Commit
·
f18be93
1
Parent(s):
46b970d
Training in progress, step 450000
Browse files- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +311 -3
- pytorch_model.bin +1 -1
- runs/Jan25_00-38-33_t1v-n-15e54913-w-0/events.out.tfevents.1674607146.t1v-n-15e54913-w-0.589238.0 +2 -2
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2693742553
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1b43161845c8e6bda7b95036487c6c5183ff73534909a2f952ed1b54df93899
|
| 3 |
size 2693742553
|
last-checkpoint/pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1346893675
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9c8b2494cce6ffb8f8fa2a74583f549bfdaa9d5b42d682399d2b1885ce2b565
|
| 3 |
size 1346893675
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13611
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e01afc5629600fbeb0ae6d44b6a9d9e243c9b03ab21b6e5cd16e2d2d04f793
|
| 3 |
size 13611
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78e735efa7e40e0dd22dcac5cb3724b0cbe120563d603ea4b62f22b0f40fc602
|
| 3 |
size 623
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 3.
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -2470,11 +2470,319 @@
|
|
| 2470 |
"eval_samples_per_second": 257.285,
|
| 2471 |
"eval_steps_per_second": 4.065,
|
| 2472 |
"step": 400000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2473 |
}
|
| 2474 |
],
|
| 2475 |
"max_steps": 1000000,
|
| 2476 |
"num_train_epochs": 9223372036854775807,
|
| 2477 |
-
"total_flos": 2.
|
| 2478 |
"trial_name": null,
|
| 2479 |
"trial_params": null
|
| 2480 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.106896,
|
| 5 |
+
"global_step": 450000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 2470 |
"eval_samples_per_second": 257.285,
|
| 2471 |
"eval_steps_per_second": 4.065,
|
| 2472 |
"step": 400000
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 3.06,
|
| 2476 |
+
"learning_rate": 6.993324133116726e-05,
|
| 2477 |
+
"loss": 0.6153,
|
| 2478 |
+
"step": 401000
|
| 2479 |
+
},
|
| 2480 |
+
{
|
| 2481 |
+
"epoch": 3.06,
|
| 2482 |
+
"learning_rate": 6.978149344295242e-05,
|
| 2483 |
+
"loss": 0.5254,
|
| 2484 |
+
"step": 402000
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 3.06,
|
| 2488 |
+
"learning_rate": 6.962952922749457e-05,
|
| 2489 |
+
"loss": 0.5045,
|
| 2490 |
+
"step": 403000
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 3.06,
|
| 2494 |
+
"learning_rate": 6.947735034665002e-05,
|
| 2495 |
+
"loss": 0.601,
|
| 2496 |
+
"step": 404000
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 3.06,
|
| 2500 |
+
"learning_rate": 6.932495846462261e-05,
|
| 2501 |
+
"loss": 0.6463,
|
| 2502 |
+
"step": 405000
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 3.06,
|
| 2506 |
+
"learning_rate": 6.917235524794558e-05,
|
| 2507 |
+
"loss": 0.6586,
|
| 2508 |
+
"step": 406000
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 3.06,
|
| 2512 |
+
"learning_rate": 6.901954236546323e-05,
|
| 2513 |
+
"loss": 0.5835,
|
| 2514 |
+
"step": 407000
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 3.06,
|
| 2518 |
+
"learning_rate": 6.886652148831279e-05,
|
| 2519 |
+
"loss": 0.5539,
|
| 2520 |
+
"step": 408000
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 3.07,
|
| 2524 |
+
"learning_rate": 6.871329428990602e-05,
|
| 2525 |
+
"loss": 0.5354,
|
| 2526 |
+
"step": 409000
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 3.07,
|
| 2530 |
+
"learning_rate": 6.855986244591104e-05,
|
| 2531 |
+
"loss": 0.6341,
|
| 2532 |
+
"step": 410000
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 3.07,
|
| 2536 |
+
"learning_rate": 6.840622763423391e-05,
|
| 2537 |
+
"loss": 0.6923,
|
| 2538 |
+
"step": 411000
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 3.07,
|
| 2542 |
+
"learning_rate": 6.825239153500029e-05,
|
| 2543 |
+
"loss": 0.6159,
|
| 2544 |
+
"step": 412000
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 3.07,
|
| 2548 |
+
"learning_rate": 6.809835583053715e-05,
|
| 2549 |
+
"loss": 0.5116,
|
| 2550 |
+
"step": 413000
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 3.07,
|
| 2554 |
+
"learning_rate": 6.794412220535426e-05,
|
| 2555 |
+
"loss": 0.5119,
|
| 2556 |
+
"step": 414000
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 3.07,
|
| 2560 |
+
"learning_rate": 6.778969234612584e-05,
|
| 2561 |
+
"loss": 0.5906,
|
| 2562 |
+
"step": 415000
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 3.07,
|
| 2566 |
+
"learning_rate": 6.763506794167208e-05,
|
| 2567 |
+
"loss": 0.6402,
|
| 2568 |
+
"step": 416000
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 3.07,
|
| 2572 |
+
"learning_rate": 6.748025068294067e-05,
|
| 2573 |
+
"loss": 0.6549,
|
| 2574 |
+
"step": 417000
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 3.07,
|
| 2578 |
+
"learning_rate": 6.732524226298841e-05,
|
| 2579 |
+
"loss": 0.5867,
|
| 2580 |
+
"step": 418000
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 3.08,
|
| 2584 |
+
"learning_rate": 6.71700443769625e-05,
|
| 2585 |
+
"loss": 0.559,
|
| 2586 |
+
"step": 419000
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 3.08,
|
| 2590 |
+
"learning_rate": 6.701465872208216e-05,
|
| 2591 |
+
"loss": 0.5155,
|
| 2592 |
+
"step": 420000
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 3.08,
|
| 2596 |
+
"learning_rate": 6.685908699762002e-05,
|
| 2597 |
+
"loss": 0.6399,
|
| 2598 |
+
"step": 421000
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 3.08,
|
| 2602 |
+
"learning_rate": 6.670333090488356e-05,
|
| 2603 |
+
"loss": 0.6847,
|
| 2604 |
+
"step": 422000
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 3.08,
|
| 2608 |
+
"learning_rate": 6.654739214719641e-05,
|
| 2609 |
+
"loss": 0.6183,
|
| 2610 |
+
"step": 423000
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 3.08,
|
| 2614 |
+
"learning_rate": 6.639127242987988e-05,
|
| 2615 |
+
"loss": 0.5072,
|
| 2616 |
+
"step": 424000
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 3.08,
|
| 2620 |
+
"learning_rate": 6.623497346023418e-05,
|
| 2621 |
+
"loss": 0.513,
|
| 2622 |
+
"step": 425000
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 3.08,
|
| 2626 |
+
"learning_rate": 6.607849694751977e-05,
|
| 2627 |
+
"loss": 0.5793,
|
| 2628 |
+
"step": 426000
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 3.08,
|
| 2632 |
+
"learning_rate": 6.592184460293877e-05,
|
| 2633 |
+
"loss": 0.6379,
|
| 2634 |
+
"step": 427000
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 3.08,
|
| 2638 |
+
"learning_rate": 6.576501813961609e-05,
|
| 2639 |
+
"loss": 0.6484,
|
| 2640 |
+
"step": 428000
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 3.09,
|
| 2644 |
+
"learning_rate": 6.56080192725808e-05,
|
| 2645 |
+
"loss": 0.5893,
|
| 2646 |
+
"step": 429000
|
| 2647 |
+
},
|
| 2648 |
+
{
|
| 2649 |
+
"epoch": 3.09,
|
| 2650 |
+
"learning_rate": 6.545084971874738e-05,
|
| 2651 |
+
"loss": 0.5657,
|
| 2652 |
+
"step": 430000
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 3.09,
|
| 2656 |
+
"learning_rate": 6.529351119689688e-05,
|
| 2657 |
+
"loss": 0.5003,
|
| 2658 |
+
"step": 431000
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 3.09,
|
| 2662 |
+
"learning_rate": 6.513600542765817e-05,
|
| 2663 |
+
"loss": 0.6413,
|
| 2664 |
+
"step": 432000
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 3.09,
|
| 2668 |
+
"learning_rate": 6.497833413348909e-05,
|
| 2669 |
+
"loss": 0.6753,
|
| 2670 |
+
"step": 433000
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 3.09,
|
| 2674 |
+
"learning_rate": 6.48204990386577e-05,
|
| 2675 |
+
"loss": 0.6138,
|
| 2676 |
+
"step": 434000
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 3.09,
|
| 2680 |
+
"learning_rate": 6.466250186922325e-05,
|
| 2681 |
+
"loss": 0.5166,
|
| 2682 |
+
"step": 435000
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 3.09,
|
| 2686 |
+
"learning_rate": 6.450434435301751e-05,
|
| 2687 |
+
"loss": 0.5174,
|
| 2688 |
+
"step": 436000
|
| 2689 |
+
},
|
| 2690 |
+
{
|
| 2691 |
+
"epoch": 3.09,
|
| 2692 |
+
"learning_rate": 6.43460282196257e-05,
|
| 2693 |
+
"loss": 0.5669,
|
| 2694 |
+
"step": 437000
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 3.09,
|
| 2698 |
+
"learning_rate": 6.418755520036775e-05,
|
| 2699 |
+
"loss": 0.634,
|
| 2700 |
+
"step": 438000
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 3.1,
|
| 2704 |
+
"learning_rate": 6.402892702827916e-05,
|
| 2705 |
+
"loss": 0.6451,
|
| 2706 |
+
"step": 439000
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 3.1,
|
| 2710 |
+
"learning_rate": 6.387014543809223e-05,
|
| 2711 |
+
"loss": 0.596,
|
| 2712 |
+
"step": 440000
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 3.1,
|
| 2716 |
+
"learning_rate": 6.371121216621698e-05,
|
| 2717 |
+
"loss": 0.5627,
|
| 2718 |
+
"step": 441000
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 3.1,
|
| 2722 |
+
"learning_rate": 6.355212895072223e-05,
|
| 2723 |
+
"loss": 0.5008,
|
| 2724 |
+
"step": 442000
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 3.1,
|
| 2728 |
+
"learning_rate": 6.339289753131649e-05,
|
| 2729 |
+
"loss": 0.6418,
|
| 2730 |
+
"step": 443000
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 3.1,
|
| 2734 |
+
"learning_rate": 6.323351964932908e-05,
|
| 2735 |
+
"loss": 0.6592,
|
| 2736 |
+
"step": 444000
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 3.1,
|
| 2740 |
+
"learning_rate": 6.307399704769099e-05,
|
| 2741 |
+
"loss": 0.6128,
|
| 2742 |
+
"step": 445000
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 3.1,
|
| 2746 |
+
"learning_rate": 6.291433147091583e-05,
|
| 2747 |
+
"loss": 0.521,
|
| 2748 |
+
"step": 446000
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 3.1,
|
| 2752 |
+
"learning_rate": 6.275452466508077e-05,
|
| 2753 |
+
"loss": 0.5179,
|
| 2754 |
+
"step": 447000
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 3.1,
|
| 2758 |
+
"learning_rate": 6.259457837780742e-05,
|
| 2759 |
+
"loss": 0.5598,
|
| 2760 |
+
"step": 448000
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 3.11,
|
| 2764 |
+
"learning_rate": 6.243449435824276e-05,
|
| 2765 |
+
"loss": 0.631,
|
| 2766 |
+
"step": 449000
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 3.11,
|
| 2770 |
+
"learning_rate": 6.227427435703997e-05,
|
| 2771 |
+
"loss": 0.6381,
|
| 2772 |
+
"step": 450000
|
| 2773 |
+
},
|
| 2774 |
+
{
|
| 2775 |
+
"epoch": 3.11,
|
| 2776 |
+
"eval_loss": 0.3511766195297241,
|
| 2777 |
+
"eval_runtime": 18.2763,
|
| 2778 |
+
"eval_samples_per_second": 273.578,
|
| 2779 |
+
"eval_steps_per_second": 4.323,
|
| 2780 |
+
"step": 450000
|
| 2781 |
}
|
| 2782 |
],
|
| 2783 |
"max_steps": 1000000,
|
| 2784 |
"num_train_epochs": 9223372036854775807,
|
| 2785 |
+
"total_flos": 2.684246915457378e+19,
|
| 2786 |
"trial_name": null,
|
| 2787 |
"trial_params": null
|
| 2788 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1346893675
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9c8b2494cce6ffb8f8fa2a74583f549bfdaa9d5b42d682399d2b1885ce2b565
|
| 3 |
size 1346893675
|
runs/Jan25_00-38-33_t1v-n-15e54913-w-0/events.out.tfevents.1674607146.t1v-n-15e54913-w-0.589238.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88cd7e4595e4d61202d20a54a751c1b6fcd33f74c1aca86faa4ec2c43066de6a
|
| 3 |
+
size 53436
|