Training in progress, step 459, 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 9823216
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a621e3f29008fb453ed2cf978e69a9ebdc967a23022e7b68588ddd7464e9f5f
|
| 3 |
size 9823216
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5963308
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:272c9698c5a63826c16a28fcfeb1bf2cca0e6c50d6284f067339e1f8383f0536
|
| 3 |
size 5963308
|
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 sha256:21b5efc09a44b04d8556fedaec1c570849e7b36226a500553357208dd76bb9f2
|
| 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 sha256:47355b5abe01a117c8c3acc65c7d1228e222f8fe09358b0c7a3a25456e8bba9b
|
| 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":
|
| 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,
|
|
@@ -2422,6 +2422,804 @@
|
|
| 2422 |
"learning_rate": 1.4767198386708997e-05,
|
| 2423 |
"loss": 1.5286,
|
| 2424 |
"step": 345
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2425 |
}
|
| 2426 |
],
|
| 2427 |
"logging_steps": 1,
|
|
@@ -2436,12 +3234,12 @@
|
|
| 2436 |
"should_evaluate": false,
|
| 2437 |
"should_log": false,
|
| 2438 |
"should_save": true,
|
| 2439 |
-
"should_training_stop":
|
| 2440 |
},
|
| 2441 |
"attributes": {}
|
| 2442 |
}
|
| 2443 |
},
|
| 2444 |
-
"total_flos":
|
| 2445 |
"train_batch_size": 4,
|
| 2446 |
"trial_name": null,
|
| 2447 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 459,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2422 |
"learning_rate": 1.4767198386708997e-05,
|
| 2423 |
"loss": 1.5286,
|
| 2424 |
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7538126361655774,
|
| 2428 |
+
"grad_norm": 0.4534967541694641,
|
| 2429 |
+
"learning_rate": 1.4522546730656129e-05,
|
| 2430 |
+
"loss": 1.4874,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7559912854030502,
|
| 2435 |
+
"grad_norm": 0.4711008369922638,
|
| 2436 |
+
"learning_rate": 1.427959386087761e-05,
|
| 2437 |
+
"loss": 1.5148,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7581699346405228,
|
| 2442 |
+
"grad_norm": 0.433192640542984,
|
| 2443 |
+
"learning_rate": 1.4038351410818434e-05,
|
| 2444 |
+
"loss": 1.6257,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7603485838779956,
|
| 2449 |
+
"grad_norm": 0.4659889042377472,
|
| 2450 |
+
"learning_rate": 1.3798830932022617e-05,
|
| 2451 |
+
"loss": 1.6029,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.7625272331154684,
|
| 2456 |
+
"grad_norm": 0.5066768527030945,
|
| 2457 |
+
"learning_rate": 1.3561043893580083e-05,
|
| 2458 |
+
"loss": 1.6275,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.7647058823529411,
|
| 2463 |
+
"grad_norm": 0.4900889992713928,
|
| 2464 |
+
"learning_rate": 1.3325001681577482e-05,
|
| 2465 |
+
"loss": 1.8466,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.7668845315904139,
|
| 2470 |
+
"grad_norm": 0.39974841475486755,
|
| 2471 |
+
"learning_rate": 1.3090715598552999e-05,
|
| 2472 |
+
"loss": 1.6942,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.7690631808278867,
|
| 2477 |
+
"grad_norm": 0.41923969984054565,
|
| 2478 |
+
"learning_rate": 1.2858196862955107e-05,
|
| 2479 |
+
"loss": 1.4645,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.7712418300653595,
|
| 2484 |
+
"grad_norm": 0.39095959067344666,
|
| 2485 |
+
"learning_rate": 1.2627456608605443e-05,
|
| 2486 |
+
"loss": 1.6083,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.7734204793028322,
|
| 2491 |
+
"grad_norm": 0.3984636068344116,
|
| 2492 |
+
"learning_rate": 1.239850588416565e-05,
|
| 2493 |
+
"loss": 1.6569,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.775599128540305,
|
| 2498 |
+
"grad_norm": 0.4196769893169403,
|
| 2499 |
+
"learning_rate": 1.217135565260833e-05,
|
| 2500 |
+
"loss": 1.6212,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.7777777777777778,
|
| 2505 |
+
"grad_norm": 0.4557774066925049,
|
| 2506 |
+
"learning_rate": 1.1946016790692094e-05,
|
| 2507 |
+
"loss": 1.6636,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.7799564270152506,
|
| 2512 |
+
"grad_norm": 0.3991740942001343,
|
| 2513 |
+
"learning_rate": 1.1722500088440768e-05,
|
| 2514 |
+
"loss": 1.4824,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.7821350762527233,
|
| 2519 |
+
"grad_norm": 0.44218507409095764,
|
| 2520 |
+
"learning_rate": 1.1500816248626712e-05,
|
| 2521 |
+
"loss": 1.5913,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.7843137254901961,
|
| 2526 |
+
"grad_norm": 0.3721107542514801,
|
| 2527 |
+
"learning_rate": 1.1280975886258293e-05,
|
| 2528 |
+
"loss": 1.5063,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.7864923747276689,
|
| 2533 |
+
"grad_norm": 0.5461967587471008,
|
| 2534 |
+
"learning_rate": 1.1062989528071682e-05,
|
| 2535 |
+
"loss": 1.7895,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.7886710239651417,
|
| 2540 |
+
"grad_norm": 0.38676172494888306,
|
| 2541 |
+
"learning_rate": 1.0846867612026745e-05,
|
| 2542 |
+
"loss": 1.4693,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.7908496732026143,
|
| 2547 |
+
"grad_norm": 0.45002123713493347,
|
| 2548 |
+
"learning_rate": 1.0632620486807243e-05,
|
| 2549 |
+
"loss": 1.5761,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.7930283224400871,
|
| 2554 |
+
"grad_norm": 0.384592741727829,
|
| 2555 |
+
"learning_rate": 1.0420258411325307e-05,
|
| 2556 |
+
"loss": 1.4709,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.7952069716775599,
|
| 2561 |
+
"grad_norm": 0.39452260732650757,
|
| 2562 |
+
"learning_rate": 1.0209791554230209e-05,
|
| 2563 |
+
"loss": 1.6416,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.7973856209150327,
|
| 2568 |
+
"grad_norm": 0.4410095512866974,
|
| 2569 |
+
"learning_rate": 1.0001229993421412e-05,
|
| 2570 |
+
"loss": 1.6746,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.7995642701525054,
|
| 2575 |
+
"grad_norm": 0.4333903193473816,
|
| 2576 |
+
"learning_rate": 9.79458371556607e-06,
|
| 2577 |
+
"loss": 1.5514,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.8017429193899782,
|
| 2582 |
+
"grad_norm": 0.38147422671318054,
|
| 2583 |
+
"learning_rate": 9.589862615620782e-06,
|
| 2584 |
+
"loss": 1.5221,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.803921568627451,
|
| 2589 |
+
"grad_norm": 0.4305866062641144,
|
| 2590 |
+
"learning_rate": 9.387076496357805e-06,
|
| 2591 |
+
"loss": 1.5458,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8061002178649237,
|
| 2596 |
+
"grad_norm": 0.42009007930755615,
|
| 2597 |
+
"learning_rate": 9.186235067895672e-06,
|
| 2598 |
+
"loss": 1.4395,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8082788671023965,
|
| 2603 |
+
"grad_norm": 0.42722266912460327,
|
| 2604 |
+
"learning_rate": 8.987347947234192e-06,
|
| 2605 |
+
"loss": 1.5633,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.8104575163398693,
|
| 2610 |
+
"grad_norm": 0.44506317377090454,
|
| 2611 |
+
"learning_rate": 8.790424657794034e-06,
|
| 2612 |
+
"loss": 1.5972,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.8126361655773421,
|
| 2617 |
+
"grad_norm": 0.44512099027633667,
|
| 2618 |
+
"learning_rate": 8.595474628960598e-06,
|
| 2619 |
+
"loss": 1.5575,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8148148148148148,
|
| 2624 |
+
"grad_norm": 0.41824987530708313,
|
| 2625 |
+
"learning_rate": 8.402507195632626e-06,
|
| 2626 |
+
"loss": 1.6123,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.8169934640522876,
|
| 2631 |
+
"grad_norm": 0.4357855021953583,
|
| 2632 |
+
"learning_rate": 8.211531597775134e-06,
|
| 2633 |
+
"loss": 1.6286,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8191721132897604,
|
| 2638 |
+
"grad_norm": 0.42029663920402527,
|
| 2639 |
+
"learning_rate": 8.022556979976992e-06,
|
| 2640 |
+
"loss": 1.4422,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8213507625272332,
|
| 2645 |
+
"grad_norm": 0.4830271005630493,
|
| 2646 |
+
"learning_rate": 7.835592391013053e-06,
|
| 2647 |
+
"loss": 1.7282,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.8235294117647058,
|
| 2652 |
+
"grad_norm": 0.45707976818084717,
|
| 2653 |
+
"learning_rate": 7.650646783410875e-06,
|
| 2654 |
+
"loss": 1.456,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8257080610021786,
|
| 2659 |
+
"grad_norm": 0.42442944645881653,
|
| 2660 |
+
"learning_rate": 7.4677290130219794e-06,
|
| 2661 |
+
"loss": 1.6168,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8278867102396514,
|
| 2666 |
+
"grad_norm": 0.448371946811676,
|
| 2667 |
+
"learning_rate": 7.286847838597905e-06,
|
| 2668 |
+
"loss": 1.6022,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8300653594771242,
|
| 2673 |
+
"grad_norm": 0.40033116936683655,
|
| 2674 |
+
"learning_rate": 7.108011921370728e-06,
|
| 2675 |
+
"loss": 1.5397,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.8322440087145969,
|
| 2680 |
+
"grad_norm": 0.45119813084602356,
|
| 2681 |
+
"learning_rate": 6.9312298246383575e-06,
|
| 2682 |
+
"loss": 1.6904,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.8344226579520697,
|
| 2687 |
+
"grad_norm": 0.43013447523117065,
|
| 2688 |
+
"learning_rate": 6.7565100133545115e-06,
|
| 2689 |
+
"loss": 1.5514,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.8366013071895425,
|
| 2694 |
+
"grad_norm": 0.42408111691474915,
|
| 2695 |
+
"learning_rate": 6.583860853723339e-06,
|
| 2696 |
+
"loss": 1.5711,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.8387799564270153,
|
| 2701 |
+
"grad_norm": 0.45047736167907715,
|
| 2702 |
+
"learning_rate": 6.413290612798884e-06,
|
| 2703 |
+
"loss": 1.5574,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.840958605664488,
|
| 2708 |
+
"grad_norm": 0.4935206472873688,
|
| 2709 |
+
"learning_rate": 6.24480745808913e-06,
|
| 2710 |
+
"loss": 1.6341,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.8431372549019608,
|
| 2715 |
+
"grad_norm": 0.4137425124645233,
|
| 2716 |
+
"learning_rate": 6.078419457165036e-06,
|
| 2717 |
+
"loss": 1.5075,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8453159041394336,
|
| 2722 |
+
"grad_norm": 0.4301162362098694,
|
| 2723 |
+
"learning_rate": 5.914134577274122e-06,
|
| 2724 |
+
"loss": 1.6337,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8474945533769063,
|
| 2729 |
+
"grad_norm": 0.4462296962738037,
|
| 2730 |
+
"learning_rate": 5.751960684959046e-06,
|
| 2731 |
+
"loss": 1.5962,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8496732026143791,
|
| 2736 |
+
"grad_norm": 0.44939714670181274,
|
| 2737 |
+
"learning_rate": 5.59190554568087e-06,
|
| 2738 |
+
"loss": 1.5191,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8518518518518519,
|
| 2743 |
+
"grad_norm": 0.4732544720172882,
|
| 2744 |
+
"learning_rate": 5.4339768234472625e-06,
|
| 2745 |
+
"loss": 1.5137,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8540305010893247,
|
| 2750 |
+
"grad_norm": 0.4495387077331543,
|
| 2751 |
+
"learning_rate": 5.27818208044551e-06,
|
| 2752 |
+
"loss": 1.5465,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8562091503267973,
|
| 2757 |
+
"grad_norm": 0.5142303109169006,
|
| 2758 |
+
"learning_rate": 5.124528776680371e-06,
|
| 2759 |
+
"loss": 1.7364,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8583877995642701,
|
| 2764 |
+
"grad_norm": 0.4349537789821625,
|
| 2765 |
+
"learning_rate": 4.9730242696169325e-06,
|
| 2766 |
+
"loss": 1.4617,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.8605664488017429,
|
| 2771 |
+
"grad_norm": 0.46584171056747437,
|
| 2772 |
+
"learning_rate": 4.8236758138282715e-06,
|
| 2773 |
+
"loss": 1.5325,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.8627450980392157,
|
| 2778 |
+
"grad_norm": 0.5047036409378052,
|
| 2779 |
+
"learning_rate": 4.676490560648067e-06,
|
| 2780 |
+
"loss": 1.5519,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.8649237472766884,
|
| 2785 |
+
"grad_norm": 0.47344809770584106,
|
| 2786 |
+
"learning_rate": 4.531475557828202e-06,
|
| 2787 |
+
"loss": 1.5025,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.8671023965141612,
|
| 2792 |
+
"grad_norm": 0.47757408022880554,
|
| 2793 |
+
"learning_rate": 4.388637749201274e-06,
|
| 2794 |
+
"loss": 1.5715,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.869281045751634,
|
| 2799 |
+
"grad_norm": 0.5043246746063232,
|
| 2800 |
+
"learning_rate": 4.247983974348096e-06,
|
| 2801 |
+
"loss": 1.6822,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.8714596949891068,
|
| 2806 |
+
"grad_norm": 0.6163721084594727,
|
| 2807 |
+
"learning_rate": 4.109520968270198e-06,
|
| 2808 |
+
"loss": 1.528,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.8736383442265795,
|
| 2813 |
+
"grad_norm": 0.41478580236434937,
|
| 2814 |
+
"learning_rate": 3.973255361067346e-06,
|
| 2815 |
+
"loss": 1.6695,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.8758169934640523,
|
| 2820 |
+
"grad_norm": 0.4069858491420746,
|
| 2821 |
+
"learning_rate": 3.839193677620028e-06,
|
| 2822 |
+
"loss": 1.572,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.8779956427015251,
|
| 2827 |
+
"grad_norm": 0.44524499773979187,
|
| 2828 |
+
"learning_rate": 3.7073423372770753e-06,
|
| 2829 |
+
"loss": 1.6545,
|
| 2830 |
+
"step": 403
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.8801742919389978,
|
| 2834 |
+
"grad_norm": 0.40681320428848267,
|
| 2835 |
+
"learning_rate": 3.577707653548229e-06,
|
| 2836 |
+
"loss": 1.6042,
|
| 2837 |
+
"step": 404
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.8823529411764706,
|
| 2841 |
+
"grad_norm": 0.39351528882980347,
|
| 2842 |
+
"learning_rate": 3.4502958338018753e-06,
|
| 2843 |
+
"loss": 1.5133,
|
| 2844 |
+
"step": 405
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.8845315904139434,
|
| 2848 |
+
"grad_norm": 0.4517339766025543,
|
| 2849 |
+
"learning_rate": 3.325112978967776e-06,
|
| 2850 |
+
"loss": 1.6433,
|
| 2851 |
+
"step": 406
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.8867102396514162,
|
| 2855 |
+
"grad_norm": 0.406124085187912,
|
| 2856 |
+
"learning_rate": 3.20216508324494e-06,
|
| 2857 |
+
"loss": 1.5657,
|
| 2858 |
+
"step": 407
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.8888888888888888,
|
| 2862 |
+
"grad_norm": 0.40703821182250977,
|
| 2863 |
+
"learning_rate": 3.081458033814627e-06,
|
| 2864 |
+
"loss": 1.6277,
|
| 2865 |
+
"step": 408
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.8910675381263616,
|
| 2869 |
+
"grad_norm": 0.44792646169662476,
|
| 2870 |
+
"learning_rate": 2.9629976105584266e-06,
|
| 2871 |
+
"loss": 1.703,
|
| 2872 |
+
"step": 409
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.8932461873638344,
|
| 2876 |
+
"grad_norm": 0.42939573526382446,
|
| 2877 |
+
"learning_rate": 2.8467894857814814e-06,
|
| 2878 |
+
"loss": 1.6469,
|
| 2879 |
+
"step": 410
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.8954248366013072,
|
| 2883 |
+
"grad_norm": 0.41463714838027954,
|
| 2884 |
+
"learning_rate": 2.732839223940914e-06,
|
| 2885 |
+
"loss": 1.4841,
|
| 2886 |
+
"step": 411
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.8976034858387799,
|
| 2890 |
+
"grad_norm": 0.44514474272727966,
|
| 2891 |
+
"learning_rate": 2.621152281379352e-06,
|
| 2892 |
+
"loss": 1.7912,
|
| 2893 |
+
"step": 412
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.8997821350762527,
|
| 2897 |
+
"grad_norm": 0.4165302515029907,
|
| 2898 |
+
"learning_rate": 2.511734006063682e-06,
|
| 2899 |
+
"loss": 1.5175,
|
| 2900 |
+
"step": 413
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9019607843137255,
|
| 2904 |
+
"grad_norm": 0.40819787979125977,
|
| 2905 |
+
"learning_rate": 2.4045896373289467e-06,
|
| 2906 |
+
"loss": 1.5655,
|
| 2907 |
+
"step": 414
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.9041394335511983,
|
| 2911 |
+
"grad_norm": 0.4505191445350647,
|
| 2912 |
+
"learning_rate": 2.299724305627482e-06,
|
| 2913 |
+
"loss": 1.5763,
|
| 2914 |
+
"step": 415
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.906318082788671,
|
| 2918 |
+
"grad_norm": 0.42655885219573975,
|
| 2919 |
+
"learning_rate": 2.1971430322832554e-06,
|
| 2920 |
+
"loss": 1.536,
|
| 2921 |
+
"step": 416
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.9084967320261438,
|
| 2925 |
+
"grad_norm": 0.4479508101940155,
|
| 2926 |
+
"learning_rate": 2.0968507292514037e-06,
|
| 2927 |
+
"loss": 1.574,
|
| 2928 |
+
"step": 417
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.9106753812636166,
|
| 2932 |
+
"grad_norm": 0.39721792936325073,
|
| 2933 |
+
"learning_rate": 1.998852198883061e-06,
|
| 2934 |
+
"loss": 1.5643,
|
| 2935 |
+
"step": 418
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9128540305010894,
|
| 2939 |
+
"grad_norm": 0.4025076627731323,
|
| 2940 |
+
"learning_rate": 1.903152133695385e-06,
|
| 2941 |
+
"loss": 1.5924,
|
| 2942 |
+
"step": 419
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9150326797385621,
|
| 2946 |
+
"grad_norm": 0.426258385181427,
|
| 2947 |
+
"learning_rate": 1.8097551161468773e-06,
|
| 2948 |
+
"loss": 1.5415,
|
| 2949 |
+
"step": 420
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.9172113289760349,
|
| 2953 |
+
"grad_norm": 0.39357250928878784,
|
| 2954 |
+
"learning_rate": 1.7186656184179472e-06,
|
| 2955 |
+
"loss": 1.5431,
|
| 2956 |
+
"step": 421
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9193899782135077,
|
| 2960 |
+
"grad_norm": 0.4268351197242737,
|
| 2961 |
+
"learning_rate": 1.6298880021967665e-06,
|
| 2962 |
+
"loss": 1.6426,
|
| 2963 |
+
"step": 422
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.9215686274509803,
|
| 2967 |
+
"grad_norm": 0.43658286333084106,
|
| 2968 |
+
"learning_rate": 1.543426518470431e-06,
|
| 2969 |
+
"loss": 1.5601,
|
| 2970 |
+
"step": 423
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9237472766884531,
|
| 2974 |
+
"grad_norm": 0.4358772337436676,
|
| 2975 |
+
"learning_rate": 1.4592853073214008e-06,
|
| 2976 |
+
"loss": 1.6446,
|
| 2977 |
+
"step": 424
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9259259259259259,
|
| 2981 |
+
"grad_norm": 0.4013616740703583,
|
| 2982 |
+
"learning_rate": 1.3774683977292424e-06,
|
| 2983 |
+
"loss": 1.4854,
|
| 2984 |
+
"step": 425
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.9281045751633987,
|
| 2988 |
+
"grad_norm": 0.4512793719768524,
|
| 2989 |
+
"learning_rate": 1.2979797073777334e-06,
|
| 2990 |
+
"loss": 1.6231,
|
| 2991 |
+
"step": 426
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.9302832244008714,
|
| 2995 |
+
"grad_norm": 0.41178834438323975,
|
| 2996 |
+
"learning_rate": 1.2208230424672562e-06,
|
| 2997 |
+
"loss": 1.4498,
|
| 2998 |
+
"step": 427
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.9324618736383442,
|
| 3002 |
+
"grad_norm": 0.45710593461990356,
|
| 3003 |
+
"learning_rate": 1.1460020975325391e-06,
|
| 3004 |
+
"loss": 1.672,
|
| 3005 |
+
"step": 428
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.934640522875817,
|
| 3009 |
+
"grad_norm": 0.43900266289711,
|
| 3010 |
+
"learning_rate": 1.0735204552657642e-06,
|
| 3011 |
+
"loss": 1.6774,
|
| 3012 |
+
"step": 429
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.9368191721132898,
|
| 3016 |
+
"grad_norm": 0.46783003211021423,
|
| 3017 |
+
"learning_rate": 1.0033815863449981e-06,
|
| 3018 |
+
"loss": 1.5636,
|
| 3019 |
+
"step": 430
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.9389978213507625,
|
| 3023 |
+
"grad_norm": 0.425657719373703,
|
| 3024 |
+
"learning_rate": 9.355888492680154e-07,
|
| 3025 |
+
"loss": 1.724,
|
| 3026 |
+
"step": 431
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.9411764705882353,
|
| 3030 |
+
"grad_norm": 0.4375082552433014,
|
| 3031 |
+
"learning_rate": 8.701454901914762e-07,
|
| 3032 |
+
"loss": 1.5833,
|
| 3033 |
+
"step": 432
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.9433551198257081,
|
| 3037 |
+
"grad_norm": 0.4675833582878113,
|
| 3038 |
+
"learning_rate": 8.0705464277549e-07,
|
| 3039 |
+
"loss": 1.5491,
|
| 3040 |
+
"step": 433
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.9455337690631809,
|
| 3044 |
+
"grad_norm": 0.40916916728019714,
|
| 3045 |
+
"learning_rate": 7.463193280335679e-07,
|
| 3046 |
+
"loss": 1.5551,
|
| 3047 |
+
"step": 434
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.9477124183006536,
|
| 3051 |
+
"grad_norm": 0.4476478099822998,
|
| 3052 |
+
"learning_rate": 6.879424541879676e-07,
|
| 3053 |
+
"loss": 1.3743,
|
| 3054 |
+
"step": 435
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9498910675381264,
|
| 3058 |
+
"grad_norm": 0.46898120641708374,
|
| 3059 |
+
"learning_rate": 6.319268165304204e-07,
|
| 3060 |
+
"loss": 1.6789,
|
| 3061 |
+
"step": 436
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9520697167755992,
|
| 3065 |
+
"grad_norm": 0.4508022665977478,
|
| 3066 |
+
"learning_rate": 5.782750972883111e-07,
|
| 3067 |
+
"loss": 1.6429,
|
| 3068 |
+
"step": 437
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.954248366013072,
|
| 3072 |
+
"grad_norm": 0.4190502464771271,
|
| 3073 |
+
"learning_rate": 5.26989865496208e-07,
|
| 3074 |
+
"loss": 1.5525,
|
| 3075 |
+
"step": 438
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.9564270152505446,
|
| 3079 |
+
"grad_norm": 0.4985792934894562,
|
| 3080 |
+
"learning_rate": 4.780735768728895e-07,
|
| 3081 |
+
"loss": 1.6118,
|
| 3082 |
+
"step": 439
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.9586056644880174,
|
| 3086 |
+
"grad_norm": 0.4321351945400238,
|
| 3087 |
+
"learning_rate": 4.3152857370371557e-07,
|
| 3088 |
+
"loss": 1.6268,
|
| 3089 |
+
"step": 440
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.9607843137254902,
|
| 3093 |
+
"grad_norm": 0.44800451397895813,
|
| 3094 |
+
"learning_rate": 3.873570847285013e-07,
|
| 3095 |
+
"loss": 1.4648,
|
| 3096 |
+
"step": 441
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.9629629629629629,
|
| 3100 |
+
"grad_norm": 0.44108688831329346,
|
| 3101 |
+
"learning_rate": 3.455612250347851e-07,
|
| 3102 |
+
"loss": 1.5555,
|
| 3103 |
+
"step": 442
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.9651416122004357,
|
| 3107 |
+
"grad_norm": 0.4282665550708771,
|
| 3108 |
+
"learning_rate": 3.061429959565487e-07,
|
| 3109 |
+
"loss": 1.4951,
|
| 3110 |
+
"step": 443
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.9673202614379085,
|
| 3114 |
+
"grad_norm": 0.4284363389015198,
|
| 3115 |
+
"learning_rate": 2.691042849783776e-07,
|
| 3116 |
+
"loss": 1.5738,
|
| 3117 |
+
"step": 444
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.9694989106753813,
|
| 3121 |
+
"grad_norm": 0.504803478717804,
|
| 3122 |
+
"learning_rate": 2.3444686564511043e-07,
|
| 3123 |
+
"loss": 1.5243,
|
| 3124 |
+
"step": 445
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.971677559912854,
|
| 3128 |
+
"grad_norm": 0.4652966558933258,
|
| 3129 |
+
"learning_rate": 2.0217239747689078e-07,
|
| 3130 |
+
"loss": 1.5924,
|
| 3131 |
+
"step": 446
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.9738562091503268,
|
| 3135 |
+
"grad_norm": 0.47274574637413025,
|
| 3136 |
+
"learning_rate": 1.722824258896971e-07,
|
| 3137 |
+
"loss": 1.4842,
|
| 3138 |
+
"step": 447
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.9760348583877996,
|
| 3142 |
+
"grad_norm": 0.4845008850097656,
|
| 3143 |
+
"learning_rate": 1.447783821213744e-07,
|
| 3144 |
+
"loss": 1.5003,
|
| 3145 |
+
"step": 448
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.9782135076252724,
|
| 3149 |
+
"grad_norm": 0.4765538275241852,
|
| 3150 |
+
"learning_rate": 1.1966158316307208e-07,
|
| 3151 |
+
"loss": 1.6308,
|
| 3152 |
+
"step": 449
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.9803921568627451,
|
| 3156 |
+
"grad_norm": 0.5732904672622681,
|
| 3157 |
+
"learning_rate": 9.693323169619462e-08,
|
| 3158 |
+
"loss": 1.6595,
|
| 3159 |
+
"step": 450
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.9825708061002179,
|
| 3163 |
+
"grad_norm": 0.4185521602630615,
|
| 3164 |
+
"learning_rate": 7.659441603481421e-08,
|
| 3165 |
+
"loss": 1.5248,
|
| 3166 |
+
"step": 451
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.9847494553376906,
|
| 3170 |
+
"grad_norm": 0.43158915638923645,
|
| 3171 |
+
"learning_rate": 5.864611007354581e-08,
|
| 3172 |
+
"loss": 1.5917,
|
| 3173 |
+
"step": 452
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.9869281045751634,
|
| 3177 |
+
"grad_norm": 0.4398334324359894,
|
| 3178 |
+
"learning_rate": 4.308917324092887e-08,
|
| 3179 |
+
"loss": 1.7976,
|
| 3180 |
+
"step": 453
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.9891067538126361,
|
| 3184 |
+
"grad_norm": 0.44609612226486206,
|
| 3185 |
+
"learning_rate": 2.9924350458271357e-08,
|
| 3186 |
+
"loss": 1.5595,
|
| 3187 |
+
"step": 454
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.9912854030501089,
|
| 3191 |
+
"grad_norm": 0.4355579912662506,
|
| 3192 |
+
"learning_rate": 1.9152272103972747e-08,
|
| 3193 |
+
"loss": 1.6064,
|
| 3194 |
+
"step": 455
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.9934640522875817,
|
| 3198 |
+
"grad_norm": 0.4061749279499054,
|
| 3199 |
+
"learning_rate": 1.0773453983342618e-08,
|
| 3200 |
+
"loss": 1.658,
|
| 3201 |
+
"step": 456
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.9956427015250545,
|
| 3205 |
+
"grad_norm": 0.41987109184265137,
|
| 3206 |
+
"learning_rate": 4.7882973039037326e-09,
|
| 3207 |
+
"loss": 1.5099,
|
| 3208 |
+
"step": 457
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.9978213507625272,
|
| 3212 |
+
"grad_norm": 0.4400121569633484,
|
| 3213 |
+
"learning_rate": 1.1970886561907258e-09,
|
| 3214 |
+
"loss": 1.5026,
|
| 3215 |
+
"step": 458
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 1.0,
|
| 3219 |
+
"grad_norm": 0.47257599234580994,
|
| 3220 |
+
"learning_rate": 0.0,
|
| 3221 |
+
"loss": 1.6942,
|
| 3222 |
+
"step": 459
|
| 3223 |
}
|
| 3224 |
],
|
| 3225 |
"logging_steps": 1,
|
|
|
|
| 3234 |
"should_evaluate": false,
|
| 3235 |
"should_log": false,
|
| 3236 |
"should_save": true,
|
| 3237 |
+
"should_training_stop": true
|
| 3238 |
},
|
| 3239 |
"attributes": {}
|
| 3240 |
}
|
| 3241 |
},
|
| 3242 |
+
"total_flos": 4931631526182912.0,
|
| 3243 |
"train_batch_size": 4,
|
| 3244 |
"trial_name": null,
|
| 3245 |
"trial_params": null
|