+
"grad_norm": 0.9790207743644714,
|
| 5530 |
+
"learning_rate": 1.989624280654839e-05,
|
| 5531 |
+
"loss": 1.7813,
|
| 5532 |
+
"step": 157800
|
| 5533 |
+
},
|
| 5534 |
+
{
|
| 5535 |
+
"epoch": 0.10988621213185511,
|
| 5536 |
+
"grad_norm": 0.6536526679992676,
|
| 5537 |
+
"learning_rate": 1.98959802842539e-05,
|
| 5538 |
+
"loss": 1.7947,
|
| 5539 |
+
"step": 158000
|
| 5540 |
+
},
|
| 5541 |
+
{
|
| 5542 |
+
"epoch": 0.11002530860290809,
|
| 5543 |
+
"grad_norm": 1.2658668756484985,
|
| 5544 |
+
"learning_rate": 1.9895717432753222e-05,
|
| 5545 |
+
"loss": 1.7699,
|
| 5546 |
+
"step": 158200
|
| 5547 |
+
},
|
| 5548 |
+
{
|
| 5549 |
+
"epoch": 0.11016440507396107,
|
| 5550 |
+
"grad_norm": 2.004922866821289,
|
| 5551 |
+
"learning_rate": 1.9895454252058903e-05,
|
| 5552 |
+
"loss": 1.7686,
|
| 5553 |
+
"step": 158400
|
| 5554 |
+
},
|
| 5555 |
+
{
|
| 5556 |
+
"epoch": 0.11030350154501405,
|
| 5557 |
+
"grad_norm": 1.3296949863433838,
|
| 5558 |
+
"learning_rate": 1.9895190742183518e-05,
|
| 5559 |
+
"loss": 1.7689,
|
| 5560 |
+
"step": 158600
|
| 5561 |
+
},
|
| 5562 |
+
{
|
| 5563 |
+
"epoch": 0.11044259801606704,
|
| 5564 |
+
"grad_norm": 1.112878680229187,
|
| 5565 |
+
"learning_rate": 1.9894926903139633e-05,
|
| 5566 |
+
"loss": 1.8022,
|
| 5567 |
+
"step": 158800
|
| 5568 |
+
},
|
| 5569 |
+
{
|
| 5570 |
+
"epoch": 0.11058169448712002,
|
| 5571 |
+
"grad_norm": 1.8036686182022095,
|
| 5572 |
+
"learning_rate": 1.9894662734939847e-05,
|
| 5573 |
+
"loss": 1.7359,
|
| 5574 |
+
"step": 159000
|
| 5575 |
+
},
|
| 5576 |
+
{
|
| 5577 |
+
"epoch": 0.110720790958173,
|
| 5578 |
+
"grad_norm": 1.7240679264068604,
|
| 5579 |
+
"learning_rate": 1.989439823759678e-05,
|
| 5580 |
+
"loss": 1.794,
|
| 5581 |
+
"step": 159200
|
| 5582 |
+
},
|
| 5583 |
+
{
|
| 5584 |
+
"epoch": 0.11085988742922598,
|
| 5585 |
+
"grad_norm": 0.8310656547546387,
|
| 5586 |
+
"learning_rate": 1.9894133411123047e-05,
|
| 5587 |
+
"loss": 1.7825,
|
| 5588 |
+
"step": 159400
|
| 5589 |
+
},
|
| 5590 |
+
{
|
| 5591 |
+
"epoch": 0.11099898390027896,
|
| 5592 |
+
"grad_norm": 1.0748811960220337,
|
| 5593 |
+
"learning_rate": 1.9893868255531295e-05,
|
| 5594 |
+
"loss": 1.7347,
|
| 5595 |
+
"step": 159600
|
| 5596 |
+
},
|
| 5597 |
+
{
|
| 5598 |
+
"epoch": 0.11113808037133194,
|
| 5599 |
+
"grad_norm": 0.7499359250068665,
|
| 5600 |
+
"learning_rate": 1.989360277083419e-05,
|
| 5601 |
+
"loss": 1.8134,
|
| 5602 |
+
"step": 159800
|
| 5603 |
+
},
|
| 5604 |
+
{
|
| 5605 |
+
"epoch": 0.11127717684238492,
|
| 5606 |
+
"grad_norm": 0.793464183807373,
|
| 5607 |
+
"learning_rate": 1.9893336957044394e-05,
|
| 5608 |
+
"loss": 1.7876,
|
| 5609 |
+
"step": 160000
|
| 5610 |
+
},
|
| 5611 |
+
{
|
| 5612 |
+
"epoch": 0.1114162733134379,
|
| 5613 |
+
"grad_norm": 1.079917073249817,
|
| 5614 |
+
"learning_rate": 1.9893070814174604e-05,
|
| 5615 |
+
"loss": 1.8132,
|
| 5616 |
+
"step": 160200
|
| 5617 |
+
},
|
| 5618 |
+
{
|
| 5619 |
+
"epoch": 0.11155536978449088,
|
| 5620 |
+
"grad_norm": 1.041785717010498,
|
| 5621 |
+
"learning_rate": 1.9892804342237518e-05,
|
| 5622 |
+
"loss": 1.7536,
|
| 5623 |
+
"step": 160400
|
| 5624 |
+
},
|
| 5625 |
+
{
|
| 5626 |
+
"epoch": 0.11169446625554387,
|
| 5627 |
+
"grad_norm": 0.9491516351699829,
|
| 5628 |
+
"learning_rate": 1.9892537541245865e-05,
|
| 5629 |
+
"loss": 1.7811,
|
| 5630 |
+
"step": 160600
|
| 5631 |
+
},
|
| 5632 |
+
{
|
| 5633 |
+
"epoch": 0.11183356272659685,
|
| 5634 |
+
"grad_norm": 1.1019930839538574,
|
| 5635 |
+
"learning_rate": 1.989227041121238e-05,
|
| 5636 |
+
"loss": 1.7485,
|
| 5637 |
+
"step": 160800
|
| 5638 |
+
},
|
| 5639 |
+
{
|
| 5640 |
+
"epoch": 0.11197265919764983,
|
| 5641 |
+
"grad_norm": 1.0586832761764526,
|
| 5642 |
+
"learning_rate": 1.9892002952149815e-05,
|
| 5643 |
+
"loss": 1.7691,
|
| 5644 |
+
"step": 161000
|
| 5645 |
}
|
| 5646 |
],
|
| 5647 |
"logging_steps": 200,
|
|
|
|
| 5661 |
"attributes": {}
|
| 5662 |
}
|
| 5663 |
},
|
| 5664 |
+
"total_flos": 8.806317924606628e+17,
|
| 5665 |
"train_batch_size": 1,
|
| 5666 |
"trial_name": null,
|
| 5667 |
"trial_params": null
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6776
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d151f8446fd7799d24b0e9a99d447d7d465b181c00e55fa7d90224c183012544
|
| 3 |
size 6776
|