Training in progress, step 678, 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 73911112
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c5e63de907b19c7910db799e3b78a42b2e9f9179854322d92c712e5afde4b89
|
| 3 |
size 73911112
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 37965684
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a14e009833d53fbe9c82d489fe0e9b4ad66c19f8ea8adae05db1259a8aef663d
|
| 3 |
size 37965684
|
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:c6b4a91354537b8cb9fcab809682ea0aa63e50d33b8c1b98a541a3219c1729c7
|
| 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:f12e5684e3c4181284222b5bfd802e16a9d019f6d061509f1ae13f7ad0785a15
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 2.048067808151245,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
-
"epoch": 1.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -4311,6 +4311,560 @@
|
|
| 4311 |
"eval_samples_per_second": 34.057,
|
| 4312 |
"eval_steps_per_second": 8.514,
|
| 4313 |
"step": 600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4314 |
}
|
| 4315 |
],
|
| 4316 |
"logging_steps": 1,
|
|
@@ -4334,12 +4888,12 @@
|
|
| 4334 |
"should_evaluate": false,
|
| 4335 |
"should_log": false,
|
| 4336 |
"should_save": true,
|
| 4337 |
-
"should_training_stop":
|
| 4338 |
},
|
| 4339 |
"attributes": {}
|
| 4340 |
}
|
| 4341 |
},
|
| 4342 |
-
"total_flos": 3.
|
| 4343 |
"train_batch_size": 4,
|
| 4344 |
"trial_name": null,
|
| 4345 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 2.048067808151245,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
+
"epoch": 1.9977900552486187,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 678,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4311 |
"eval_samples_per_second": 34.057,
|
| 4312 |
"eval_steps_per_second": 8.514,
|
| 4313 |
"step": 600
|
| 4314 |
+
},
|
| 4315 |
+
{
|
| 4316 |
+
"epoch": 1.770902394106814,
|
| 4317 |
+
"grad_norm": 0.3590349853038788,
|
| 4318 |
+
"learning_rate": 6.485553382910026e-06,
|
| 4319 |
+
"loss": 1.9617,
|
| 4320 |
+
"step": 601
|
| 4321 |
+
},
|
| 4322 |
+
{
|
| 4323 |
+
"epoch": 1.7738489871086556,
|
| 4324 |
+
"grad_norm": 0.35265564918518066,
|
| 4325 |
+
"learning_rate": 6.319977191187232e-06,
|
| 4326 |
+
"loss": 2.0291,
|
| 4327 |
+
"step": 602
|
| 4328 |
+
},
|
| 4329 |
+
{
|
| 4330 |
+
"epoch": 1.7767955801104973,
|
| 4331 |
+
"grad_norm": 0.34210672974586487,
|
| 4332 |
+
"learning_rate": 6.156473015002029e-06,
|
| 4333 |
+
"loss": 1.9461,
|
| 4334 |
+
"step": 603
|
| 4335 |
+
},
|
| 4336 |
+
{
|
| 4337 |
+
"epoch": 1.7797421731123388,
|
| 4338 |
+
"grad_norm": 0.36943700909614563,
|
| 4339 |
+
"learning_rate": 5.995044470741151e-06,
|
| 4340 |
+
"loss": 1.9437,
|
| 4341 |
+
"step": 604
|
| 4342 |
+
},
|
| 4343 |
+
{
|
| 4344 |
+
"epoch": 1.7826887661141804,
|
| 4345 |
+
"grad_norm": 0.3616660237312317,
|
| 4346 |
+
"learning_rate": 5.835695128882513e-06,
|
| 4347 |
+
"loss": 1.9987,
|
| 4348 |
+
"step": 605
|
| 4349 |
+
},
|
| 4350 |
+
{
|
| 4351 |
+
"epoch": 1.7856353591160221,
|
| 4352 |
+
"grad_norm": 0.35595205426216125,
|
| 4353 |
+
"learning_rate": 5.678428513916212e-06,
|
| 4354 |
+
"loss": 2.0232,
|
| 4355 |
+
"step": 606
|
| 4356 |
+
},
|
| 4357 |
+
{
|
| 4358 |
+
"epoch": 1.7885819521178639,
|
| 4359 |
+
"grad_norm": 0.3314264416694641,
|
| 4360 |
+
"learning_rate": 5.5232481042665764e-06,
|
| 4361 |
+
"loss": 1.7963,
|
| 4362 |
+
"step": 607
|
| 4363 |
+
},
|
| 4364 |
+
{
|
| 4365 |
+
"epoch": 1.7915285451197054,
|
| 4366 |
+
"grad_norm": 0.3526301681995392,
|
| 4367 |
+
"learning_rate": 5.370157332215209e-06,
|
| 4368 |
+
"loss": 2.0898,
|
| 4369 |
+
"step": 608
|
| 4370 |
+
},
|
| 4371 |
+
{
|
| 4372 |
+
"epoch": 1.794475138121547,
|
| 4373 |
+
"grad_norm": 0.3451070785522461,
|
| 4374 |
+
"learning_rate": 5.219159583825106e-06,
|
| 4375 |
+
"loss": 1.8876,
|
| 4376 |
+
"step": 609
|
| 4377 |
+
},
|
| 4378 |
+
{
|
| 4379 |
+
"epoch": 1.7974217311233884,
|
| 4380 |
+
"grad_norm": 0.34398043155670166,
|
| 4381 |
+
"learning_rate": 5.07025819886574e-06,
|
| 4382 |
+
"loss": 1.9447,
|
| 4383 |
+
"step": 610
|
| 4384 |
+
},
|
| 4385 |
+
{
|
| 4386 |
+
"epoch": 1.8003683241252302,
|
| 4387 |
+
"grad_norm": 0.33297380805015564,
|
| 4388 |
+
"learning_rate": 4.923456470739219e-06,
|
| 4389 |
+
"loss": 1.9111,
|
| 4390 |
+
"step": 611
|
| 4391 |
+
},
|
| 4392 |
+
{
|
| 4393 |
+
"epoch": 1.803314917127072,
|
| 4394 |
+
"grad_norm": 0.35569652915000916,
|
| 4395 |
+
"learning_rate": 4.778757646407362e-06,
|
| 4396 |
+
"loss": 2.052,
|
| 4397 |
+
"step": 612
|
| 4398 |
+
},
|
| 4399 |
+
{
|
| 4400 |
+
"epoch": 1.8062615101289135,
|
| 4401 |
+
"grad_norm": 0.3721804618835449,
|
| 4402 |
+
"learning_rate": 4.636164926320064e-06,
|
| 4403 |
+
"loss": 2.1146,
|
| 4404 |
+
"step": 613
|
| 4405 |
+
},
|
| 4406 |
+
{
|
| 4407 |
+
"epoch": 1.809208103130755,
|
| 4408 |
+
"grad_norm": 0.3460250198841095,
|
| 4409 |
+
"learning_rate": 4.495681464344259e-06,
|
| 4410 |
+
"loss": 1.9138,
|
| 4411 |
+
"step": 614
|
| 4412 |
+
},
|
| 4413 |
+
{
|
| 4414 |
+
"epoch": 1.8121546961325967,
|
| 4415 |
+
"grad_norm": 0.38758331537246704,
|
| 4416 |
+
"learning_rate": 4.357310367694378e-06,
|
| 4417 |
+
"loss": 1.975,
|
| 4418 |
+
"step": 615
|
| 4419 |
+
},
|
| 4420 |
+
{
|
| 4421 |
+
"epoch": 1.8151012891344385,
|
| 4422 |
+
"grad_norm": 0.37770211696624756,
|
| 4423 |
+
"learning_rate": 4.22105469686348e-06,
|
| 4424 |
+
"loss": 2.0591,
|
| 4425 |
+
"step": 616
|
| 4426 |
+
},
|
| 4427 |
+
{
|
| 4428 |
+
"epoch": 1.81804788213628,
|
| 4429 |
+
"grad_norm": 0.3766396641731262,
|
| 4430 |
+
"learning_rate": 4.086917465555662e-06,
|
| 4431 |
+
"loss": 2.0694,
|
| 4432 |
+
"step": 617
|
| 4433 |
+
},
|
| 4434 |
+
{
|
| 4435 |
+
"epoch": 1.8209944751381215,
|
| 4436 |
+
"grad_norm": 0.3475191593170166,
|
| 4437 |
+
"learning_rate": 3.954901640619368e-06,
|
| 4438 |
+
"loss": 1.9057,
|
| 4439 |
+
"step": 618
|
| 4440 |
+
},
|
| 4441 |
+
{
|
| 4442 |
+
"epoch": 1.823941068139963,
|
| 4443 |
+
"grad_norm": 0.3547375798225403,
|
| 4444 |
+
"learning_rate": 3.825010141981677e-06,
|
| 4445 |
+
"loss": 2.0484,
|
| 4446 |
+
"step": 619
|
| 4447 |
+
},
|
| 4448 |
+
{
|
| 4449 |
+
"epoch": 1.8268876611418048,
|
| 4450 |
+
"grad_norm": 0.35701024532318115,
|
| 4451 |
+
"learning_rate": 3.6972458425838962e-06,
|
| 4452 |
+
"loss": 2.084,
|
| 4453 |
+
"step": 620
|
| 4454 |
+
},
|
| 4455 |
+
{
|
| 4456 |
+
"epoch": 1.8298342541436465,
|
| 4457 |
+
"grad_norm": 0.3376360535621643,
|
| 4458 |
+
"learning_rate": 3.571611568317856e-06,
|
| 4459 |
+
"loss": 1.8885,
|
| 4460 |
+
"step": 621
|
| 4461 |
+
},
|
| 4462 |
+
{
|
| 4463 |
+
"epoch": 1.832780847145488,
|
| 4464 |
+
"grad_norm": 0.34869226813316345,
|
| 4465 |
+
"learning_rate": 3.4481100979635306e-06,
|
| 4466 |
+
"loss": 2.1957,
|
| 4467 |
+
"step": 622
|
| 4468 |
+
},
|
| 4469 |
+
{
|
| 4470 |
+
"epoch": 1.8357274401473296,
|
| 4471 |
+
"grad_norm": 0.3375036418437958,
|
| 4472 |
+
"learning_rate": 3.32674416312746e-06,
|
| 4473 |
+
"loss": 1.8859,
|
| 4474 |
+
"step": 623
|
| 4475 |
+
},
|
| 4476 |
+
{
|
| 4477 |
+
"epoch": 1.838674033149171,
|
| 4478 |
+
"grad_norm": 0.3811117708683014,
|
| 4479 |
+
"learning_rate": 3.207516448182435e-06,
|
| 4480 |
+
"loss": 2.147,
|
| 4481 |
+
"step": 624
|
| 4482 |
+
},
|
| 4483 |
+
{
|
| 4484 |
+
"epoch": 1.8416206261510129,
|
| 4485 |
+
"grad_norm": 0.3357301652431488,
|
| 4486 |
+
"learning_rate": 3.0904295902080326e-06,
|
| 4487 |
+
"loss": 1.7973,
|
| 4488 |
+
"step": 625
|
| 4489 |
+
},
|
| 4490 |
+
{
|
| 4491 |
+
"epoch": 1.8445672191528546,
|
| 4492 |
+
"grad_norm": 0.38550111651420593,
|
| 4493 |
+
"learning_rate": 2.9754861789324073e-06,
|
| 4494 |
+
"loss": 2.1196,
|
| 4495 |
+
"step": 626
|
| 4496 |
+
},
|
| 4497 |
+
{
|
| 4498 |
+
"epoch": 1.8475138121546961,
|
| 4499 |
+
"grad_norm": 0.3680182099342346,
|
| 4500 |
+
"learning_rate": 2.8626887566748807e-06,
|
| 4501 |
+
"loss": 2.0742,
|
| 4502 |
+
"step": 627
|
| 4503 |
+
},
|
| 4504 |
+
{
|
| 4505 |
+
"epoch": 1.8504604051565376,
|
| 4506 |
+
"grad_norm": 0.3618917763233185,
|
| 4507 |
+
"learning_rate": 2.752039818289809e-06,
|
| 4508 |
+
"loss": 2.0812,
|
| 4509 |
+
"step": 628
|
| 4510 |
+
},
|
| 4511 |
+
{
|
| 4512 |
+
"epoch": 1.8534069981583794,
|
| 4513 |
+
"grad_norm": 0.3539152443408966,
|
| 4514 |
+
"learning_rate": 2.6435418111113276e-06,
|
| 4515 |
+
"loss": 1.9961,
|
| 4516 |
+
"step": 629
|
| 4517 |
+
},
|
| 4518 |
+
{
|
| 4519 |
+
"epoch": 1.8563535911602211,
|
| 4520 |
+
"grad_norm": 0.35489708185195923,
|
| 4521 |
+
"learning_rate": 2.537197134899294e-06,
|
| 4522 |
+
"loss": 2.0572,
|
| 4523 |
+
"step": 630
|
| 4524 |
+
},
|
| 4525 |
+
{
|
| 4526 |
+
"epoch": 1.8593001841620627,
|
| 4527 |
+
"grad_norm": 0.34133121371269226,
|
| 4528 |
+
"learning_rate": 2.433008141786153e-06,
|
| 4529 |
+
"loss": 1.9718,
|
| 4530 |
+
"step": 631
|
| 4531 |
+
},
|
| 4532 |
+
{
|
| 4533 |
+
"epoch": 1.8622467771639042,
|
| 4534 |
+
"grad_norm": 0.34101560711860657,
|
| 4535 |
+
"learning_rate": 2.330977136224932e-06,
|
| 4536 |
+
"loss": 1.881,
|
| 4537 |
+
"step": 632
|
| 4538 |
+
},
|
| 4539 |
+
{
|
| 4540 |
+
"epoch": 1.8651933701657457,
|
| 4541 |
+
"grad_norm": 0.3772835433483124,
|
| 4542 |
+
"learning_rate": 2.2311063749382742e-06,
|
| 4543 |
+
"loss": 2.0331,
|
| 4544 |
+
"step": 633
|
| 4545 |
+
},
|
| 4546 |
+
{
|
| 4547 |
+
"epoch": 1.8681399631675875,
|
| 4548 |
+
"grad_norm": 0.3728832006454468,
|
| 4549 |
+
"learning_rate": 2.1333980668685414e-06,
|
| 4550 |
+
"loss": 1.9764,
|
| 4551 |
+
"step": 634
|
| 4552 |
+
},
|
| 4553 |
+
{
|
| 4554 |
+
"epoch": 1.8710865561694292,
|
| 4555 |
+
"grad_norm": 0.36289656162261963,
|
| 4556 |
+
"learning_rate": 2.037854373128889e-06,
|
| 4557 |
+
"loss": 2.006,
|
| 4558 |
+
"step": 635
|
| 4559 |
+
},
|
| 4560 |
+
{
|
| 4561 |
+
"epoch": 1.8740331491712707,
|
| 4562 |
+
"grad_norm": 0.36972489953041077,
|
| 4563 |
+
"learning_rate": 1.9444774069555694e-06,
|
| 4564 |
+
"loss": 2.1053,
|
| 4565 |
+
"step": 636
|
| 4566 |
+
},
|
| 4567 |
+
{
|
| 4568 |
+
"epoch": 1.8769797421731123,
|
| 4569 |
+
"grad_norm": 0.34683194756507874,
|
| 4570 |
+
"learning_rate": 1.8532692336611035e-06,
|
| 4571 |
+
"loss": 2.0312,
|
| 4572 |
+
"step": 637
|
| 4573 |
+
},
|
| 4574 |
+
{
|
| 4575 |
+
"epoch": 1.879926335174954,
|
| 4576 |
+
"grad_norm": 0.3639572858810425,
|
| 4577 |
+
"learning_rate": 1.7642318705886286e-06,
|
| 4578 |
+
"loss": 1.9183,
|
| 4579 |
+
"step": 638
|
| 4580 |
+
},
|
| 4581 |
+
{
|
| 4582 |
+
"epoch": 1.8828729281767957,
|
| 4583 |
+
"grad_norm": 0.3998357355594635,
|
| 4584 |
+
"learning_rate": 1.6773672870673218e-06,
|
| 4585 |
+
"loss": 2.0961,
|
| 4586 |
+
"step": 639
|
| 4587 |
+
},
|
| 4588 |
+
{
|
| 4589 |
+
"epoch": 1.8858195211786373,
|
| 4590 |
+
"grad_norm": 0.3396947383880615,
|
| 4591 |
+
"learning_rate": 1.5926774043687365e-06,
|
| 4592 |
+
"loss": 1.9543,
|
| 4593 |
+
"step": 640
|
| 4594 |
+
},
|
| 4595 |
+
{
|
| 4596 |
+
"epoch": 1.8887661141804788,
|
| 4597 |
+
"grad_norm": 0.3566475510597229,
|
| 4598 |
+
"learning_rate": 1.510164095664457e-06,
|
| 4599 |
+
"loss": 2.0163,
|
| 4600 |
+
"step": 641
|
| 4601 |
+
},
|
| 4602 |
+
{
|
| 4603 |
+
"epoch": 1.8917127071823203,
|
| 4604 |
+
"grad_norm": 0.33658653497695923,
|
| 4605 |
+
"learning_rate": 1.4298291859845214e-06,
|
| 4606 |
+
"loss": 1.9524,
|
| 4607 |
+
"step": 642
|
| 4608 |
+
},
|
| 4609 |
+
{
|
| 4610 |
+
"epoch": 1.894659300184162,
|
| 4611 |
+
"grad_norm": 0.35298457741737366,
|
| 4612 |
+
"learning_rate": 1.351674452177143e-06,
|
| 4613 |
+
"loss": 2.0429,
|
| 4614 |
+
"step": 643
|
| 4615 |
+
},
|
| 4616 |
+
{
|
| 4617 |
+
"epoch": 1.8976058931860038,
|
| 4618 |
+
"grad_norm": 0.353605180978775,
|
| 4619 |
+
"learning_rate": 1.2757016228693964e-06,
|
| 4620 |
+
"loss": 2.0201,
|
| 4621 |
+
"step": 644
|
| 4622 |
+
},
|
| 4623 |
+
{
|
| 4624 |
+
"epoch": 1.9005524861878453,
|
| 4625 |
+
"grad_norm": 0.3473069667816162,
|
| 4626 |
+
"learning_rate": 1.2019123784289488e-06,
|
| 4627 |
+
"loss": 2.054,
|
| 4628 |
+
"step": 645
|
| 4629 |
+
},
|
| 4630 |
+
{
|
| 4631 |
+
"epoch": 1.9034990791896869,
|
| 4632 |
+
"grad_norm": 0.373976469039917,
|
| 4633 |
+
"learning_rate": 1.1303083509269452e-06,
|
| 4634 |
+
"loss": 1.9479,
|
| 4635 |
+
"step": 646
|
| 4636 |
+
},
|
| 4637 |
+
{
|
| 4638 |
+
"epoch": 1.9064456721915284,
|
| 4639 |
+
"grad_norm": 0.3614707887172699,
|
| 4640 |
+
"learning_rate": 1.0608911241018594e-06,
|
| 4641 |
+
"loss": 1.9979,
|
| 4642 |
+
"step": 647
|
| 4643 |
+
},
|
| 4644 |
+
{
|
| 4645 |
+
"epoch": 1.9093922651933701,
|
| 4646 |
+
"grad_norm": 0.35335227847099304,
|
| 4647 |
+
"learning_rate": 9.936622333245104e-07,
|
| 4648 |
+
"loss": 1.9355,
|
| 4649 |
+
"step": 648
|
| 4650 |
+
},
|
| 4651 |
+
{
|
| 4652 |
+
"epoch": 1.9123388581952119,
|
| 4653 |
+
"grad_norm": 0.3640158474445343,
|
| 4654 |
+
"learning_rate": 9.286231655640799e-07,
|
| 4655 |
+
"loss": 2.0179,
|
| 4656 |
+
"step": 649
|
| 4657 |
+
},
|
| 4658 |
+
{
|
| 4659 |
+
"epoch": 1.9152854511970534,
|
| 4660 |
+
"grad_norm": 0.3631095290184021,
|
| 4661 |
+
"learning_rate": 8.657753593552143e-07,
|
| 4662 |
+
"loss": 1.9934,
|
| 4663 |
+
"step": 650
|
| 4664 |
+
},
|
| 4665 |
+
{
|
| 4666 |
+
"epoch": 1.9152854511970534,
|
| 4667 |
+
"eval_loss": 2.0474448204040527,
|
| 4668 |
+
"eval_runtime": 33.5976,
|
| 4669 |
+
"eval_samples_per_second": 34.05,
|
| 4670 |
+
"eval_steps_per_second": 8.513,
|
| 4671 |
+
"step": 650
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 1.918232044198895,
|
| 4675 |
+
"grad_norm": 0.3535584509372711,
|
| 4676 |
+
"learning_rate": 8.051202047662187e-07,
|
| 4677 |
+
"loss": 1.9061,
|
| 4678 |
+
"step": 651
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 1.9211786372007367,
|
| 4682 |
+
"grad_norm": 0.3503694534301758,
|
| 4683 |
+
"learning_rate": 7.466590433683251e-07,
|
| 4684 |
+
"loss": 1.9775,
|
| 4685 |
+
"step": 652
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 1.9241252302025784,
|
| 4689 |
+
"grad_norm": 0.3380708396434784,
|
| 4690 |
+
"learning_rate": 6.903931682059827e-07,
|
| 4691 |
+
"loss": 1.8295,
|
| 4692 |
+
"step": 653
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 1.92707182320442,
|
| 4696 |
+
"grad_norm": 0.3362259268760681,
|
| 4697 |
+
"learning_rate": 6.363238237683033e-07,
|
| 4698 |
+
"loss": 1.7838,
|
| 4699 |
+
"step": 654
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 1.9300184162062615,
|
| 4703 |
+
"grad_norm": 0.3610450327396393,
|
| 4704 |
+
"learning_rate": 5.844522059614943e-07,
|
| 4705 |
+
"loss": 2.1624,
|
| 4706 |
+
"step": 655
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 1.932965009208103,
|
| 4710 |
+
"grad_norm": 0.34263625741004944,
|
| 4711 |
+
"learning_rate": 5.347794620824576e-07,
|
| 4712 |
+
"loss": 1.9548,
|
| 4713 |
+
"step": 656
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 1.9359116022099447,
|
| 4717 |
+
"grad_norm": 0.35191041231155396,
|
| 4718 |
+
"learning_rate": 4.873066907933543e-07,
|
| 4719 |
+
"loss": 2.0965,
|
| 4720 |
+
"step": 657
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 1.9388581952117865,
|
| 4724 |
+
"grad_norm": 0.361122727394104,
|
| 4725 |
+
"learning_rate": 4.4203494209733576e-07,
|
| 4726 |
+
"loss": 1.9781,
|
| 4727 |
+
"step": 658
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 1.941804788213628,
|
| 4731 |
+
"grad_norm": 0.35918036103248596,
|
| 4732 |
+
"learning_rate": 3.9896521731532797e-07,
|
| 4733 |
+
"loss": 1.8921,
|
| 4734 |
+
"step": 659
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 1.9447513812154695,
|
| 4738 |
+
"grad_norm": 0.37808507680892944,
|
| 4739 |
+
"learning_rate": 3.580984690638611e-07,
|
| 4740 |
+
"loss": 2.2418,
|
| 4741 |
+
"step": 660
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 1.9476979742173113,
|
| 4745 |
+
"grad_norm": 0.34101665019989014,
|
| 4746 |
+
"learning_rate": 3.194356012340305e-07,
|
| 4747 |
+
"loss": 1.8401,
|
| 4748 |
+
"step": 661
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 1.9506445672191528,
|
| 4752 |
+
"grad_norm": 0.360929936170578,
|
| 4753 |
+
"learning_rate": 2.8297746897146816e-07,
|
| 4754 |
+
"loss": 2.0246,
|
| 4755 |
+
"step": 662
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 1.9535911602209945,
|
| 4759 |
+
"grad_norm": 0.36307036876678467,
|
| 4760 |
+
"learning_rate": 2.48724878657447e-07,
|
| 4761 |
+
"loss": 2.0241,
|
| 4762 |
+
"step": 663
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 1.956537753222836,
|
| 4766 |
+
"grad_norm": 0.3703291118144989,
|
| 4767 |
+
"learning_rate": 2.1667858789105043e-07,
|
| 4768 |
+
"loss": 2.1115,
|
| 4769 |
+
"step": 664
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 1.9594843462246776,
|
| 4773 |
+
"grad_norm": 0.3611985445022583,
|
| 4774 |
+
"learning_rate": 1.8683930547243045e-07,
|
| 4775 |
+
"loss": 1.9565,
|
| 4776 |
+
"step": 665
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 1.9624309392265193,
|
| 4780 |
+
"grad_norm": 0.351419061422348,
|
| 4781 |
+
"learning_rate": 1.5920769138706438e-07,
|
| 4782 |
+
"loss": 1.8914,
|
| 4783 |
+
"step": 666
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 1.965377532228361,
|
| 4787 |
+
"grad_norm": 0.36377424001693726,
|
| 4788 |
+
"learning_rate": 1.3378435679122226e-07,
|
| 4789 |
+
"loss": 2.0512,
|
| 4790 |
+
"step": 667
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 1.9683241252302026,
|
| 4794 |
+
"grad_norm": 0.36848321557044983,
|
| 4795 |
+
"learning_rate": 1.1056986399845537e-07,
|
| 4796 |
+
"loss": 2.1475,
|
| 4797 |
+
"step": 668
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 1.9712707182320441,
|
| 4801 |
+
"grad_norm": 0.3410709500312805,
|
| 4802 |
+
"learning_rate": 8.956472646710623e-08,
|
| 4803 |
+
"loss": 1.9034,
|
| 4804 |
+
"step": 669
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 1.9742173112338857,
|
| 4808 |
+
"grad_norm": 0.37526410818099976,
|
| 4809 |
+
"learning_rate": 7.07694087889621e-08,
|
| 4810 |
+
"loss": 2.1324,
|
| 4811 |
+
"step": 670
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 1.9771639042357274,
|
| 4815 |
+
"grad_norm": 0.3631035089492798,
|
| 4816 |
+
"learning_rate": 5.418432667905204e-08,
|
| 4817 |
+
"loss": 1.9301,
|
| 4818 |
+
"step": 671
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 1.9801104972375692,
|
| 4822 |
+
"grad_norm": 0.3520108759403229,
|
| 4823 |
+
"learning_rate": 3.980984696634327e-08,
|
| 4824 |
+
"loss": 1.9756,
|
| 4825 |
+
"step": 672
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 1.9830570902394107,
|
| 4829 |
+
"grad_norm": 0.35726797580718994,
|
| 4830 |
+
"learning_rate": 2.764628758570309e-08,
|
| 4831 |
+
"loss": 1.9707,
|
| 4832 |
+
"step": 673
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 1.9860036832412522,
|
| 4836 |
+
"grad_norm": 0.35579240322113037,
|
| 4837 |
+
"learning_rate": 1.7693917570837936e-08,
|
| 4838 |
+
"loss": 1.9214,
|
| 4839 |
+
"step": 674
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 1.988950276243094,
|
| 4843 |
+
"grad_norm": 0.3749707043170929,
|
| 4844 |
+
"learning_rate": 9.95295704835364e-09,
|
| 4845 |
+
"loss": 2.0993,
|
| 4846 |
+
"step": 675
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 1.9918968692449357,
|
| 4850 |
+
"grad_norm": 0.35399675369262695,
|
| 4851 |
+
"learning_rate": 4.42357723288156e-09,
|
| 4852 |
+
"loss": 1.9808,
|
| 4853 |
+
"step": 676
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 1.9948434622467772,
|
| 4857 |
+
"grad_norm": 0.34812989830970764,
|
| 4858 |
+
"learning_rate": 1.1059004233038296e-09,
|
| 4859 |
+
"loss": 1.9791,
|
| 4860 |
+
"step": 677
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 1.9977900552486187,
|
| 4864 |
+
"grad_norm": 0.3504377603530884,
|
| 4865 |
+
"learning_rate": 0.0,
|
| 4866 |
+
"loss": 1.9639,
|
| 4867 |
+
"step": 678
|
| 4868 |
}
|
| 4869 |
],
|
| 4870 |
"logging_steps": 1,
|
|
|
|
| 4888 |
"should_evaluate": false,
|
| 4889 |
"should_log": false,
|
| 4890 |
"should_save": true,
|
| 4891 |
+
"should_training_stop": true
|
| 4892 |
},
|
| 4893 |
"attributes": {}
|
| 4894 |
}
|
| 4895 |
},
|
| 4896 |
+
"total_flos": 3.544397205114716e+17,
|
| 4897 |
"train_batch_size": 4,
|
| 4898 |
"trial_name": null,
|
| 4899 |
"trial_params": null
|