Training in progress, step 900, 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 671149168
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1c47d79ddcde62aa6fbf20930b648d737a9449a97c4acec79234f3cfa542bf3d
|
3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 341314644
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9f3ce5310800259d79bf505b8026b04c299b4d6bf18f03aa21d7ac106f5a030e
|
3 |
size 341314644
|
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 sha256:22eb5252c4f5e97c7bb8fb6473bf2e47fb44e7385129e0d532d3cbf45f087cce
|
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 sha256:8a1d54ec4929657be14dbe7bdd312b878c26facb4723f06fcb7f80a15ec48efe
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"best_metric": 0.
|
3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
4 |
-
"epoch": 0.
|
5 |
"eval_steps": 150,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -5305,6 +5305,1064 @@
|
|
5305 |
"eval_samples_per_second": 21.265,
|
5306 |
"eval_steps_per_second": 2.659,
|
5307 |
"step": 750
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5308 |
}
|
5309 |
],
|
5310 |
"logging_steps": 1,
|
@@ -5333,7 +6391,7 @@
|
|
5333 |
"attributes": {}
|
5334 |
}
|
5335 |
},
|
5336 |
-
"total_flos":
|
5337 |
"train_batch_size": 12,
|
5338 |
"trial_name": null,
|
5339 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.9426594972610474,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-900",
|
4 |
+
"epoch": 0.45184503388837755,
|
5 |
"eval_steps": 150,
|
6 |
+
"global_step": 900,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
5305 |
"eval_samples_per_second": 21.265,
|
5306 |
"eval_steps_per_second": 2.659,
|
5307 |
"step": 750
|
5308 |
+
},
|
5309 |
+
{
|
5310 |
+
"epoch": 0.37703957827796836,
|
5311 |
+
"grad_norm": 0.4330700933933258,
|
5312 |
+
"learning_rate": 5.095515587893884e-05,
|
5313 |
+
"loss": 1.2318,
|
5314 |
+
"step": 751
|
5315 |
+
},
|
5316 |
+
{
|
5317 |
+
"epoch": 0.3775416283156221,
|
5318 |
+
"grad_norm": 0.3779419958591461,
|
5319 |
+
"learning_rate": 5.084903828741312e-05,
|
5320 |
+
"loss": 1.2228,
|
5321 |
+
"step": 752
|
5322 |
+
},
|
5323 |
+
{
|
5324 |
+
"epoch": 0.3780436783532759,
|
5325 |
+
"grad_norm": 0.376594603061676,
|
5326 |
+
"learning_rate": 5.0742916870253334e-05,
|
5327 |
+
"loss": 1.1351,
|
5328 |
+
"step": 753
|
5329 |
+
},
|
5330 |
+
{
|
5331 |
+
"epoch": 0.37854572839092965,
|
5332 |
+
"grad_norm": 0.3838042914867401,
|
5333 |
+
"learning_rate": 5.063679210562602e-05,
|
5334 |
+
"loss": 1.1161,
|
5335 |
+
"step": 754
|
5336 |
+
},
|
5337 |
+
{
|
5338 |
+
"epoch": 0.3790477784285834,
|
5339 |
+
"grad_norm": 0.37450775504112244,
|
5340 |
+
"learning_rate": 5.053066447171282e-05,
|
5341 |
+
"loss": 1.0012,
|
5342 |
+
"step": 755
|
5343 |
+
},
|
5344 |
+
{
|
5345 |
+
"epoch": 0.37954982846623714,
|
5346 |
+
"grad_norm": 0.37315741181373596,
|
5347 |
+
"learning_rate": 5.042453444670828e-05,
|
5348 |
+
"loss": 1.1146,
|
5349 |
+
"step": 756
|
5350 |
+
},
|
5351 |
+
{
|
5352 |
+
"epoch": 0.3800518785038909,
|
5353 |
+
"grad_norm": 0.3619626760482788,
|
5354 |
+
"learning_rate": 5.031840250881776e-05,
|
5355 |
+
"loss": 1.0954,
|
5356 |
+
"step": 757
|
5357 |
+
},
|
5358 |
+
{
|
5359 |
+
"epoch": 0.38055392854154463,
|
5360 |
+
"grad_norm": 0.3665991723537445,
|
5361 |
+
"learning_rate": 5.021226913625522e-05,
|
5362 |
+
"loss": 1.0704,
|
5363 |
+
"step": 758
|
5364 |
+
},
|
5365 |
+
{
|
5366 |
+
"epoch": 0.3810559785791984,
|
5367 |
+
"grad_norm": 0.3833234906196594,
|
5368 |
+
"learning_rate": 5.0106134807241045e-05,
|
5369 |
+
"loss": 1.0973,
|
5370 |
+
"step": 759
|
5371 |
+
},
|
5372 |
+
{
|
5373 |
+
"epoch": 0.3815580286168522,
|
5374 |
+
"grad_norm": 0.37826788425445557,
|
5375 |
+
"learning_rate": 5e-05,
|
5376 |
+
"loss": 1.016,
|
5377 |
+
"step": 760
|
5378 |
+
},
|
5379 |
+
{
|
5380 |
+
"epoch": 0.3820600786545059,
|
5381 |
+
"grad_norm": 0.3752281665802002,
|
5382 |
+
"learning_rate": 4.989386519275895e-05,
|
5383 |
+
"loss": 1.0214,
|
5384 |
+
"step": 761
|
5385 |
+
},
|
5386 |
+
{
|
5387 |
+
"epoch": 0.38256212869215966,
|
5388 |
+
"grad_norm": 0.35231512784957886,
|
5389 |
+
"learning_rate": 4.978773086374479e-05,
|
5390 |
+
"loss": 0.9812,
|
5391 |
+
"step": 762
|
5392 |
+
},
|
5393 |
+
{
|
5394 |
+
"epoch": 0.3830641787298134,
|
5395 |
+
"grad_norm": 0.34861356019973755,
|
5396 |
+
"learning_rate": 4.968159749118223e-05,
|
5397 |
+
"loss": 0.9588,
|
5398 |
+
"step": 763
|
5399 |
+
},
|
5400 |
+
{
|
5401 |
+
"epoch": 0.38356622876746715,
|
5402 |
+
"grad_norm": 0.3637848198413849,
|
5403 |
+
"learning_rate": 4.957546555329173e-05,
|
5404 |
+
"loss": 0.9808,
|
5405 |
+
"step": 764
|
5406 |
+
},
|
5407 |
+
{
|
5408 |
+
"epoch": 0.3840682788051209,
|
5409 |
+
"grad_norm": 0.38542938232421875,
|
5410 |
+
"learning_rate": 4.94693355282872e-05,
|
5411 |
+
"loss": 1.0052,
|
5412 |
+
"step": 765
|
5413 |
+
},
|
5414 |
+
{
|
5415 |
+
"epoch": 0.38457032884277464,
|
5416 |
+
"grad_norm": 0.3675108850002289,
|
5417 |
+
"learning_rate": 4.9363207894374e-05,
|
5418 |
+
"loss": 0.9797,
|
5419 |
+
"step": 766
|
5420 |
+
},
|
5421 |
+
{
|
5422 |
+
"epoch": 0.3850723788804284,
|
5423 |
+
"grad_norm": 0.3529476523399353,
|
5424 |
+
"learning_rate": 4.925708312974667e-05,
|
5425 |
+
"loss": 1.0427,
|
5426 |
+
"step": 767
|
5427 |
+
},
|
5428 |
+
{
|
5429 |
+
"epoch": 0.3855744289180822,
|
5430 |
+
"grad_norm": 0.35466766357421875,
|
5431 |
+
"learning_rate": 4.9150961712586895e-05,
|
5432 |
+
"loss": 1.0076,
|
5433 |
+
"step": 768
|
5434 |
+
},
|
5435 |
+
{
|
5436 |
+
"epoch": 0.38607647895573594,
|
5437 |
+
"grad_norm": 0.3574579358100891,
|
5438 |
+
"learning_rate": 4.904484412106117e-05,
|
5439 |
+
"loss": 1.0206,
|
5440 |
+
"step": 769
|
5441 |
+
},
|
5442 |
+
{
|
5443 |
+
"epoch": 0.3865785289933897,
|
5444 |
+
"grad_norm": 0.35434436798095703,
|
5445 |
+
"learning_rate": 4.893873083331882e-05,
|
5446 |
+
"loss": 0.944,
|
5447 |
+
"step": 770
|
5448 |
+
},
|
5449 |
+
{
|
5450 |
+
"epoch": 0.3870805790310434,
|
5451 |
+
"grad_norm": 0.37650713324546814,
|
5452 |
+
"learning_rate": 4.88326223274898e-05,
|
5453 |
+
"loss": 0.9769,
|
5454 |
+
"step": 771
|
5455 |
+
},
|
5456 |
+
{
|
5457 |
+
"epoch": 0.38758262906869717,
|
5458 |
+
"grad_norm": 0.3571126461029053,
|
5459 |
+
"learning_rate": 4.8726519081682444e-05,
|
5460 |
+
"loss": 0.996,
|
5461 |
+
"step": 772
|
5462 |
+
},
|
5463 |
+
{
|
5464 |
+
"epoch": 0.3880846791063509,
|
5465 |
+
"grad_norm": 0.3663455843925476,
|
5466 |
+
"learning_rate": 4.862042157398146e-05,
|
5467 |
+
"loss": 0.908,
|
5468 |
+
"step": 773
|
5469 |
+
},
|
5470 |
+
{
|
5471 |
+
"epoch": 0.38858672914400466,
|
5472 |
+
"grad_norm": 0.380512535572052,
|
5473 |
+
"learning_rate": 4.851433028244562e-05,
|
5474 |
+
"loss": 1.0196,
|
5475 |
+
"step": 774
|
5476 |
+
},
|
5477 |
+
{
|
5478 |
+
"epoch": 0.38908877918165846,
|
5479 |
+
"grad_norm": 0.38776859641075134,
|
5480 |
+
"learning_rate": 4.840824568510579e-05,
|
5481 |
+
"loss": 0.9251,
|
5482 |
+
"step": 775
|
5483 |
+
},
|
5484 |
+
{
|
5485 |
+
"epoch": 0.3895908292193122,
|
5486 |
+
"grad_norm": 0.39721420407295227,
|
5487 |
+
"learning_rate": 4.830216825996257e-05,
|
5488 |
+
"loss": 0.9202,
|
5489 |
+
"step": 776
|
5490 |
+
},
|
5491 |
+
{
|
5492 |
+
"epoch": 0.39009287925696595,
|
5493 |
+
"grad_norm": 0.3933786153793335,
|
5494 |
+
"learning_rate": 4.8196098484984305e-05,
|
5495 |
+
"loss": 0.944,
|
5496 |
+
"step": 777
|
5497 |
+
},
|
5498 |
+
{
|
5499 |
+
"epoch": 0.3905949292946197,
|
5500 |
+
"grad_norm": 0.3744068741798401,
|
5501 |
+
"learning_rate": 4.809003683810486e-05,
|
5502 |
+
"loss": 0.9442,
|
5503 |
+
"step": 778
|
5504 |
+
},
|
5505 |
+
{
|
5506 |
+
"epoch": 0.39109697933227344,
|
5507 |
+
"grad_norm": 0.39798104763031006,
|
5508 |
+
"learning_rate": 4.798398379722147e-05,
|
5509 |
+
"loss": 0.9739,
|
5510 |
+
"step": 779
|
5511 |
+
},
|
5512 |
+
{
|
5513 |
+
"epoch": 0.3915990293699272,
|
5514 |
+
"grad_norm": 0.3898034691810608,
|
5515 |
+
"learning_rate": 4.78779398401926e-05,
|
5516 |
+
"loss": 0.9401,
|
5517 |
+
"step": 780
|
5518 |
+
},
|
5519 |
+
{
|
5520 |
+
"epoch": 0.39210107940758093,
|
5521 |
+
"grad_norm": 0.3922993540763855,
|
5522 |
+
"learning_rate": 4.777190544483574e-05,
|
5523 |
+
"loss": 0.9504,
|
5524 |
+
"step": 781
|
5525 |
+
},
|
5526 |
+
{
|
5527 |
+
"epoch": 0.39260312944523473,
|
5528 |
+
"grad_norm": 0.38821038603782654,
|
5529 |
+
"learning_rate": 4.7665881088925376e-05,
|
5530 |
+
"loss": 0.9617,
|
5531 |
+
"step": 782
|
5532 |
+
},
|
5533 |
+
{
|
5534 |
+
"epoch": 0.3931051794828885,
|
5535 |
+
"grad_norm": 0.3955070674419403,
|
5536 |
+
"learning_rate": 4.75598672501907e-05,
|
5537 |
+
"loss": 0.9072,
|
5538 |
+
"step": 783
|
5539 |
+
},
|
5540 |
+
{
|
5541 |
+
"epoch": 0.3936072295205422,
|
5542 |
+
"grad_norm": 0.38435256481170654,
|
5543 |
+
"learning_rate": 4.7453864406313544e-05,
|
5544 |
+
"loss": 0.9285,
|
5545 |
+
"step": 784
|
5546 |
+
},
|
5547 |
+
{
|
5548 |
+
"epoch": 0.39410927955819597,
|
5549 |
+
"grad_norm": 0.40070778131484985,
|
5550 |
+
"learning_rate": 4.734787303492615e-05,
|
5551 |
+
"loss": 0.9422,
|
5552 |
+
"step": 785
|
5553 |
+
},
|
5554 |
+
{
|
5555 |
+
"epoch": 0.3946113295958497,
|
5556 |
+
"grad_norm": 0.4178116023540497,
|
5557 |
+
"learning_rate": 4.7241893613609126e-05,
|
5558 |
+
"loss": 0.9361,
|
5559 |
+
"step": 786
|
5560 |
+
},
|
5561 |
+
{
|
5562 |
+
"epoch": 0.39511337963350346,
|
5563 |
+
"grad_norm": 0.4187740087509155,
|
5564 |
+
"learning_rate": 4.7135926619889226e-05,
|
5565 |
+
"loss": 0.8883,
|
5566 |
+
"step": 787
|
5567 |
+
},
|
5568 |
+
{
|
5569 |
+
"epoch": 0.3956154296711572,
|
5570 |
+
"grad_norm": 0.42808717489242554,
|
5571 |
+
"learning_rate": 4.702997253123716e-05,
|
5572 |
+
"loss": 0.8763,
|
5573 |
+
"step": 788
|
5574 |
+
},
|
5575 |
+
{
|
5576 |
+
"epoch": 0.396117479708811,
|
5577 |
+
"grad_norm": 0.4418085813522339,
|
5578 |
+
"learning_rate": 4.6924031825065566e-05,
|
5579 |
+
"loss": 0.9475,
|
5580 |
+
"step": 789
|
5581 |
+
},
|
5582 |
+
{
|
5583 |
+
"epoch": 0.39661952974646475,
|
5584 |
+
"grad_norm": 0.4347171485424042,
|
5585 |
+
"learning_rate": 4.6818104978726685e-05,
|
5586 |
+
"loss": 0.7853,
|
5587 |
+
"step": 790
|
5588 |
+
},
|
5589 |
+
{
|
5590 |
+
"epoch": 0.3971215797841185,
|
5591 |
+
"grad_norm": 0.4366185665130615,
|
5592 |
+
"learning_rate": 4.6712192469510425e-05,
|
5593 |
+
"loss": 0.8485,
|
5594 |
+
"step": 791
|
5595 |
+
},
|
5596 |
+
{
|
5597 |
+
"epoch": 0.39762362982177224,
|
5598 |
+
"grad_norm": 0.4427374601364136,
|
5599 |
+
"learning_rate": 4.6606294774641966e-05,
|
5600 |
+
"loss": 0.8737,
|
5601 |
+
"step": 792
|
5602 |
+
},
|
5603 |
+
{
|
5604 |
+
"epoch": 0.398125679859426,
|
5605 |
+
"grad_norm": 0.4442150890827179,
|
5606 |
+
"learning_rate": 4.6500412371279836e-05,
|
5607 |
+
"loss": 0.8032,
|
5608 |
+
"step": 793
|
5609 |
+
},
|
5610 |
+
{
|
5611 |
+
"epoch": 0.39862772989707973,
|
5612 |
+
"grad_norm": 0.4936541020870209,
|
5613 |
+
"learning_rate": 4.6394545736513634e-05,
|
5614 |
+
"loss": 0.8794,
|
5615 |
+
"step": 794
|
5616 |
+
},
|
5617 |
+
{
|
5618 |
+
"epoch": 0.3991297799347335,
|
5619 |
+
"grad_norm": 0.47061917185783386,
|
5620 |
+
"learning_rate": 4.628869534736187e-05,
|
5621 |
+
"loss": 0.8568,
|
5622 |
+
"step": 795
|
5623 |
+
},
|
5624 |
+
{
|
5625 |
+
"epoch": 0.3996318299723872,
|
5626 |
+
"grad_norm": 0.525748610496521,
|
5627 |
+
"learning_rate": 4.618286168076993e-05,
|
5628 |
+
"loss": 0.8513,
|
5629 |
+
"step": 796
|
5630 |
+
},
|
5631 |
+
{
|
5632 |
+
"epoch": 0.400133880010041,
|
5633 |
+
"grad_norm": 0.4828825891017914,
|
5634 |
+
"learning_rate": 4.607704521360776e-05,
|
5635 |
+
"loss": 0.8328,
|
5636 |
+
"step": 797
|
5637 |
+
},
|
5638 |
+
{
|
5639 |
+
"epoch": 0.40063593004769477,
|
5640 |
+
"grad_norm": 0.4649796187877655,
|
5641 |
+
"learning_rate": 4.597124642266788e-05,
|
5642 |
+
"loss": 0.7556,
|
5643 |
+
"step": 798
|
5644 |
+
},
|
5645 |
+
{
|
5646 |
+
"epoch": 0.4011379800853485,
|
5647 |
+
"grad_norm": 0.5552456974983215,
|
5648 |
+
"learning_rate": 4.5865465784663114e-05,
|
5649 |
+
"loss": 0.8184,
|
5650 |
+
"step": 799
|
5651 |
+
},
|
5652 |
+
{
|
5653 |
+
"epoch": 0.40164003012300226,
|
5654 |
+
"grad_norm": 0.706791341304779,
|
5655 |
+
"learning_rate": 4.575970377622456e-05,
|
5656 |
+
"loss": 0.7444,
|
5657 |
+
"step": 800
|
5658 |
+
},
|
5659 |
+
{
|
5660 |
+
"epoch": 0.402142080160656,
|
5661 |
+
"grad_norm": 0.4323110282421112,
|
5662 |
+
"learning_rate": 4.565396087389927e-05,
|
5663 |
+
"loss": 1.1972,
|
5664 |
+
"step": 801
|
5665 |
+
},
|
5666 |
+
{
|
5667 |
+
"epoch": 0.40264413019830975,
|
5668 |
+
"grad_norm": 0.354783833026886,
|
5669 |
+
"learning_rate": 4.554823755414829e-05,
|
5670 |
+
"loss": 1.1179,
|
5671 |
+
"step": 802
|
5672 |
+
},
|
5673 |
+
{
|
5674 |
+
"epoch": 0.4031461802359635,
|
5675 |
+
"grad_norm": 0.3601534068584442,
|
5676 |
+
"learning_rate": 4.544253429334444e-05,
|
5677 |
+
"loss": 1.1264,
|
5678 |
+
"step": 803
|
5679 |
+
},
|
5680 |
+
{
|
5681 |
+
"epoch": 0.4036482302736173,
|
5682 |
+
"grad_norm": 0.3654196858406067,
|
5683 |
+
"learning_rate": 4.5336851567770076e-05,
|
5684 |
+
"loss": 1.0834,
|
5685 |
+
"step": 804
|
5686 |
+
},
|
5687 |
+
{
|
5688 |
+
"epoch": 0.40415028031127104,
|
5689 |
+
"grad_norm": 0.3873622715473175,
|
5690 |
+
"learning_rate": 4.52311898536151e-05,
|
5691 |
+
"loss": 1.0247,
|
5692 |
+
"step": 805
|
5693 |
+
},
|
5694 |
+
{
|
5695 |
+
"epoch": 0.4046523303489248,
|
5696 |
+
"grad_norm": 0.37240368127822876,
|
5697 |
+
"learning_rate": 4.5125549626974696e-05,
|
5698 |
+
"loss": 1.0396,
|
5699 |
+
"step": 806
|
5700 |
+
},
|
5701 |
+
{
|
5702 |
+
"epoch": 0.4051543803865785,
|
5703 |
+
"grad_norm": 0.36485597491264343,
|
5704 |
+
"learning_rate": 4.5019931363847275e-05,
|
5705 |
+
"loss": 1.0249,
|
5706 |
+
"step": 807
|
5707 |
+
},
|
5708 |
+
{
|
5709 |
+
"epoch": 0.4056564304242323,
|
5710 |
+
"grad_norm": 0.38187476992607117,
|
5711 |
+
"learning_rate": 4.491433554013221e-05,
|
5712 |
+
"loss": 1.0405,
|
5713 |
+
"step": 808
|
5714 |
+
},
|
5715 |
+
{
|
5716 |
+
"epoch": 0.406158480461886,
|
5717 |
+
"grad_norm": 0.36962300539016724,
|
5718 |
+
"learning_rate": 4.480876263162783e-05,
|
5719 |
+
"loss": 1.0253,
|
5720 |
+
"step": 809
|
5721 |
+
},
|
5722 |
+
{
|
5723 |
+
"epoch": 0.40666053049953976,
|
5724 |
+
"grad_norm": 0.34921392798423767,
|
5725 |
+
"learning_rate": 4.47032131140292e-05,
|
5726 |
+
"loss": 1.016,
|
5727 |
+
"step": 810
|
5728 |
+
},
|
5729 |
+
{
|
5730 |
+
"epoch": 0.40716258053719356,
|
5731 |
+
"grad_norm": 0.3537079691886902,
|
5732 |
+
"learning_rate": 4.459768746292597e-05,
|
5733 |
+
"loss": 1.0478,
|
5734 |
+
"step": 811
|
5735 |
+
},
|
5736 |
+
{
|
5737 |
+
"epoch": 0.4076646305748473,
|
5738 |
+
"grad_norm": 0.3565637767314911,
|
5739 |
+
"learning_rate": 4.449218615380029e-05,
|
5740 |
+
"loss": 1.0148,
|
5741 |
+
"step": 812
|
5742 |
+
},
|
5743 |
+
{
|
5744 |
+
"epoch": 0.40816668061250105,
|
5745 |
+
"grad_norm": 0.35647860169410706,
|
5746 |
+
"learning_rate": 4.4386709662024544e-05,
|
5747 |
+
"loss": 0.9924,
|
5748 |
+
"step": 813
|
5749 |
+
},
|
5750 |
+
{
|
5751 |
+
"epoch": 0.4086687306501548,
|
5752 |
+
"grad_norm": 0.34907302260398865,
|
5753 |
+
"learning_rate": 4.4281258462859396e-05,
|
5754 |
+
"loss": 1.0018,
|
5755 |
+
"step": 814
|
5756 |
+
},
|
5757 |
+
{
|
5758 |
+
"epoch": 0.40917078068780854,
|
5759 |
+
"grad_norm": 0.3495464026927948,
|
5760 |
+
"learning_rate": 4.4175833031451473e-05,
|
5761 |
+
"loss": 0.9449,
|
5762 |
+
"step": 815
|
5763 |
+
},
|
5764 |
+
{
|
5765 |
+
"epoch": 0.4096728307254623,
|
5766 |
+
"grad_norm": 0.3409779369831085,
|
5767 |
+
"learning_rate": 4.407043384283136e-05,
|
5768 |
+
"loss": 0.9676,
|
5769 |
+
"step": 816
|
5770 |
+
},
|
5771 |
+
{
|
5772 |
+
"epoch": 0.41017488076311603,
|
5773 |
+
"grad_norm": 0.3575940430164337,
|
5774 |
+
"learning_rate": 4.396506137191131e-05,
|
5775 |
+
"loss": 0.9863,
|
5776 |
+
"step": 817
|
5777 |
+
},
|
5778 |
+
{
|
5779 |
+
"epoch": 0.41067693080076983,
|
5780 |
+
"grad_norm": 0.36198464035987854,
|
5781 |
+
"learning_rate": 4.3859716093483245e-05,
|
5782 |
+
"loss": 0.9905,
|
5783 |
+
"step": 818
|
5784 |
+
},
|
5785 |
+
{
|
5786 |
+
"epoch": 0.4111789808384236,
|
5787 |
+
"grad_norm": 0.34198319911956787,
|
5788 |
+
"learning_rate": 4.3754398482216606e-05,
|
5789 |
+
"loss": 0.9482,
|
5790 |
+
"step": 819
|
5791 |
+
},
|
5792 |
+
{
|
5793 |
+
"epoch": 0.4116810308760773,
|
5794 |
+
"grad_norm": 0.3572383224964142,
|
5795 |
+
"learning_rate": 4.364910901265606e-05,
|
5796 |
+
"loss": 0.934,
|
5797 |
+
"step": 820
|
5798 |
+
},
|
5799 |
+
{
|
5800 |
+
"epoch": 0.41218308091373107,
|
5801 |
+
"grad_norm": 0.3588048219680786,
|
5802 |
+
"learning_rate": 4.354384815921958e-05,
|
5803 |
+
"loss": 0.9856,
|
5804 |
+
"step": 821
|
5805 |
+
},
|
5806 |
+
{
|
5807 |
+
"epoch": 0.4126851309513848,
|
5808 |
+
"grad_norm": 0.3628753125667572,
|
5809 |
+
"learning_rate": 4.343861639619611e-05,
|
5810 |
+
"loss": 0.9762,
|
5811 |
+
"step": 822
|
5812 |
+
},
|
5813 |
+
{
|
5814 |
+
"epoch": 0.41318718098903856,
|
5815 |
+
"grad_norm": 0.3723025321960449,
|
5816 |
+
"learning_rate": 4.3333414197743595e-05,
|
5817 |
+
"loss": 0.9704,
|
5818 |
+
"step": 823
|
5819 |
+
},
|
5820 |
+
{
|
5821 |
+
"epoch": 0.4136892310266923,
|
5822 |
+
"grad_norm": 0.3608042597770691,
|
5823 |
+
"learning_rate": 4.322824203788669e-05,
|
5824 |
+
"loss": 0.951,
|
5825 |
+
"step": 824
|
5826 |
+
},
|
5827 |
+
{
|
5828 |
+
"epoch": 0.4141912810643461,
|
5829 |
+
"grad_norm": 0.3752797245979309,
|
5830 |
+
"learning_rate": 4.3123100390514756e-05,
|
5831 |
+
"loss": 0.9878,
|
5832 |
+
"step": 825
|
5833 |
+
},
|
5834 |
+
{
|
5835 |
+
"epoch": 0.41469333110199985,
|
5836 |
+
"grad_norm": 0.37421780824661255,
|
5837 |
+
"learning_rate": 4.3017989729379675e-05,
|
5838 |
+
"loss": 0.9776,
|
5839 |
+
"step": 826
|
5840 |
+
},
|
5841 |
+
{
|
5842 |
+
"epoch": 0.4151953811396536,
|
5843 |
+
"grad_norm": 0.3613242506980896,
|
5844 |
+
"learning_rate": 4.291291052809366e-05,
|
5845 |
+
"loss": 0.9205,
|
5846 |
+
"step": 827
|
5847 |
+
},
|
5848 |
+
{
|
5849 |
+
"epoch": 0.41569743117730734,
|
5850 |
+
"grad_norm": 0.3855215609073639,
|
5851 |
+
"learning_rate": 4.280786326012723e-05,
|
5852 |
+
"loss": 0.986,
|
5853 |
+
"step": 828
|
5854 |
+
},
|
5855 |
+
{
|
5856 |
+
"epoch": 0.4161994812149611,
|
5857 |
+
"grad_norm": 0.41651931405067444,
|
5858 |
+
"learning_rate": 4.2702848398806956e-05,
|
5859 |
+
"loss": 0.9639,
|
5860 |
+
"step": 829
|
5861 |
+
},
|
5862 |
+
{
|
5863 |
+
"epoch": 0.41670153125261483,
|
5864 |
+
"grad_norm": 0.3905417323112488,
|
5865 |
+
"learning_rate": 4.2597866417313436e-05,
|
5866 |
+
"loss": 0.9319,
|
5867 |
+
"step": 830
|
5868 |
+
},
|
5869 |
+
{
|
5870 |
+
"epoch": 0.4172035812902686,
|
5871 |
+
"grad_norm": 0.4226928651332855,
|
5872 |
+
"learning_rate": 4.249291778867909e-05,
|
5873 |
+
"loss": 0.9213,
|
5874 |
+
"step": 831
|
5875 |
+
},
|
5876 |
+
{
|
5877 |
+
"epoch": 0.4177056313279223,
|
5878 |
+
"grad_norm": 0.382017582654953,
|
5879 |
+
"learning_rate": 4.23880029857861e-05,
|
5880 |
+
"loss": 0.8846,
|
5881 |
+
"step": 832
|
5882 |
+
},
|
5883 |
+
{
|
5884 |
+
"epoch": 0.4182076813655761,
|
5885 |
+
"grad_norm": 0.417928546667099,
|
5886 |
+
"learning_rate": 4.2283122481364144e-05,
|
5887 |
+
"loss": 0.9288,
|
5888 |
+
"step": 833
|
5889 |
+
},
|
5890 |
+
{
|
5891 |
+
"epoch": 0.41870973140322987,
|
5892 |
+
"grad_norm": 0.41737717390060425,
|
5893 |
+
"learning_rate": 4.2178276747988446e-05,
|
5894 |
+
"loss": 0.9423,
|
5895 |
+
"step": 834
|
5896 |
+
},
|
5897 |
+
{
|
5898 |
+
"epoch": 0.4192117814408836,
|
5899 |
+
"grad_norm": 0.39423155784606934,
|
5900 |
+
"learning_rate": 4.207346625807756e-05,
|
5901 |
+
"loss": 0.8894,
|
5902 |
+
"step": 835
|
5903 |
+
},
|
5904 |
+
{
|
5905 |
+
"epoch": 0.41971383147853736,
|
5906 |
+
"grad_norm": 0.427852064371109,
|
5907 |
+
"learning_rate": 4.196869148389114e-05,
|
5908 |
+
"loss": 0.9639,
|
5909 |
+
"step": 836
|
5910 |
+
},
|
5911 |
+
{
|
5912 |
+
"epoch": 0.4202158815161911,
|
5913 |
+
"grad_norm": 0.4028894007205963,
|
5914 |
+
"learning_rate": 4.1863952897528e-05,
|
5915 |
+
"loss": 0.9309,
|
5916 |
+
"step": 837
|
5917 |
+
},
|
5918 |
+
{
|
5919 |
+
"epoch": 0.42071793155384485,
|
5920 |
+
"grad_norm": 0.42165279388427734,
|
5921 |
+
"learning_rate": 4.175925097092388e-05,
|
5922 |
+
"loss": 0.9514,
|
5923 |
+
"step": 838
|
5924 |
+
},
|
5925 |
+
{
|
5926 |
+
"epoch": 0.4212199815914986,
|
5927 |
+
"grad_norm": 0.4179115295410156,
|
5928 |
+
"learning_rate": 4.165458617584933e-05,
|
5929 |
+
"loss": 0.8544,
|
5930 |
+
"step": 839
|
5931 |
+
},
|
5932 |
+
{
|
5933 |
+
"epoch": 0.4217220316291524,
|
5934 |
+
"grad_norm": 0.479951947927475,
|
5935 |
+
"learning_rate": 4.1549958983907555e-05,
|
5936 |
+
"loss": 0.811,
|
5937 |
+
"step": 840
|
5938 |
+
},
|
5939 |
+
{
|
5940 |
+
"epoch": 0.42222408166680614,
|
5941 |
+
"grad_norm": 0.45290902256965637,
|
5942 |
+
"learning_rate": 4.144536986653239e-05,
|
5943 |
+
"loss": 0.8243,
|
5944 |
+
"step": 841
|
5945 |
+
},
|
5946 |
+
{
|
5947 |
+
"epoch": 0.4227261317044599,
|
5948 |
+
"grad_norm": 0.4473222494125366,
|
5949 |
+
"learning_rate": 4.1340819294986076e-05,
|
5950 |
+
"loss": 0.8137,
|
5951 |
+
"step": 842
|
5952 |
+
},
|
5953 |
+
{
|
5954 |
+
"epoch": 0.42322818174211363,
|
5955 |
+
"grad_norm": 0.42771241068840027,
|
5956 |
+
"learning_rate": 4.1236307740357173e-05,
|
5957 |
+
"loss": 0.8189,
|
5958 |
+
"step": 843
|
5959 |
+
},
|
5960 |
+
{
|
5961 |
+
"epoch": 0.4237302317797674,
|
5962 |
+
"grad_norm": 0.45651838183403015,
|
5963 |
+
"learning_rate": 4.113183567355846e-05,
|
5964 |
+
"loss": 0.8224,
|
5965 |
+
"step": 844
|
5966 |
+
},
|
5967 |
+
{
|
5968 |
+
"epoch": 0.4242322818174211,
|
5969 |
+
"grad_norm": 0.4706350266933441,
|
5970 |
+
"learning_rate": 4.102740356532473e-05,
|
5971 |
+
"loss": 0.8297,
|
5972 |
+
"step": 845
|
5973 |
+
},
|
5974 |
+
{
|
5975 |
+
"epoch": 0.42473433185507486,
|
5976 |
+
"grad_norm": 0.4705800712108612,
|
5977 |
+
"learning_rate": 4.092301188621084e-05,
|
5978 |
+
"loss": 0.7732,
|
5979 |
+
"step": 846
|
5980 |
+
},
|
5981 |
+
{
|
5982 |
+
"epoch": 0.42523638189272867,
|
5983 |
+
"grad_norm": 0.5137692093849182,
|
5984 |
+
"learning_rate": 4.081866110658934e-05,
|
5985 |
+
"loss": 0.8374,
|
5986 |
+
"step": 847
|
5987 |
+
},
|
5988 |
+
{
|
5989 |
+
"epoch": 0.4257384319303824,
|
5990 |
+
"grad_norm": 0.5054532885551453,
|
5991 |
+
"learning_rate": 4.0714351696648614e-05,
|
5992 |
+
"loss": 0.8556,
|
5993 |
+
"step": 848
|
5994 |
+
},
|
5995 |
+
{
|
5996 |
+
"epoch": 0.42624048196803616,
|
5997 |
+
"grad_norm": 0.5825408697128296,
|
5998 |
+
"learning_rate": 4.061008412639055e-05,
|
5999 |
+
"loss": 0.8321,
|
6000 |
+
"step": 849
|
6001 |
+
},
|
6002 |
+
{
|
6003 |
+
"epoch": 0.4267425320056899,
|
6004 |
+
"grad_norm": 0.6395136117935181,
|
6005 |
+
"learning_rate": 4.050585886562858e-05,
|
6006 |
+
"loss": 0.721,
|
6007 |
+
"step": 850
|
6008 |
+
},
|
6009 |
+
{
|
6010 |
+
"epoch": 0.42724458204334365,
|
6011 |
+
"grad_norm": 0.5878275632858276,
|
6012 |
+
"learning_rate": 4.0401676383985484e-05,
|
6013 |
+
"loss": 1.3045,
|
6014 |
+
"step": 851
|
6015 |
+
},
|
6016 |
+
{
|
6017 |
+
"epoch": 0.4277466320809974,
|
6018 |
+
"grad_norm": 0.3765466511249542,
|
6019 |
+
"learning_rate": 4.0297537150891235e-05,
|
6020 |
+
"loss": 1.1244,
|
6021 |
+
"step": 852
|
6022 |
+
},
|
6023 |
+
{
|
6024 |
+
"epoch": 0.42824868211865114,
|
6025 |
+
"grad_norm": 0.38248923420906067,
|
6026 |
+
"learning_rate": 4.0193441635581e-05,
|
6027 |
+
"loss": 1.1962,
|
6028 |
+
"step": 853
|
6029 |
+
},
|
6030 |
+
{
|
6031 |
+
"epoch": 0.42875073215630494,
|
6032 |
+
"grad_norm": 0.3714083433151245,
|
6033 |
+
"learning_rate": 4.008939030709291e-05,
|
6034 |
+
"loss": 1.026,
|
6035 |
+
"step": 854
|
6036 |
+
},
|
6037 |
+
{
|
6038 |
+
"epoch": 0.4292527821939587,
|
6039 |
+
"grad_norm": 0.3839676082134247,
|
6040 |
+
"learning_rate": 3.998538363426605e-05,
|
6041 |
+
"loss": 1.101,
|
6042 |
+
"step": 855
|
6043 |
+
},
|
6044 |
+
{
|
6045 |
+
"epoch": 0.4297548322316124,
|
6046 |
+
"grad_norm": 0.3552037477493286,
|
6047 |
+
"learning_rate": 3.988142208573822e-05,
|
6048 |
+
"loss": 1.0671,
|
6049 |
+
"step": 856
|
6050 |
+
},
|
6051 |
+
{
|
6052 |
+
"epoch": 0.43025688226926617,
|
6053 |
+
"grad_norm": 0.36277374625205994,
|
6054 |
+
"learning_rate": 3.977750612994396e-05,
|
6055 |
+
"loss": 1.115,
|
6056 |
+
"step": 857
|
6057 |
+
},
|
6058 |
+
{
|
6059 |
+
"epoch": 0.4307589323069199,
|
6060 |
+
"grad_norm": 0.3462297022342682,
|
6061 |
+
"learning_rate": 3.9673636235112376e-05,
|
6062 |
+
"loss": 1.0309,
|
6063 |
+
"step": 858
|
6064 |
+
},
|
6065 |
+
{
|
6066 |
+
"epoch": 0.43126098234457366,
|
6067 |
+
"grad_norm": 0.3610150218009949,
|
6068 |
+
"learning_rate": 3.956981286926498e-05,
|
6069 |
+
"loss": 1.0359,
|
6070 |
+
"step": 859
|
6071 |
+
},
|
6072 |
+
{
|
6073 |
+
"epoch": 0.4317630323822274,
|
6074 |
+
"grad_norm": 0.35921838879585266,
|
6075 |
+
"learning_rate": 3.94660365002137e-05,
|
6076 |
+
"loss": 1.0397,
|
6077 |
+
"step": 860
|
6078 |
+
},
|
6079 |
+
{
|
6080 |
+
"epoch": 0.4322650824198812,
|
6081 |
+
"grad_norm": 0.3716135621070862,
|
6082 |
+
"learning_rate": 3.93623075955586e-05,
|
6083 |
+
"loss": 1.0673,
|
6084 |
+
"step": 861
|
6085 |
+
},
|
6086 |
+
{
|
6087 |
+
"epoch": 0.43276713245753495,
|
6088 |
+
"grad_norm": 0.37005794048309326,
|
6089 |
+
"learning_rate": 3.925862662268602e-05,
|
6090 |
+
"loss": 1.0354,
|
6091 |
+
"step": 862
|
6092 |
+
},
|
6093 |
+
{
|
6094 |
+
"epoch": 0.4332691824951887,
|
6095 |
+
"grad_norm": 0.34723684191703796,
|
6096 |
+
"learning_rate": 3.9154994048766184e-05,
|
6097 |
+
"loss": 1.0334,
|
6098 |
+
"step": 863
|
6099 |
+
},
|
6100 |
+
{
|
6101 |
+
"epoch": 0.43377123253284244,
|
6102 |
+
"grad_norm": 0.3506997525691986,
|
6103 |
+
"learning_rate": 3.905141034075135e-05,
|
6104 |
+
"loss": 0.9656,
|
6105 |
+
"step": 864
|
6106 |
+
},
|
6107 |
+
{
|
6108 |
+
"epoch": 0.4342732825704962,
|
6109 |
+
"grad_norm": 0.37688568234443665,
|
6110 |
+
"learning_rate": 3.894787596537352e-05,
|
6111 |
+
"loss": 0.9302,
|
6112 |
+
"step": 865
|
6113 |
+
},
|
6114 |
+
{
|
6115 |
+
"epoch": 0.43477533260814993,
|
6116 |
+
"grad_norm": 0.3472607433795929,
|
6117 |
+
"learning_rate": 3.884439138914243e-05,
|
6118 |
+
"loss": 0.9686,
|
6119 |
+
"step": 866
|
6120 |
+
},
|
6121 |
+
{
|
6122 |
+
"epoch": 0.4352773826458037,
|
6123 |
+
"grad_norm": 0.35843560099601746,
|
6124 |
+
"learning_rate": 3.874095707834349e-05,
|
6125 |
+
"loss": 0.9701,
|
6126 |
+
"step": 867
|
6127 |
+
},
|
6128 |
+
{
|
6129 |
+
"epoch": 0.4357794326834574,
|
6130 |
+
"grad_norm": 0.3564199209213257,
|
6131 |
+
"learning_rate": 3.863757349903551e-05,
|
6132 |
+
"loss": 1.0456,
|
6133 |
+
"step": 868
|
6134 |
+
},
|
6135 |
+
{
|
6136 |
+
"epoch": 0.4362814827211112,
|
6137 |
+
"grad_norm": 0.38524752855300903,
|
6138 |
+
"learning_rate": 3.853424111704879e-05,
|
6139 |
+
"loss": 0.9603,
|
6140 |
+
"step": 869
|
6141 |
+
},
|
6142 |
+
{
|
6143 |
+
"epoch": 0.43678353275876497,
|
6144 |
+
"grad_norm": 0.3552170693874359,
|
6145 |
+
"learning_rate": 3.843096039798293e-05,
|
6146 |
+
"loss": 0.9274,
|
6147 |
+
"step": 870
|
6148 |
+
},
|
6149 |
+
{
|
6150 |
+
"epoch": 0.4372855827964187,
|
6151 |
+
"grad_norm": 0.37275344133377075,
|
6152 |
+
"learning_rate": 3.832773180720475e-05,
|
6153 |
+
"loss": 1.0213,
|
6154 |
+
"step": 871
|
6155 |
+
},
|
6156 |
+
{
|
6157 |
+
"epoch": 0.43778763283407246,
|
6158 |
+
"grad_norm": 0.3630153238773346,
|
6159 |
+
"learning_rate": 3.822455580984613e-05,
|
6160 |
+
"loss": 0.9482,
|
6161 |
+
"step": 872
|
6162 |
+
},
|
6163 |
+
{
|
6164 |
+
"epoch": 0.4382896828717262,
|
6165 |
+
"grad_norm": 0.36190661787986755,
|
6166 |
+
"learning_rate": 3.8121432870802045e-05,
|
6167 |
+
"loss": 0.881,
|
6168 |
+
"step": 873
|
6169 |
+
},
|
6170 |
+
{
|
6171 |
+
"epoch": 0.43879173290937995,
|
6172 |
+
"grad_norm": 0.3701936602592468,
|
6173 |
+
"learning_rate": 3.801836345472841e-05,
|
6174 |
+
"loss": 1.0065,
|
6175 |
+
"step": 874
|
6176 |
+
},
|
6177 |
+
{
|
6178 |
+
"epoch": 0.4392937829470337,
|
6179 |
+
"grad_norm": 0.4397743344306946,
|
6180 |
+
"learning_rate": 3.791534802603988e-05,
|
6181 |
+
"loss": 0.9938,
|
6182 |
+
"step": 875
|
6183 |
+
},
|
6184 |
+
{
|
6185 |
+
"epoch": 0.4397958329846875,
|
6186 |
+
"grad_norm": 0.36815145611763,
|
6187 |
+
"learning_rate": 3.781238704890793e-05,
|
6188 |
+
"loss": 0.9628,
|
6189 |
+
"step": 876
|
6190 |
+
},
|
6191 |
+
{
|
6192 |
+
"epoch": 0.44029788302234124,
|
6193 |
+
"grad_norm": 0.3762166500091553,
|
6194 |
+
"learning_rate": 3.7709480987258636e-05,
|
6195 |
+
"loss": 0.9478,
|
6196 |
+
"step": 877
|
6197 |
+
},
|
6198 |
+
{
|
6199 |
+
"epoch": 0.440799933059995,
|
6200 |
+
"grad_norm": 0.39231258630752563,
|
6201 |
+
"learning_rate": 3.760663030477072e-05,
|
6202 |
+
"loss": 1.0166,
|
6203 |
+
"step": 878
|
6204 |
+
},
|
6205 |
+
{
|
6206 |
+
"epoch": 0.44130198309764873,
|
6207 |
+
"grad_norm": 0.38583433628082275,
|
6208 |
+
"learning_rate": 3.750383546487324e-05,
|
6209 |
+
"loss": 0.9232,
|
6210 |
+
"step": 879
|
6211 |
+
},
|
6212 |
+
{
|
6213 |
+
"epoch": 0.4418040331353025,
|
6214 |
+
"grad_norm": 0.3934246301651001,
|
6215 |
+
"learning_rate": 3.740109693074375e-05,
|
6216 |
+
"loss": 0.9657,
|
6217 |
+
"step": 880
|
6218 |
+
},
|
6219 |
+
{
|
6220 |
+
"epoch": 0.4423060831729562,
|
6221 |
+
"grad_norm": 0.4055297374725342,
|
6222 |
+
"learning_rate": 3.729841516530604e-05,
|
6223 |
+
"loss": 0.9054,
|
6224 |
+
"step": 881
|
6225 |
+
},
|
6226 |
+
{
|
6227 |
+
"epoch": 0.44280813321060997,
|
6228 |
+
"grad_norm": 0.4082297682762146,
|
6229 |
+
"learning_rate": 3.7195790631228136e-05,
|
6230 |
+
"loss": 0.9365,
|
6231 |
+
"step": 882
|
6232 |
+
},
|
6233 |
+
{
|
6234 |
+
"epoch": 0.44331018324826377,
|
6235 |
+
"grad_norm": 0.39798596501350403,
|
6236 |
+
"learning_rate": 3.709322379092019e-05,
|
6237 |
+
"loss": 0.9023,
|
6238 |
+
"step": 883
|
6239 |
+
},
|
6240 |
+
{
|
6241 |
+
"epoch": 0.4438122332859175,
|
6242 |
+
"grad_norm": 0.418045312166214,
|
6243 |
+
"learning_rate": 3.6990715106532356e-05,
|
6244 |
+
"loss": 0.9233,
|
6245 |
+
"step": 884
|
6246 |
+
},
|
6247 |
+
{
|
6248 |
+
"epoch": 0.44431428332357126,
|
6249 |
+
"grad_norm": 0.4316072463989258,
|
6250 |
+
"learning_rate": 3.68882650399528e-05,
|
6251 |
+
"loss": 0.8931,
|
6252 |
+
"step": 885
|
6253 |
+
},
|
6254 |
+
{
|
6255 |
+
"epoch": 0.444816333361225,
|
6256 |
+
"grad_norm": 0.42850467562675476,
|
6257 |
+
"learning_rate": 3.6785874052805516e-05,
|
6258 |
+
"loss": 0.8839,
|
6259 |
+
"step": 886
|
6260 |
+
},
|
6261 |
+
{
|
6262 |
+
"epoch": 0.44531838339887875,
|
6263 |
+
"grad_norm": 0.4238118529319763,
|
6264 |
+
"learning_rate": 3.6683542606448347e-05,
|
6265 |
+
"loss": 0.9291,
|
6266 |
+
"step": 887
|
6267 |
+
},
|
6268 |
+
{
|
6269 |
+
"epoch": 0.4458204334365325,
|
6270 |
+
"grad_norm": 0.415999174118042,
|
6271 |
+
"learning_rate": 3.658127116197079e-05,
|
6272 |
+
"loss": 0.9257,
|
6273 |
+
"step": 888
|
6274 |
+
},
|
6275 |
+
{
|
6276 |
+
"epoch": 0.44632248347418624,
|
6277 |
+
"grad_norm": 0.4444602131843567,
|
6278 |
+
"learning_rate": 3.6479060180192034e-05,
|
6279 |
+
"loss": 0.8785,
|
6280 |
+
"step": 889
|
6281 |
+
},
|
6282 |
+
{
|
6283 |
+
"epoch": 0.44682453351184004,
|
6284 |
+
"grad_norm": 0.4339217245578766,
|
6285 |
+
"learning_rate": 3.637691012165886e-05,
|
6286 |
+
"loss": 0.7952,
|
6287 |
+
"step": 890
|
6288 |
+
},
|
6289 |
+
{
|
6290 |
+
"epoch": 0.4473265835494938,
|
6291 |
+
"grad_norm": 0.4458482563495636,
|
6292 |
+
"learning_rate": 3.627482144664344e-05,
|
6293 |
+
"loss": 0.8247,
|
6294 |
+
"step": 891
|
6295 |
+
},
|
6296 |
+
{
|
6297 |
+
"epoch": 0.44782863358714753,
|
6298 |
+
"grad_norm": 0.4593295454978943,
|
6299 |
+
"learning_rate": 3.6172794615141446e-05,
|
6300 |
+
"loss": 0.8401,
|
6301 |
+
"step": 892
|
6302 |
+
},
|
6303 |
+
{
|
6304 |
+
"epoch": 0.4483306836248013,
|
6305 |
+
"grad_norm": 0.47604867815971375,
|
6306 |
+
"learning_rate": 3.607083008686985e-05,
|
6307 |
+
"loss": 0.8271,
|
6308 |
+
"step": 893
|
6309 |
+
},
|
6310 |
+
{
|
6311 |
+
"epoch": 0.448832733662455,
|
6312 |
+
"grad_norm": 0.45923951268196106,
|
6313 |
+
"learning_rate": 3.596892832126494e-05,
|
6314 |
+
"loss": 0.858,
|
6315 |
+
"step": 894
|
6316 |
+
},
|
6317 |
+
{
|
6318 |
+
"epoch": 0.44933478370010876,
|
6319 |
+
"grad_norm": 0.4550018608570099,
|
6320 |
+
"learning_rate": 3.586708977748012e-05,
|
6321 |
+
"loss": 0.7788,
|
6322 |
+
"step": 895
|
6323 |
+
},
|
6324 |
+
{
|
6325 |
+
"epoch": 0.4498368337377625,
|
6326 |
+
"grad_norm": 0.4726627469062805,
|
6327 |
+
"learning_rate": 3.5765314914384026e-05,
|
6328 |
+
"loss": 0.8576,
|
6329 |
+
"step": 896
|
6330 |
+
},
|
6331 |
+
{
|
6332 |
+
"epoch": 0.4503388837754163,
|
6333 |
+
"grad_norm": 0.4911380708217621,
|
6334 |
+
"learning_rate": 3.5663604190558296e-05,
|
6335 |
+
"loss": 0.8507,
|
6336 |
+
"step": 897
|
6337 |
+
},
|
6338 |
+
{
|
6339 |
+
"epoch": 0.45084093381307005,
|
6340 |
+
"grad_norm": 0.5006689429283142,
|
6341 |
+
"learning_rate": 3.556195806429559e-05,
|
6342 |
+
"loss": 0.7908,
|
6343 |
+
"step": 898
|
6344 |
+
},
|
6345 |
+
{
|
6346 |
+
"epoch": 0.4513429838507238,
|
6347 |
+
"grad_norm": 0.6167373061180115,
|
6348 |
+
"learning_rate": 3.546037699359751e-05,
|
6349 |
+
"loss": 0.7922,
|
6350 |
+
"step": 899
|
6351 |
+
},
|
6352 |
+
{
|
6353 |
+
"epoch": 0.45184503388837755,
|
6354 |
+
"grad_norm": 0.6547103524208069,
|
6355 |
+
"learning_rate": 3.5358861436172485e-05,
|
6356 |
+
"loss": 0.6946,
|
6357 |
+
"step": 900
|
6358 |
+
},
|
6359 |
+
{
|
6360 |
+
"epoch": 0.45184503388837755,
|
6361 |
+
"eval_loss": 0.9426594972610474,
|
6362 |
+
"eval_runtime": 710.4868,
|
6363 |
+
"eval_samples_per_second": 21.249,
|
6364 |
+
"eval_steps_per_second": 2.657,
|
6365 |
+
"step": 900
|
6366 |
}
|
6367 |
],
|
6368 |
"logging_steps": 1,
|
|
|
6391 |
"attributes": {}
|
6392 |
}
|
6393 |
},
|
6394 |
+
"total_flos": 6.249771756744081e+18,
|
6395 |
"train_batch_size": 12,
|
6396 |
"trial_name": null,
|
6397 |
"trial_params": null
|