Training in progress, step 475, 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 167832240
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c444a5739e2bec3b1e3f04bd30a01356e1edcb8c01b2ba0dd74328accd2825b2
|
| 3 |
size 167832240
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 85723732
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:08d25728ac5fe47a0ad27b3da864cab99af91f3548996d69185914cd67c672bb
|
| 3 |
size 85723732
|
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 sha256:1126b5614f51e5fa36121a9290d6b01e210532249374ad3c426067b6e3d80cfd
|
| 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 sha256:5d3ab8b8dc5babc32b4adc3c596b50dd0fcac27b238d3838d86c3c68054c541d
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0
|
| 5 |
"eval_steps": 119,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2538,6 +2538,832 @@
|
|
| 2538 |
"eval_samples_per_second": 4.073,
|
| 2539 |
"eval_steps_per_second": 4.073,
|
| 2540 |
"step": 357
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2541 |
}
|
| 2542 |
],
|
| 2543 |
"logging_steps": 1,
|
|
@@ -2552,12 +3378,12 @@
|
|
| 2552 |
"should_evaluate": false,
|
| 2553 |
"should_log": false,
|
| 2554 |
"should_save": true,
|
| 2555 |
-
"should_training_stop":
|
| 2556 |
},
|
| 2557 |
"attributes": {}
|
| 2558 |
}
|
| 2559 |
},
|
| 2560 |
-
"total_flos":
|
| 2561 |
"train_batch_size": 1,
|
| 2562 |
"trial_name": null,
|
| 2563 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 119,
|
| 6 |
+
"global_step": 475,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2538 |
"eval_samples_per_second": 4.073,
|
| 2539 |
"eval_steps_per_second": 4.073,
|
| 2540 |
"step": 357
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 0.7536842105263157,
|
| 2544 |
+
"grad_norm": 0.8696005940437317,
|
| 2545 |
+
"learning_rate": 2.964852511800519e-05,
|
| 2546 |
+
"loss": 1.0904,
|
| 2547 |
+
"step": 358
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.7557894736842106,
|
| 2551 |
+
"grad_norm": 0.7804014086723328,
|
| 2552 |
+
"learning_rate": 2.9169990463359555e-05,
|
| 2553 |
+
"loss": 1.0226,
|
| 2554 |
+
"step": 359
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 0.7578947368421053,
|
| 2558 |
+
"grad_norm": 0.8228606581687927,
|
| 2559 |
+
"learning_rate": 2.869468883687798e-05,
|
| 2560 |
+
"loss": 1.3508,
|
| 2561 |
+
"step": 360
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 0.76,
|
| 2565 |
+
"grad_norm": 0.6658822298049927,
|
| 2566 |
+
"learning_rate": 2.8222641933652117e-05,
|
| 2567 |
+
"loss": 1.1766,
|
| 2568 |
+
"step": 361
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 0.7621052631578947,
|
| 2572 |
+
"grad_norm": 0.7910031676292419,
|
| 2573 |
+
"learning_rate": 2.7753871300212142e-05,
|
| 2574 |
+
"loss": 0.7955,
|
| 2575 |
+
"step": 362
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 0.7642105263157895,
|
| 2579 |
+
"grad_norm": 0.7152170538902283,
|
| 2580 |
+
"learning_rate": 2.7288398333543064e-05,
|
| 2581 |
+
"loss": 0.9675,
|
| 2582 |
+
"step": 363
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 0.7663157894736842,
|
| 2586 |
+
"grad_norm": 0.5529299974441528,
|
| 2587 |
+
"learning_rate": 2.6826244280108437e-05,
|
| 2588 |
+
"loss": 0.9956,
|
| 2589 |
+
"step": 364
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 0.7684210526315789,
|
| 2593 |
+
"grad_norm": 0.7928904891014099,
|
| 2594 |
+
"learning_rate": 2.6367430234880284e-05,
|
| 2595 |
+
"loss": 0.963,
|
| 2596 |
+
"step": 365
|
| 2597 |
+
},
|
| 2598 |
+
{
|
| 2599 |
+
"epoch": 0.7705263157894737,
|
| 2600 |
+
"grad_norm": 0.8560205101966858,
|
| 2601 |
+
"learning_rate": 2.591197714037631e-05,
|
| 2602 |
+
"loss": 1.164,
|
| 2603 |
+
"step": 366
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 0.7726315789473684,
|
| 2607 |
+
"grad_norm": 0.5956188440322876,
|
| 2608 |
+
"learning_rate": 2.5459905785704042e-05,
|
| 2609 |
+
"loss": 1.1437,
|
| 2610 |
+
"step": 367
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 0.7747368421052632,
|
| 2614 |
+
"grad_norm": 1.018763542175293,
|
| 2615 |
+
"learning_rate": 2.5011236805611814e-05,
|
| 2616 |
+
"loss": 1.1257,
|
| 2617 |
+
"step": 368
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 0.7768421052631579,
|
| 2621 |
+
"grad_norm": 2.8923377990722656,
|
| 2622 |
+
"learning_rate": 2.4565990679546914e-05,
|
| 2623 |
+
"loss": 0.8183,
|
| 2624 |
+
"step": 369
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 0.7789473684210526,
|
| 2628 |
+
"grad_norm": 0.7362539172172546,
|
| 2629 |
+
"learning_rate": 2.4124187730720917e-05,
|
| 2630 |
+
"loss": 1.2692,
|
| 2631 |
+
"step": 370
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 0.7810526315789473,
|
| 2635 |
+
"grad_norm": 0.9826866388320923,
|
| 2636 |
+
"learning_rate": 2.368584812518184e-05,
|
| 2637 |
+
"loss": 0.8437,
|
| 2638 |
+
"step": 371
|
| 2639 |
+
},
|
| 2640 |
+
{
|
| 2641 |
+
"epoch": 0.783157894736842,
|
| 2642 |
+
"grad_norm": 0.8725941181182861,
|
| 2643 |
+
"learning_rate": 2.3250991870893835e-05,
|
| 2644 |
+
"loss": 1.0641,
|
| 2645 |
+
"step": 372
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 0.7852631578947369,
|
| 2649 |
+
"grad_norm": 2.081977605819702,
|
| 2650 |
+
"learning_rate": 2.2819638816823797e-05,
|
| 2651 |
+
"loss": 1.1523,
|
| 2652 |
+
"step": 373
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 0.7873684210526316,
|
| 2656 |
+
"grad_norm": 2.948880434036255,
|
| 2657 |
+
"learning_rate": 2.2391808652035517e-05,
|
| 2658 |
+
"loss": 0.6608,
|
| 2659 |
+
"step": 374
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 0.7894736842105263,
|
| 2663 |
+
"grad_norm": 0.845608651638031,
|
| 2664 |
+
"learning_rate": 2.1967520904790827e-05,
|
| 2665 |
+
"loss": 1.0555,
|
| 2666 |
+
"step": 375
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 0.791578947368421,
|
| 2670 |
+
"grad_norm": 0.850184440612793,
|
| 2671 |
+
"learning_rate": 2.154679494165829e-05,
|
| 2672 |
+
"loss": 1.1188,
|
| 2673 |
+
"step": 376
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 0.7936842105263158,
|
| 2677 |
+
"grad_norm": 0.8228728175163269,
|
| 2678 |
+
"learning_rate": 2.1129649966629184e-05,
|
| 2679 |
+
"loss": 0.9344,
|
| 2680 |
+
"step": 377
|
| 2681 |
+
},
|
| 2682 |
+
{
|
| 2683 |
+
"epoch": 0.7957894736842105,
|
| 2684 |
+
"grad_norm": 0.9157503843307495,
|
| 2685 |
+
"learning_rate": 2.0716105020241072e-05,
|
| 2686 |
+
"loss": 0.7635,
|
| 2687 |
+
"step": 378
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 0.7978947368421052,
|
| 2691 |
+
"grad_norm": 1.4381229877471924,
|
| 2692 |
+
"learning_rate": 2.0306178978708514e-05,
|
| 2693 |
+
"loss": 1.258,
|
| 2694 |
+
"step": 379
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 0.8,
|
| 2698 |
+
"grad_norm": 0.6052026748657227,
|
| 2699 |
+
"learning_rate": 1.9899890553061562e-05,
|
| 2700 |
+
"loss": 1.1365,
|
| 2701 |
+
"step": 380
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 0.8021052631578948,
|
| 2705 |
+
"grad_norm": 0.8866299390792847,
|
| 2706 |
+
"learning_rate": 1.9497258288291654e-05,
|
| 2707 |
+
"loss": 1.0385,
|
| 2708 |
+
"step": 381
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 0.8042105263157895,
|
| 2712 |
+
"grad_norm": 0.743588387966156,
|
| 2713 |
+
"learning_rate": 1.9098300562505266e-05,
|
| 2714 |
+
"loss": 1.0008,
|
| 2715 |
+
"step": 382
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 0.8063157894736842,
|
| 2719 |
+
"grad_norm": 0.797770082950592,
|
| 2720 |
+
"learning_rate": 1.8703035586084816e-05,
|
| 2721 |
+
"loss": 0.82,
|
| 2722 |
+
"step": 383
|
| 2723 |
+
},
|
| 2724 |
+
{
|
| 2725 |
+
"epoch": 0.8084210526315789,
|
| 2726 |
+
"grad_norm": 1.2986137866973877,
|
| 2727 |
+
"learning_rate": 1.831148140085762e-05,
|
| 2728 |
+
"loss": 0.9923,
|
| 2729 |
+
"step": 384
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 0.8105263157894737,
|
| 2733 |
+
"grad_norm": 1.0393884181976318,
|
| 2734 |
+
"learning_rate": 1.7923655879272393e-05,
|
| 2735 |
+
"loss": 1.4782,
|
| 2736 |
+
"step": 385
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 0.8126315789473684,
|
| 2740 |
+
"grad_norm": 0.961280882358551,
|
| 2741 |
+
"learning_rate": 1.753957672358324e-05,
|
| 2742 |
+
"loss": 0.9412,
|
| 2743 |
+
"step": 386
|
| 2744 |
+
},
|
| 2745 |
+
{
|
| 2746 |
+
"epoch": 0.8147368421052632,
|
| 2747 |
+
"grad_norm": 0.5964862108230591,
|
| 2748 |
+
"learning_rate": 1.7159261465041952e-05,
|
| 2749 |
+
"loss": 0.9657,
|
| 2750 |
+
"step": 387
|
| 2751 |
+
},
|
| 2752 |
+
{
|
| 2753 |
+
"epoch": 0.8168421052631579,
|
| 2754 |
+
"grad_norm": 1.1498229503631592,
|
| 2755 |
+
"learning_rate": 1.6782727463097624e-05,
|
| 2756 |
+
"loss": 0.8533,
|
| 2757 |
+
"step": 388
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 0.8189473684210526,
|
| 2761 |
+
"grad_norm": 1.617012619972229,
|
| 2762 |
+
"learning_rate": 1.6409991904604173e-05,
|
| 2763 |
+
"loss": 0.6844,
|
| 2764 |
+
"step": 389
|
| 2765 |
+
},
|
| 2766 |
+
{
|
| 2767 |
+
"epoch": 0.8210526315789474,
|
| 2768 |
+
"grad_norm": 0.642407238483429,
|
| 2769 |
+
"learning_rate": 1.60410718030361e-05,
|
| 2770 |
+
"loss": 0.7084,
|
| 2771 |
+
"step": 390
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 0.8231578947368421,
|
| 2775 |
+
"grad_norm": 0.6749714016914368,
|
| 2776 |
+
"learning_rate": 1.5675983997711795e-05,
|
| 2777 |
+
"loss": 1.1988,
|
| 2778 |
+
"step": 391
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 0.8252631578947368,
|
| 2782 |
+
"grad_norm": 1.0748071670532227,
|
| 2783 |
+
"learning_rate": 1.5314745153024766e-05,
|
| 2784 |
+
"loss": 0.856,
|
| 2785 |
+
"step": 392
|
| 2786 |
+
},
|
| 2787 |
+
{
|
| 2788 |
+
"epoch": 0.8273684210526315,
|
| 2789 |
+
"grad_norm": 1.0286914110183716,
|
| 2790 |
+
"learning_rate": 1.495737175768326e-05,
|
| 2791 |
+
"loss": 1.1481,
|
| 2792 |
+
"step": 393
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 0.8294736842105264,
|
| 2796 |
+
"grad_norm": 0.8149781227111816,
|
| 2797 |
+
"learning_rate": 1.4603880123957447e-05,
|
| 2798 |
+
"loss": 0.7169,
|
| 2799 |
+
"step": 394
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 0.8315789473684211,
|
| 2803 |
+
"grad_norm": 1.1764047145843506,
|
| 2804 |
+
"learning_rate": 1.425428638693489e-05,
|
| 2805 |
+
"loss": 1.014,
|
| 2806 |
+
"step": 395
|
| 2807 |
+
},
|
| 2808 |
+
{
|
| 2809 |
+
"epoch": 0.8336842105263158,
|
| 2810 |
+
"grad_norm": 1.1029857397079468,
|
| 2811 |
+
"learning_rate": 1.3908606503784139e-05,
|
| 2812 |
+
"loss": 1.2927,
|
| 2813 |
+
"step": 396
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 0.8357894736842105,
|
| 2817 |
+
"grad_norm": 1.063981056213379,
|
| 2818 |
+
"learning_rate": 1.356685625302625e-05,
|
| 2819 |
+
"loss": 0.5832,
|
| 2820 |
+
"step": 397
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 0.8378947368421052,
|
| 2824 |
+
"grad_norm": 0.8528422117233276,
|
| 2825 |
+
"learning_rate": 1.3229051233814637e-05,
|
| 2826 |
+
"loss": 0.8959,
|
| 2827 |
+
"step": 398
|
| 2828 |
+
},
|
| 2829 |
+
{
|
| 2830 |
+
"epoch": 0.84,
|
| 2831 |
+
"grad_norm": 0.8783009052276611,
|
| 2832 |
+
"learning_rate": 1.2895206865223064e-05,
|
| 2833 |
+
"loss": 0.8633,
|
| 2834 |
+
"step": 399
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 0.8421052631578947,
|
| 2838 |
+
"grad_norm": 1.0344438552856445,
|
| 2839 |
+
"learning_rate": 1.2565338385541792e-05,
|
| 2840 |
+
"loss": 1.7439,
|
| 2841 |
+
"step": 400
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 0.8442105263157895,
|
| 2845 |
+
"grad_norm": 0.9089331030845642,
|
| 2846 |
+
"learning_rate": 1.2239460851582118e-05,
|
| 2847 |
+
"loss": 0.9778,
|
| 2848 |
+
"step": 401
|
| 2849 |
+
},
|
| 2850 |
+
{
|
| 2851 |
+
"epoch": 0.8463157894736842,
|
| 2852 |
+
"grad_norm": 1.5010464191436768,
|
| 2853 |
+
"learning_rate": 1.1917589137989005e-05,
|
| 2854 |
+
"loss": 0.8401,
|
| 2855 |
+
"step": 402
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 0.848421052631579,
|
| 2859 |
+
"grad_norm": 1.1659886837005615,
|
| 2860 |
+
"learning_rate": 1.1599737936562149e-05,
|
| 2861 |
+
"loss": 1.1214,
|
| 2862 |
+
"step": 403
|
| 2863 |
+
},
|
| 2864 |
+
{
|
| 2865 |
+
"epoch": 0.8505263157894737,
|
| 2866 |
+
"grad_norm": 0.781868040561676,
|
| 2867 |
+
"learning_rate": 1.1285921755585504e-05,
|
| 2868 |
+
"loss": 1.1831,
|
| 2869 |
+
"step": 404
|
| 2870 |
+
},
|
| 2871 |
+
{
|
| 2872 |
+
"epoch": 0.8526315789473684,
|
| 2873 |
+
"grad_norm": 0.8510306477546692,
|
| 2874 |
+
"learning_rate": 1.097615491916485e-05,
|
| 2875 |
+
"loss": 1.0715,
|
| 2876 |
+
"step": 405
|
| 2877 |
+
},
|
| 2878 |
+
{
|
| 2879 |
+
"epoch": 0.8547368421052631,
|
| 2880 |
+
"grad_norm": 0.7362077236175537,
|
| 2881 |
+
"learning_rate": 1.0670451566574102e-05,
|
| 2882 |
+
"loss": 0.9367,
|
| 2883 |
+
"step": 406
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 0.8568421052631578,
|
| 2887 |
+
"grad_norm": 1.3995620012283325,
|
| 2888 |
+
"learning_rate": 1.0368825651609893e-05,
|
| 2889 |
+
"loss": 0.8952,
|
| 2890 |
+
"step": 407
|
| 2891 |
+
},
|
| 2892 |
+
{
|
| 2893 |
+
"epoch": 0.8589473684210527,
|
| 2894 |
+
"grad_norm": 0.8937683701515198,
|
| 2895 |
+
"learning_rate": 1.007129094195468e-05,
|
| 2896 |
+
"loss": 1.2489,
|
| 2897 |
+
"step": 408
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 0.8610526315789474,
|
| 2901 |
+
"grad_norm": 2.044285774230957,
|
| 2902 |
+
"learning_rate": 9.777861018548251e-06,
|
| 2903 |
+
"loss": 1.658,
|
| 2904 |
+
"step": 409
|
| 2905 |
+
},
|
| 2906 |
+
{
|
| 2907 |
+
"epoch": 0.8631578947368421,
|
| 2908 |
+
"grad_norm": 1.3045108318328857,
|
| 2909 |
+
"learning_rate": 9.488549274967872e-06,
|
| 2910 |
+
"loss": 0.8359,
|
| 2911 |
+
"step": 410
|
| 2912 |
+
},
|
| 2913 |
+
{
|
| 2914 |
+
"epoch": 0.8652631578947368,
|
| 2915 |
+
"grad_norm": 0.6791781783103943,
|
| 2916 |
+
"learning_rate": 9.203368916817012e-06,
|
| 2917 |
+
"loss": 1.0212,
|
| 2918 |
+
"step": 411
|
| 2919 |
+
},
|
| 2920 |
+
{
|
| 2921 |
+
"epoch": 0.8673684210526316,
|
| 2922 |
+
"grad_norm": 1.7533940076828003,
|
| 2923 |
+
"learning_rate": 8.92233296112236e-06,
|
| 2924 |
+
"loss": 0.7901,
|
| 2925 |
+
"step": 412
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 0.8694736842105263,
|
| 2929 |
+
"grad_norm": 0.5513384938240051,
|
| 2930 |
+
"learning_rate": 8.645454235739903e-06,
|
| 2931 |
+
"loss": 0.8964,
|
| 2932 |
+
"step": 413
|
| 2933 |
+
},
|
| 2934 |
+
{
|
| 2935 |
+
"epoch": 0.871578947368421,
|
| 2936 |
+
"grad_norm": 1.2993557453155518,
|
| 2937 |
+
"learning_rate": 8.372745378769309e-06,
|
| 2938 |
+
"loss": 0.9348,
|
| 2939 |
+
"step": 414
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 0.8736842105263158,
|
| 2943 |
+
"grad_norm": 1.2195082902908325,
|
| 2944 |
+
"learning_rate": 8.10421883797694e-06,
|
| 2945 |
+
"loss": 1.2758,
|
| 2946 |
+
"step": 415
|
| 2947 |
+
},
|
| 2948 |
+
{
|
| 2949 |
+
"epoch": 0.8757894736842106,
|
| 2950 |
+
"grad_norm": 1.2428478002548218,
|
| 2951 |
+
"learning_rate": 7.839886870227909e-06,
|
| 2952 |
+
"loss": 0.9534,
|
| 2953 |
+
"step": 416
|
| 2954 |
+
},
|
| 2955 |
+
{
|
| 2956 |
+
"epoch": 0.8778947368421053,
|
| 2957 |
+
"grad_norm": 0.5317150950431824,
|
| 2958 |
+
"learning_rate": 7.5797615409264335e-06,
|
| 2959 |
+
"loss": 0.7835,
|
| 2960 |
+
"step": 417
|
| 2961 |
+
},
|
| 2962 |
+
{
|
| 2963 |
+
"epoch": 0.88,
|
| 2964 |
+
"grad_norm": 0.7354329228401184,
|
| 2965 |
+
"learning_rate": 7.32385472346514e-06,
|
| 2966 |
+
"loss": 1.1074,
|
| 2967 |
+
"step": 418
|
| 2968 |
+
},
|
| 2969 |
+
{
|
| 2970 |
+
"epoch": 0.8821052631578947,
|
| 2971 |
+
"grad_norm": 1.009934663772583,
|
| 2972 |
+
"learning_rate": 7.072178098683246e-06,
|
| 2973 |
+
"loss": 1.253,
|
| 2974 |
+
"step": 419
|
| 2975 |
+
},
|
| 2976 |
+
{
|
| 2977 |
+
"epoch": 0.8842105263157894,
|
| 2978 |
+
"grad_norm": 0.8174380660057068,
|
| 2979 |
+
"learning_rate": 6.824743154333157e-06,
|
| 2980 |
+
"loss": 1.0671,
|
| 2981 |
+
"step": 420
|
| 2982 |
+
},
|
| 2983 |
+
{
|
| 2984 |
+
"epoch": 0.8863157894736842,
|
| 2985 |
+
"grad_norm": 0.777045726776123,
|
| 2986 |
+
"learning_rate": 6.581561184556295e-06,
|
| 2987 |
+
"loss": 1.0806,
|
| 2988 |
+
"step": 421
|
| 2989 |
+
},
|
| 2990 |
+
{
|
| 2991 |
+
"epoch": 0.888421052631579,
|
| 2992 |
+
"grad_norm": 1.013598918914795,
|
| 2993 |
+
"learning_rate": 6.342643289367522e-06,
|
| 2994 |
+
"loss": 1.1415,
|
| 2995 |
+
"step": 422
|
| 2996 |
+
},
|
| 2997 |
+
{
|
| 2998 |
+
"epoch": 0.8905263157894737,
|
| 2999 |
+
"grad_norm": 0.9286550283432007,
|
| 3000 |
+
"learning_rate": 6.108000374148448e-06,
|
| 3001 |
+
"loss": 0.9374,
|
| 3002 |
+
"step": 423
|
| 3003 |
+
},
|
| 3004 |
+
{
|
| 3005 |
+
"epoch": 0.8926315789473684,
|
| 3006 |
+
"grad_norm": 0.8254793286323547,
|
| 3007 |
+
"learning_rate": 5.87764314914967e-06,
|
| 3008 |
+
"loss": 0.8427,
|
| 3009 |
+
"step": 424
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 0.8947368421052632,
|
| 3013 |
+
"grad_norm": 1.3116326332092285,
|
| 3014 |
+
"learning_rate": 5.651582129001986e-06,
|
| 3015 |
+
"loss": 1.2204,
|
| 3016 |
+
"step": 425
|
| 3017 |
+
},
|
| 3018 |
+
{
|
| 3019 |
+
"epoch": 0.8968421052631579,
|
| 3020 |
+
"grad_norm": 0.5832639932632446,
|
| 3021 |
+
"learning_rate": 5.429827632236284e-06,
|
| 3022 |
+
"loss": 0.7114,
|
| 3023 |
+
"step": 426
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 0.8989473684210526,
|
| 3027 |
+
"grad_norm": 0.719308078289032,
|
| 3028 |
+
"learning_rate": 5.212389780812732e-06,
|
| 3029 |
+
"loss": 0.8413,
|
| 3030 |
+
"step": 427
|
| 3031 |
+
},
|
| 3032 |
+
{
|
| 3033 |
+
"epoch": 0.9010526315789473,
|
| 3034 |
+
"grad_norm": 0.8241797089576721,
|
| 3035 |
+
"learning_rate": 4.999278499658666e-06,
|
| 3036 |
+
"loss": 0.9538,
|
| 3037 |
+
"step": 428
|
| 3038 |
+
},
|
| 3039 |
+
{
|
| 3040 |
+
"epoch": 0.9031578947368422,
|
| 3041 |
+
"grad_norm": 0.68552565574646,
|
| 3042 |
+
"learning_rate": 4.790503516215572e-06,
|
| 3043 |
+
"loss": 1.0962,
|
| 3044 |
+
"step": 429
|
| 3045 |
+
},
|
| 3046 |
+
{
|
| 3047 |
+
"epoch": 0.9052631578947369,
|
| 3048 |
+
"grad_norm": 1.2444804906845093,
|
| 3049 |
+
"learning_rate": 4.586074359995119e-06,
|
| 3050 |
+
"loss": 1.2615,
|
| 3051 |
+
"step": 430
|
| 3052 |
+
},
|
| 3053 |
+
{
|
| 3054 |
+
"epoch": 0.9073684210526316,
|
| 3055 |
+
"grad_norm": 1.2105733156204224,
|
| 3056 |
+
"learning_rate": 4.386000362144138e-06,
|
| 3057 |
+
"loss": 1.1771,
|
| 3058 |
+
"step": 431
|
| 3059 |
+
},
|
| 3060 |
+
{
|
| 3061 |
+
"epoch": 0.9094736842105263,
|
| 3062 |
+
"grad_norm": 0.7618857026100159,
|
| 3063 |
+
"learning_rate": 4.190290655018736e-06,
|
| 3064 |
+
"loss": 1.0337,
|
| 3065 |
+
"step": 432
|
| 3066 |
+
},
|
| 3067 |
+
{
|
| 3068 |
+
"epoch": 0.911578947368421,
|
| 3069 |
+
"grad_norm": 0.6650370955467224,
|
| 3070 |
+
"learning_rate": 3.998954171767422e-06,
|
| 3071 |
+
"loss": 1.0167,
|
| 3072 |
+
"step": 433
|
| 3073 |
+
},
|
| 3074 |
+
{
|
| 3075 |
+
"epoch": 0.9136842105263158,
|
| 3076 |
+
"grad_norm": 0.8902232050895691,
|
| 3077 |
+
"learning_rate": 3.811999645923414e-06,
|
| 3078 |
+
"loss": 1.0959,
|
| 3079 |
+
"step": 434
|
| 3080 |
+
},
|
| 3081 |
+
{
|
| 3082 |
+
"epoch": 0.9157894736842105,
|
| 3083 |
+
"grad_norm": 0.8037033081054688,
|
| 3084 |
+
"learning_rate": 3.6294356110059157e-06,
|
| 3085 |
+
"loss": 1.0283,
|
| 3086 |
+
"step": 435
|
| 3087 |
+
},
|
| 3088 |
+
{
|
| 3089 |
+
"epoch": 0.9178947368421052,
|
| 3090 |
+
"grad_norm": 0.7740287780761719,
|
| 3091 |
+
"learning_rate": 3.451270400130646e-06,
|
| 3092 |
+
"loss": 0.9952,
|
| 3093 |
+
"step": 436
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 0.92,
|
| 3097 |
+
"grad_norm": 0.6354364156723022,
|
| 3098 |
+
"learning_rate": 3.277512145629502e-06,
|
| 3099 |
+
"loss": 0.7446,
|
| 3100 |
+
"step": 437
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 0.9221052631578948,
|
| 3104 |
+
"grad_norm": 1.0572397708892822,
|
| 3105 |
+
"learning_rate": 3.10816877867931e-06,
|
| 3106 |
+
"loss": 1.0779,
|
| 3107 |
+
"step": 438
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"epoch": 0.9242105263157895,
|
| 3111 |
+
"grad_norm": 1.2464516162872314,
|
| 3112 |
+
"learning_rate": 2.943248028939838e-06,
|
| 3113 |
+
"loss": 1.1318,
|
| 3114 |
+
"step": 439
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 0.9263157894736842,
|
| 3118 |
+
"grad_norm": 1.017857313156128,
|
| 3119 |
+
"learning_rate": 2.7827574242009437e-06,
|
| 3120 |
+
"loss": 0.7569,
|
| 3121 |
+
"step": 440
|
| 3122 |
+
},
|
| 3123 |
+
{
|
| 3124 |
+
"epoch": 0.9284210526315789,
|
| 3125 |
+
"grad_norm": 0.8179978132247925,
|
| 3126 |
+
"learning_rate": 2.626704290039017e-06,
|
| 3127 |
+
"loss": 0.9221,
|
| 3128 |
+
"step": 441
|
| 3129 |
+
},
|
| 3130 |
+
{
|
| 3131 |
+
"epoch": 0.9305263157894736,
|
| 3132 |
+
"grad_norm": 0.7406687140464783,
|
| 3133 |
+
"learning_rate": 2.4750957494826033e-06,
|
| 3134 |
+
"loss": 0.9411,
|
| 3135 |
+
"step": 442
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 0.9326315789473684,
|
| 3139 |
+
"grad_norm": 0.7985221147537231,
|
| 3140 |
+
"learning_rate": 2.327938722687184e-06,
|
| 3141 |
+
"loss": 0.5579,
|
| 3142 |
+
"step": 443
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 0.9347368421052632,
|
| 3146 |
+
"grad_norm": 0.8860235214233398,
|
| 3147 |
+
"learning_rate": 2.1852399266194314e-06,
|
| 3148 |
+
"loss": 1.2353,
|
| 3149 |
+
"step": 444
|
| 3150 |
+
},
|
| 3151 |
+
{
|
| 3152 |
+
"epoch": 0.9368421052631579,
|
| 3153 |
+
"grad_norm": 0.8313080072402954,
|
| 3154 |
+
"learning_rate": 2.0470058747505516e-06,
|
| 3155 |
+
"loss": 0.9344,
|
| 3156 |
+
"step": 445
|
| 3157 |
+
},
|
| 3158 |
+
{
|
| 3159 |
+
"epoch": 0.9389473684210526,
|
| 3160 |
+
"grad_norm": 3.333054780960083,
|
| 3161 |
+
"learning_rate": 1.9132428767589473e-06,
|
| 3162 |
+
"loss": 1.4123,
|
| 3163 |
+
"step": 446
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 0.9410526315789474,
|
| 3167 |
+
"grad_norm": 1.0423784255981445,
|
| 3168 |
+
"learning_rate": 1.7839570382422787e-06,
|
| 3169 |
+
"loss": 1.0753,
|
| 3170 |
+
"step": 447
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 0.9431578947368421,
|
| 3174 |
+
"grad_norm": 0.7692586183547974,
|
| 3175 |
+
"learning_rate": 1.6591542604387445e-06,
|
| 3176 |
+
"loss": 0.669,
|
| 3177 |
+
"step": 448
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 0.9452631578947368,
|
| 3181 |
+
"grad_norm": 0.7186653017997742,
|
| 3182 |
+
"learning_rate": 1.538840239957684e-06,
|
| 3183 |
+
"loss": 1.2113,
|
| 3184 |
+
"step": 449
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 0.9473684210526315,
|
| 3188 |
+
"grad_norm": 0.7315964698791504,
|
| 3189 |
+
"learning_rate": 1.4230204685196203e-06,
|
| 3190 |
+
"loss": 1.2322,
|
| 3191 |
+
"step": 450
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 0.9494736842105264,
|
| 3195 |
+
"grad_norm": 1.0726860761642456,
|
| 3196 |
+
"learning_rate": 1.3117002327055927e-06,
|
| 3197 |
+
"loss": 0.9252,
|
| 3198 |
+
"step": 451
|
| 3199 |
+
},
|
| 3200 |
+
{
|
| 3201 |
+
"epoch": 0.9515789473684211,
|
| 3202 |
+
"grad_norm": 0.6930716037750244,
|
| 3203 |
+
"learning_rate": 1.20488461371574e-06,
|
| 3204 |
+
"loss": 0.9903,
|
| 3205 |
+
"step": 452
|
| 3206 |
+
},
|
| 3207 |
+
{
|
| 3208 |
+
"epoch": 0.9536842105263158,
|
| 3209 |
+
"grad_norm": 1.1421600580215454,
|
| 3210 |
+
"learning_rate": 1.102578487137529e-06,
|
| 3211 |
+
"loss": 1.0869,
|
| 3212 |
+
"step": 453
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 0.9557894736842105,
|
| 3216 |
+
"grad_norm": 0.6389675736427307,
|
| 3217 |
+
"learning_rate": 1.004786522723089e-06,
|
| 3218 |
+
"loss": 0.7617,
|
| 3219 |
+
"step": 454
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 0.9578947368421052,
|
| 3223 |
+
"grad_norm": 1.449903964996338,
|
| 3224 |
+
"learning_rate": 9.11513184176116e-07,
|
| 3225 |
+
"loss": 1.0412,
|
| 3226 |
+
"step": 455
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 0.96,
|
| 3230 |
+
"grad_norm": 0.7298595905303955,
|
| 3231 |
+
"learning_rate": 8.227627289481121e-07,
|
| 3232 |
+
"loss": 0.9267,
|
| 3233 |
+
"step": 456
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 0.9621052631578947,
|
| 3237 |
+
"grad_norm": 0.988619327545166,
|
| 3238 |
+
"learning_rate": 7.385392080440534e-07,
|
| 3239 |
+
"loss": 0.9785,
|
| 3240 |
+
"step": 457
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 0.9642105263157895,
|
| 3244 |
+
"grad_norm": 1.0173723697662354,
|
| 3245 |
+
"learning_rate": 6.588464658374815e-07,
|
| 3246 |
+
"loss": 1.4313,
|
| 3247 |
+
"step": 458
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 0.9663157894736842,
|
| 3251 |
+
"grad_norm": 0.8676223158836365,
|
| 3252 |
+
"learning_rate": 5.836881398950667e-07,
|
| 3253 |
+
"loss": 1.1253,
|
| 3254 |
+
"step": 459
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 0.968421052631579,
|
| 3258 |
+
"grad_norm": 2.1744537353515625,
|
| 3259 |
+
"learning_rate": 5.130676608104845e-07,
|
| 3260 |
+
"loss": 0.9635,
|
| 3261 |
+
"step": 460
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 0.9705263157894737,
|
| 3265 |
+
"grad_norm": 0.9125698804855347,
|
| 3266 |
+
"learning_rate": 4.469882520479196e-07,
|
| 3267 |
+
"loss": 1.2048,
|
| 3268 |
+
"step": 461
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 0.9726315789473684,
|
| 3272 |
+
"grad_norm": 0.8515905141830444,
|
| 3273 |
+
"learning_rate": 3.8545292979486057e-07,
|
| 3274 |
+
"loss": 0.9358,
|
| 3275 |
+
"step": 462
|
| 3276 |
+
},
|
| 3277 |
+
{
|
| 3278 |
+
"epoch": 0.9747368421052631,
|
| 3279 |
+
"grad_norm": 0.8601663708686829,
|
| 3280 |
+
"learning_rate": 3.2846450282447703e-07,
|
| 3281 |
+
"loss": 1.0253,
|
| 3282 |
+
"step": 463
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 0.9768421052631578,
|
| 3286 |
+
"grad_norm": 0.8213205337524414,
|
| 3287 |
+
"learning_rate": 2.760255723673888e-07,
|
| 3288 |
+
"loss": 0.8577,
|
| 3289 |
+
"step": 464
|
| 3290 |
+
},
|
| 3291 |
+
{
|
| 3292 |
+
"epoch": 0.9789473684210527,
|
| 3293 |
+
"grad_norm": 0.7541871666908264,
|
| 3294 |
+
"learning_rate": 2.2813853199292746e-07,
|
| 3295 |
+
"loss": 0.9517,
|
| 3296 |
+
"step": 465
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 0.9810526315789474,
|
| 3300 |
+
"grad_norm": 0.7226133346557617,
|
| 3301 |
+
"learning_rate": 1.8480556749991274e-07,
|
| 3302 |
+
"loss": 1.0434,
|
| 3303 |
+
"step": 466
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 0.9831578947368421,
|
| 3307 |
+
"grad_norm": 1.1320992708206177,
|
| 3308 |
+
"learning_rate": 1.460286568168212e-07,
|
| 3309 |
+
"loss": 0.7791,
|
| 3310 |
+
"step": 467
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 0.9852631578947368,
|
| 3314 |
+
"grad_norm": 0.9081979990005493,
|
| 3315 |
+
"learning_rate": 1.1180956991160286e-07,
|
| 3316 |
+
"loss": 0.3705,
|
| 3317 |
+
"step": 468
|
| 3318 |
+
},
|
| 3319 |
+
{
|
| 3320 |
+
"epoch": 0.9873684210526316,
|
| 3321 |
+
"grad_norm": 2.7810940742492676,
|
| 3322 |
+
"learning_rate": 8.214986871076802e-08,
|
| 3323 |
+
"loss": 1.2752,
|
| 3324 |
+
"step": 469
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 0.9894736842105263,
|
| 3328 |
+
"grad_norm": 0.7669661045074463,
|
| 3329 |
+
"learning_rate": 5.705090702819993e-08,
|
| 3330 |
+
"loss": 0.7446,
|
| 3331 |
+
"step": 470
|
| 3332 |
+
},
|
| 3333 |
+
{
|
| 3334 |
+
"epoch": 0.991578947368421,
|
| 3335 |
+
"grad_norm": 1.0128413438796997,
|
| 3336 |
+
"learning_rate": 3.6513830503293045e-08,
|
| 3337 |
+
"loss": 1.2359,
|
| 3338 |
+
"step": 471
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 0.9936842105263158,
|
| 3342 |
+
"grad_norm": 1.728697657585144,
|
| 3343 |
+
"learning_rate": 2.0539576548717076e-08,
|
| 3344 |
+
"loss": 0.7628,
|
| 3345 |
+
"step": 472
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 0.9957894736842106,
|
| 3349 |
+
"grad_norm": 0.7068576812744141,
|
| 3350 |
+
"learning_rate": 9.128874307551272e-09,
|
| 3351 |
+
"loss": 1.063,
|
| 3352 |
+
"step": 473
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 0.9978947368421053,
|
| 3356 |
+
"grad_norm": 0.7579674124717712,
|
| 3357 |
+
"learning_rate": 2.282244620088747e-09,
|
| 3358 |
+
"loss": 0.8835,
|
| 3359 |
+
"step": 474
|
| 3360 |
+
},
|
| 3361 |
+
{
|
| 3362 |
+
"epoch": 1.0,
|
| 3363 |
+
"grad_norm": 0.7143834829330444,
|
| 3364 |
+
"learning_rate": 0.0,
|
| 3365 |
+
"loss": 1.1322,
|
| 3366 |
+
"step": 475
|
| 3367 |
}
|
| 3368 |
],
|
| 3369 |
"logging_steps": 1,
|
|
|
|
| 3378 |
"should_evaluate": false,
|
| 3379 |
"should_log": false,
|
| 3380 |
"should_save": true,
|
| 3381 |
+
"should_training_stop": true
|
| 3382 |
},
|
| 3383 |
"attributes": {}
|
| 3384 |
}
|
| 3385 |
},
|
| 3386 |
+
"total_flos": 4.425821325990298e+16,
|
| 3387 |
"train_batch_size": 1,
|
| 3388 |
"trial_name": null,
|
| 3389 |
"trial_params": null
|