Muhammad Khalifa
commited on
Commit
•
ed42aba
1
Parent(s):
9e05026
add instrucode llama-2 models
Browse files- with_input/decomp_code_with_intermediates/llama-2-13b/best_model/adapter_model.bin +1 -1
- with_input/decomp_code_with_intermediates/llama-2-13b/best_model/optimizer.pt +1 -1
- with_input/decomp_code_with_intermediates/llama-2-13b/best_model/rng_state.pth +1 -1
- with_input/decomp_code_with_intermediates/llama-2-13b/best_model/scheduler.pt +1 -1
- with_input/decomp_code_with_intermediates/llama-2-13b/best_model/trainer_state.json +389 -5
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/adapter_config.json +21 -0
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/adapter_model.bin +3 -0
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/optimizer.pt +3 -0
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/rng_state.pth +3 -0
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/scheduler.pt +3 -0
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/trainer_state.json +2963 -0
- with_input/decomp_code_with_intermediates/llama-2-7b/best_model/training_args.bin +3 -0
with_input/decomp_code_with_intermediates/llama-2-13b/best_model/adapter_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 104973389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1be8f8694836a75efc5677c7fbf5a2347f9119984a1fb487d97f560d2e4ba3c
|
| 3 |
size 104973389
|
with_input/decomp_code_with_intermediates/llama-2-13b/best_model/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 209984517
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ba202a5dd9f6e829efcc949f64a6920d1eb8cae18a5770a3a04f6beab758de6
|
| 3 |
size 209984517
|
with_input/decomp_code_with_intermediates/llama-2-13b/best_model/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14575
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32a7e257ee0a424339e28f278d725f4ca822b8498210079485ba1bab78524a38
|
| 3 |
size 14575
|
with_input/decomp_code_with_intermediates/llama-2-13b/best_model/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b2d3080c8e2e4ab287873ffeaf2b06cd1df976c27dba320c4b0ac6a28dd37d5
|
| 3 |
size 627
|
with_input/decomp_code_with_intermediates/llama-2-13b/best_model/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "checkpoints/instrucode/with_input/decomp_code_with_intermediates/llama-2-13b/checkpoint-
|
| 4 |
-
"epoch": 2.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2567,13 +2567,397 @@
|
|
| 2567 |
"eval_samples_per_second": 2.461,
|
| 2568 |
"eval_steps_per_second": 0.308,
|
| 2569 |
"step": 4000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2570 |
}
|
| 2571 |
],
|
| 2572 |
"logging_steps": 10,
|
| 2573 |
"max_steps": 7775,
|
| 2574 |
"num_train_epochs": 5,
|
| 2575 |
"save_steps": 200,
|
| 2576 |
-
"total_flos":
|
| 2577 |
"trial_name": null,
|
| 2578 |
"trial_params": null
|
| 2579 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.29191577434539795,
|
| 3 |
+
"best_model_checkpoint": "checkpoints/instrucode/with_input/decomp_code_with_intermediates/llama-2-13b/checkpoint-4600",
|
| 4 |
+
"epoch": 2.958199356913183,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 4600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2567 |
"eval_samples_per_second": 2.461,
|
| 2568 |
"eval_steps_per_second": 0.308,
|
| 2569 |
"step": 4000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 2.58,
|
| 2573 |
+
"learning_rate": 0.00014724429967426708,
|
| 2574 |
+
"loss": 0.266,
|
| 2575 |
+
"step": 4010
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 2.59,
|
| 2579 |
+
"learning_rate": 0.00014685342019543973,
|
| 2580 |
+
"loss": 0.2664,
|
| 2581 |
+
"step": 4020
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 2.59,
|
| 2585 |
+
"learning_rate": 0.00014646254071661236,
|
| 2586 |
+
"loss": 0.2735,
|
| 2587 |
+
"step": 4030
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.6,
|
| 2591 |
+
"learning_rate": 0.000146071661237785,
|
| 2592 |
+
"loss": 0.2636,
|
| 2593 |
+
"step": 4040
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 2.6,
|
| 2597 |
+
"learning_rate": 0.00014568078175895765,
|
| 2598 |
+
"loss": 0.2698,
|
| 2599 |
+
"step": 4050
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 2.61,
|
| 2603 |
+
"learning_rate": 0.00014528990228013027,
|
| 2604 |
+
"loss": 0.2805,
|
| 2605 |
+
"step": 4060
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 2.62,
|
| 2609 |
+
"learning_rate": 0.00014489902280130293,
|
| 2610 |
+
"loss": 0.2821,
|
| 2611 |
+
"step": 4070
|
| 2612 |
+
},
|
| 2613 |
+
{
|
| 2614 |
+
"epoch": 2.62,
|
| 2615 |
+
"learning_rate": 0.00014450814332247556,
|
| 2616 |
+
"loss": 0.2724,
|
| 2617 |
+
"step": 4080
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 2.63,
|
| 2621 |
+
"learning_rate": 0.00014411726384364819,
|
| 2622 |
+
"loss": 0.2673,
|
| 2623 |
+
"step": 4090
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 2.64,
|
| 2627 |
+
"learning_rate": 0.00014372638436482084,
|
| 2628 |
+
"loss": 0.2737,
|
| 2629 |
+
"step": 4100
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.64,
|
| 2633 |
+
"learning_rate": 0.00014333550488599347,
|
| 2634 |
+
"loss": 0.2808,
|
| 2635 |
+
"step": 4110
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 2.65,
|
| 2639 |
+
"learning_rate": 0.00014294462540716612,
|
| 2640 |
+
"loss": 0.2605,
|
| 2641 |
+
"step": 4120
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 2.66,
|
| 2645 |
+
"learning_rate": 0.00014255374592833875,
|
| 2646 |
+
"loss": 0.2592,
|
| 2647 |
+
"step": 4130
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 2.66,
|
| 2651 |
+
"learning_rate": 0.00014216286644951138,
|
| 2652 |
+
"loss": 0.2703,
|
| 2653 |
+
"step": 4140
|
| 2654 |
+
},
|
| 2655 |
+
{
|
| 2656 |
+
"epoch": 2.67,
|
| 2657 |
+
"learning_rate": 0.00014177198697068404,
|
| 2658 |
+
"loss": 0.269,
|
| 2659 |
+
"step": 4150
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 2.68,
|
| 2663 |
+
"learning_rate": 0.00014138110749185666,
|
| 2664 |
+
"loss": 0.2741,
|
| 2665 |
+
"step": 4160
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 2.68,
|
| 2669 |
+
"learning_rate": 0.0001409902280130293,
|
| 2670 |
+
"loss": 0.2778,
|
| 2671 |
+
"step": 4170
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.69,
|
| 2675 |
+
"learning_rate": 0.00014059934853420195,
|
| 2676 |
+
"loss": 0.2691,
|
| 2677 |
+
"step": 4180
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 2.69,
|
| 2681 |
+
"learning_rate": 0.00014020846905537458,
|
| 2682 |
+
"loss": 0.2734,
|
| 2683 |
+
"step": 4190
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 2.7,
|
| 2687 |
+
"learning_rate": 0.00013981758957654723,
|
| 2688 |
+
"loss": 0.2631,
|
| 2689 |
+
"step": 4200
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 2.7,
|
| 2693 |
+
"eval_loss": 0.29299652576446533,
|
| 2694 |
+
"eval_runtime": 812.6678,
|
| 2695 |
+
"eval_samples_per_second": 2.461,
|
| 2696 |
+
"eval_steps_per_second": 0.308,
|
| 2697 |
+
"step": 4200
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 2.71,
|
| 2701 |
+
"learning_rate": 0.00013942671009771986,
|
| 2702 |
+
"loss": 0.2685,
|
| 2703 |
+
"step": 4210
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 2.71,
|
| 2707 |
+
"learning_rate": 0.0001390358306188925,
|
| 2708 |
+
"loss": 0.2822,
|
| 2709 |
+
"step": 4220
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 2.72,
|
| 2713 |
+
"learning_rate": 0.00013864495114006514,
|
| 2714 |
+
"loss": 0.268,
|
| 2715 |
+
"step": 4230
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 2.73,
|
| 2719 |
+
"learning_rate": 0.00013825407166123777,
|
| 2720 |
+
"loss": 0.2691,
|
| 2721 |
+
"step": 4240
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 2.73,
|
| 2725 |
+
"learning_rate": 0.00013786319218241043,
|
| 2726 |
+
"loss": 0.2872,
|
| 2727 |
+
"step": 4250
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.74,
|
| 2731 |
+
"learning_rate": 0.00013747231270358305,
|
| 2732 |
+
"loss": 0.2747,
|
| 2733 |
+
"step": 4260
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 2.75,
|
| 2737 |
+
"learning_rate": 0.00013708143322475568,
|
| 2738 |
+
"loss": 0.2593,
|
| 2739 |
+
"step": 4270
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 2.75,
|
| 2743 |
+
"learning_rate": 0.00013669055374592834,
|
| 2744 |
+
"loss": 0.2702,
|
| 2745 |
+
"step": 4280
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 2.76,
|
| 2749 |
+
"learning_rate": 0.00013629967426710097,
|
| 2750 |
+
"loss": 0.2705,
|
| 2751 |
+
"step": 4290
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 2.77,
|
| 2755 |
+
"learning_rate": 0.0001359087947882736,
|
| 2756 |
+
"loss": 0.2846,
|
| 2757 |
+
"step": 4300
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 2.77,
|
| 2761 |
+
"learning_rate": 0.00013551791530944622,
|
| 2762 |
+
"loss": 0.2708,
|
| 2763 |
+
"step": 4310
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 2.78,
|
| 2767 |
+
"learning_rate": 0.00013512703583061888,
|
| 2768 |
+
"loss": 0.2765,
|
| 2769 |
+
"step": 4320
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.78,
|
| 2773 |
+
"learning_rate": 0.00013473615635179153,
|
| 2774 |
+
"loss": 0.2763,
|
| 2775 |
+
"step": 4330
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 2.79,
|
| 2779 |
+
"learning_rate": 0.00013434527687296416,
|
| 2780 |
+
"loss": 0.2695,
|
| 2781 |
+
"step": 4340
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.8,
|
| 2785 |
+
"learning_rate": 0.0001339543973941368,
|
| 2786 |
+
"loss": 0.2776,
|
| 2787 |
+
"step": 4350
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 2.8,
|
| 2791 |
+
"learning_rate": 0.00013356351791530944,
|
| 2792 |
+
"loss": 0.2685,
|
| 2793 |
+
"step": 4360
|
| 2794 |
+
},
|
| 2795 |
+
{
|
| 2796 |
+
"epoch": 2.81,
|
| 2797 |
+
"learning_rate": 0.00013317263843648207,
|
| 2798 |
+
"loss": 0.2632,
|
| 2799 |
+
"step": 4370
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 2.82,
|
| 2803 |
+
"learning_rate": 0.00013278175895765473,
|
| 2804 |
+
"loss": 0.2725,
|
| 2805 |
+
"step": 4380
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 2.82,
|
| 2809 |
+
"learning_rate": 0.00013239087947882735,
|
| 2810 |
+
"loss": 0.2761,
|
| 2811 |
+
"step": 4390
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.83,
|
| 2815 |
+
"learning_rate": 0.00013199999999999998,
|
| 2816 |
+
"loss": 0.2725,
|
| 2817 |
+
"step": 4400
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 2.83,
|
| 2821 |
+
"eval_loss": 0.29228949546813965,
|
| 2822 |
+
"eval_runtime": 813.002,
|
| 2823 |
+
"eval_samples_per_second": 2.46,
|
| 2824 |
+
"eval_steps_per_second": 0.308,
|
| 2825 |
+
"step": 4400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.84,
|
| 2829 |
+
"learning_rate": 0.00013160912052117264,
|
| 2830 |
+
"loss": 0.2816,
|
| 2831 |
+
"step": 4410
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 2.84,
|
| 2835 |
+
"learning_rate": 0.00013121824104234527,
|
| 2836 |
+
"loss": 0.2777,
|
| 2837 |
+
"step": 4420
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 2.85,
|
| 2841 |
+
"learning_rate": 0.0001308273615635179,
|
| 2842 |
+
"loss": 0.2579,
|
| 2843 |
+
"step": 4430
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 2.86,
|
| 2847 |
+
"learning_rate": 0.00013043648208469052,
|
| 2848 |
+
"loss": 0.274,
|
| 2849 |
+
"step": 4440
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 2.86,
|
| 2853 |
+
"learning_rate": 0.00013004560260586318,
|
| 2854 |
+
"loss": 0.2737,
|
| 2855 |
+
"step": 4450
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 2.87,
|
| 2859 |
+
"learning_rate": 0.00012965472312703583,
|
| 2860 |
+
"loss": 0.269,
|
| 2861 |
+
"step": 4460
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 2.87,
|
| 2865 |
+
"learning_rate": 0.00012926384364820846,
|
| 2866 |
+
"loss": 0.2683,
|
| 2867 |
+
"step": 4470
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.88,
|
| 2871 |
+
"learning_rate": 0.0001288729641693811,
|
| 2872 |
+
"loss": 0.2697,
|
| 2873 |
+
"step": 4480
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 2.89,
|
| 2877 |
+
"learning_rate": 0.00012848208469055372,
|
| 2878 |
+
"loss": 0.2724,
|
| 2879 |
+
"step": 4490
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 2.89,
|
| 2883 |
+
"learning_rate": 0.00012809120521172637,
|
| 2884 |
+
"loss": 0.2692,
|
| 2885 |
+
"step": 4500
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 2.9,
|
| 2889 |
+
"learning_rate": 0.00012770032573289903,
|
| 2890 |
+
"loss": 0.2734,
|
| 2891 |
+
"step": 4510
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 2.91,
|
| 2895 |
+
"learning_rate": 0.00012730944625407166,
|
| 2896 |
+
"loss": 0.256,
|
| 2897 |
+
"step": 4520
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 2.91,
|
| 2901 |
+
"learning_rate": 0.00012691856677524428,
|
| 2902 |
+
"loss": 0.2695,
|
| 2903 |
+
"step": 4530
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 2.92,
|
| 2907 |
+
"learning_rate": 0.00012652768729641694,
|
| 2908 |
+
"loss": 0.2743,
|
| 2909 |
+
"step": 4540
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.93,
|
| 2913 |
+
"learning_rate": 0.00012613680781758957,
|
| 2914 |
+
"loss": 0.2709,
|
| 2915 |
+
"step": 4550
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 2.93,
|
| 2919 |
+
"learning_rate": 0.0001257459283387622,
|
| 2920 |
+
"loss": 0.2662,
|
| 2921 |
+
"step": 4560
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.94,
|
| 2925 |
+
"learning_rate": 0.00012535504885993482,
|
| 2926 |
+
"loss": 0.2664,
|
| 2927 |
+
"step": 4570
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 2.95,
|
| 2931 |
+
"learning_rate": 0.00012496416938110748,
|
| 2932 |
+
"loss": 0.2567,
|
| 2933 |
+
"step": 4580
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 2.95,
|
| 2937 |
+
"learning_rate": 0.00012457328990228013,
|
| 2938 |
+
"loss": 0.2645,
|
| 2939 |
+
"step": 4590
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 2.96,
|
| 2943 |
+
"learning_rate": 0.00012418241042345276,
|
| 2944 |
+
"loss": 0.2708,
|
| 2945 |
+
"step": 4600
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 2.96,
|
| 2949 |
+
"eval_loss": 0.29191577434539795,
|
| 2950 |
+
"eval_runtime": 812.9344,
|
| 2951 |
+
"eval_samples_per_second": 2.46,
|
| 2952 |
+
"eval_steps_per_second": 0.308,
|
| 2953 |
+
"step": 4600
|
| 2954 |
}
|
| 2955 |
],
|
| 2956 |
"logging_steps": 10,
|
| 2957 |
"max_steps": 7775,
|
| 2958 |
"num_train_epochs": 5,
|
| 2959 |
"save_steps": 200,
|
| 2960 |
+
"total_flos": 8.106140807566295e+18,
|
| 2961 |
"trial_name": null,
|
| 2962 |
"trial_params": null
|
| 2963 |
}
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/adapter_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"enable_lora": null,
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"lora_alpha": 16,
|
| 9 |
+
"lora_dropout": 0.05,
|
| 10 |
+
"merge_weights": false,
|
| 11 |
+
"modules_to_save": null,
|
| 12 |
+
"peft_type": "LORA",
|
| 13 |
+
"r": 16,
|
| 14 |
+
"target_modules": [
|
| 15 |
+
"q_proj",
|
| 16 |
+
"k_proj",
|
| 17 |
+
"v_proj",
|
| 18 |
+
"o_proj"
|
| 19 |
+
],
|
| 20 |
+
"task_type": "CAUSAL_LM"
|
| 21 |
+
}
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e0405d88cb30e39ecc34606d067e7d4c47d582098529effa8deac41622b18ca
|
| 3 |
+
size 67201357
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2d224ff38e2edcdd9e6ed1a6ca7ccd31de023b72f29912527934c65d33c783b
|
| 3 |
+
size 134433093
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d048c6aad44c269b2e352d34467f9c6e448b72d27c99c5b0723b3dc385b552ac
|
| 3 |
+
size 14575
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ef4538bb4c4ccd639f82a857dcd9692f818d5cfe746cb7b8a3f0a2708086ade
|
| 3 |
+
size 627
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/trainer_state.json
ADDED
|
@@ -0,0 +1,2963 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.3047821521759033,
|
| 3 |
+
"best_model_checkpoint": "checkpoints/instrucode/with_input/decomp_code_with_intermediates/llama-2-7b/checkpoint-4600",
|
| 4 |
+
"epoch": 2.958199356913183,
|
| 5 |
+
"eval_steps": 200,
|
| 6 |
+
"global_step": 4600,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 2.9999999999999997e-05,
|
| 14 |
+
"loss": 1.0271,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 5.9999999999999995e-05,
|
| 20 |
+
"loss": 0.9546,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.02,
|
| 25 |
+
"learning_rate": 8.999999999999999e-05,
|
| 26 |
+
"loss": 0.8146,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.03,
|
| 31 |
+
"learning_rate": 0.00011999999999999999,
|
| 32 |
+
"loss": 0.575,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.03,
|
| 37 |
+
"learning_rate": 0.00015,
|
| 38 |
+
"loss": 0.4684,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.04,
|
| 43 |
+
"learning_rate": 0.00017999999999999998,
|
| 44 |
+
"loss": 0.4319,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.05,
|
| 49 |
+
"learning_rate": 0.00020999999999999998,
|
| 50 |
+
"loss": 0.4216,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05,
|
| 55 |
+
"learning_rate": 0.00023999999999999998,
|
| 56 |
+
"loss": 0.4052,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.06,
|
| 61 |
+
"learning_rate": 0.00027,
|
| 62 |
+
"loss": 0.3835,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.06,
|
| 67 |
+
"learning_rate": 0.0003,
|
| 68 |
+
"loss": 0.3884,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.07,
|
| 73 |
+
"learning_rate": 0.00029960912052117263,
|
| 74 |
+
"loss": 0.3753,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08,
|
| 79 |
+
"learning_rate": 0.0002992182410423453,
|
| 80 |
+
"loss": 0.3825,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.08,
|
| 85 |
+
"learning_rate": 0.0002988273615635179,
|
| 86 |
+
"loss": 0.3745,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.09,
|
| 91 |
+
"learning_rate": 0.00029843648208469054,
|
| 92 |
+
"loss": 0.3676,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1,
|
| 97 |
+
"learning_rate": 0.00029804560260586314,
|
| 98 |
+
"loss": 0.3813,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.1,
|
| 103 |
+
"learning_rate": 0.0002976547231270358,
|
| 104 |
+
"loss": 0.3692,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.11,
|
| 109 |
+
"learning_rate": 0.00029726384364820845,
|
| 110 |
+
"loss": 0.3661,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.12,
|
| 115 |
+
"learning_rate": 0.00029687296416938105,
|
| 116 |
+
"loss": 0.3591,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.12,
|
| 121 |
+
"learning_rate": 0.0002964820846905537,
|
| 122 |
+
"loss": 0.3638,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.13,
|
| 127 |
+
"learning_rate": 0.00029609120521172636,
|
| 128 |
+
"loss": 0.3624,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.13,
|
| 133 |
+
"eval_loss": 0.35484859347343445,
|
| 134 |
+
"eval_runtime": 622.7911,
|
| 135 |
+
"eval_samples_per_second": 3.211,
|
| 136 |
+
"eval_steps_per_second": 0.401,
|
| 137 |
+
"step": 200
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.14,
|
| 141 |
+
"learning_rate": 0.00029573941368078174,
|
| 142 |
+
"loss": 0.3593,
|
| 143 |
+
"step": 210
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.14,
|
| 147 |
+
"learning_rate": 0.0002953485342019544,
|
| 148 |
+
"loss": 0.3482,
|
| 149 |
+
"step": 220
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.15,
|
| 153 |
+
"learning_rate": 0.00029495765472312705,
|
| 154 |
+
"loss": 0.3452,
|
| 155 |
+
"step": 230
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.15,
|
| 159 |
+
"learning_rate": 0.00029456677524429965,
|
| 160 |
+
"loss": 0.3469,
|
| 161 |
+
"step": 240
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.16,
|
| 165 |
+
"learning_rate": 0.0002941758957654723,
|
| 166 |
+
"loss": 0.3535,
|
| 167 |
+
"step": 250
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.17,
|
| 171 |
+
"learning_rate": 0.0002937850162866449,
|
| 172 |
+
"loss": 0.3562,
|
| 173 |
+
"step": 260
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.17,
|
| 177 |
+
"learning_rate": 0.00029339413680781756,
|
| 178 |
+
"loss": 0.3434,
|
| 179 |
+
"step": 270
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.18,
|
| 183 |
+
"learning_rate": 0.0002930032573289902,
|
| 184 |
+
"loss": 0.3482,
|
| 185 |
+
"step": 280
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.19,
|
| 189 |
+
"learning_rate": 0.0002926123778501628,
|
| 190 |
+
"loss": 0.3601,
|
| 191 |
+
"step": 290
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.19,
|
| 195 |
+
"learning_rate": 0.00029222149837133547,
|
| 196 |
+
"loss": 0.3501,
|
| 197 |
+
"step": 300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.2,
|
| 201 |
+
"learning_rate": 0.0002918306188925081,
|
| 202 |
+
"loss": 0.3532,
|
| 203 |
+
"step": 310
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.21,
|
| 207 |
+
"learning_rate": 0.0002914397394136808,
|
| 208 |
+
"loss": 0.352,
|
| 209 |
+
"step": 320
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 0.21,
|
| 213 |
+
"learning_rate": 0.0002910488599348534,
|
| 214 |
+
"loss": 0.3417,
|
| 215 |
+
"step": 330
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.22,
|
| 219 |
+
"learning_rate": 0.00029065798045602604,
|
| 220 |
+
"loss": 0.3354,
|
| 221 |
+
"step": 340
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.23,
|
| 225 |
+
"learning_rate": 0.00029026710097719864,
|
| 226 |
+
"loss": 0.3534,
|
| 227 |
+
"step": 350
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.23,
|
| 231 |
+
"learning_rate": 0.0002898762214983713,
|
| 232 |
+
"loss": 0.3395,
|
| 233 |
+
"step": 360
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.24,
|
| 237 |
+
"learning_rate": 0.00028948534201954395,
|
| 238 |
+
"loss": 0.3537,
|
| 239 |
+
"step": 370
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.24,
|
| 243 |
+
"learning_rate": 0.0002890944625407166,
|
| 244 |
+
"loss": 0.3423,
|
| 245 |
+
"step": 380
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.25,
|
| 249 |
+
"learning_rate": 0.0002887035830618892,
|
| 250 |
+
"loss": 0.3323,
|
| 251 |
+
"step": 390
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.26,
|
| 255 |
+
"learning_rate": 0.00028831270358306186,
|
| 256 |
+
"loss": 0.3461,
|
| 257 |
+
"step": 400
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.26,
|
| 261 |
+
"eval_loss": 0.3401348292827606,
|
| 262 |
+
"eval_runtime": 622.7746,
|
| 263 |
+
"eval_samples_per_second": 3.211,
|
| 264 |
+
"eval_steps_per_second": 0.401,
|
| 265 |
+
"step": 400
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.26,
|
| 269 |
+
"learning_rate": 0.0002879218241042345,
|
| 270 |
+
"loss": 0.3458,
|
| 271 |
+
"step": 410
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 0.27,
|
| 275 |
+
"learning_rate": 0.0002875309446254071,
|
| 276 |
+
"loss": 0.3286,
|
| 277 |
+
"step": 420
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.28,
|
| 281 |
+
"learning_rate": 0.0002871400651465798,
|
| 282 |
+
"loss": 0.3352,
|
| 283 |
+
"step": 430
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.28,
|
| 287 |
+
"learning_rate": 0.00028674918566775243,
|
| 288 |
+
"loss": 0.3406,
|
| 289 |
+
"step": 440
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.29,
|
| 293 |
+
"learning_rate": 0.00028635830618892503,
|
| 294 |
+
"loss": 0.3382,
|
| 295 |
+
"step": 450
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.3,
|
| 299 |
+
"learning_rate": 0.0002859674267100977,
|
| 300 |
+
"loss": 0.3501,
|
| 301 |
+
"step": 460
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.3,
|
| 305 |
+
"learning_rate": 0.00028557654723127034,
|
| 306 |
+
"loss": 0.3494,
|
| 307 |
+
"step": 470
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 0.31,
|
| 311 |
+
"learning_rate": 0.000285185667752443,
|
| 312 |
+
"loss": 0.3364,
|
| 313 |
+
"step": 480
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.32,
|
| 317 |
+
"learning_rate": 0.00028483387622149837,
|
| 318 |
+
"loss": 0.3453,
|
| 319 |
+
"step": 490
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.32,
|
| 323 |
+
"learning_rate": 0.00028444299674267097,
|
| 324 |
+
"loss": 0.3407,
|
| 325 |
+
"step": 500
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.33,
|
| 329 |
+
"learning_rate": 0.0002840521172638436,
|
| 330 |
+
"loss": 0.3369,
|
| 331 |
+
"step": 510
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.33,
|
| 335 |
+
"learning_rate": 0.0002836612377850163,
|
| 336 |
+
"loss": 0.3297,
|
| 337 |
+
"step": 520
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 0.34,
|
| 341 |
+
"learning_rate": 0.0002832703583061889,
|
| 342 |
+
"loss": 0.3347,
|
| 343 |
+
"step": 530
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 0.35,
|
| 347 |
+
"learning_rate": 0.00028287947882736154,
|
| 348 |
+
"loss": 0.3424,
|
| 349 |
+
"step": 540
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.35,
|
| 353 |
+
"learning_rate": 0.00028248859934853414,
|
| 354 |
+
"loss": 0.3341,
|
| 355 |
+
"step": 550
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 0.36,
|
| 359 |
+
"learning_rate": 0.0002820977198697068,
|
| 360 |
+
"loss": 0.3383,
|
| 361 |
+
"step": 560
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.37,
|
| 365 |
+
"learning_rate": 0.00028170684039087945,
|
| 366 |
+
"loss": 0.3328,
|
| 367 |
+
"step": 570
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.37,
|
| 371 |
+
"learning_rate": 0.0002813159609120521,
|
| 372 |
+
"loss": 0.3388,
|
| 373 |
+
"step": 580
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.38,
|
| 377 |
+
"learning_rate": 0.00028092508143322476,
|
| 378 |
+
"loss": 0.325,
|
| 379 |
+
"step": 590
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 0.39,
|
| 383 |
+
"learning_rate": 0.00028053420195439736,
|
| 384 |
+
"loss": 0.3366,
|
| 385 |
+
"step": 600
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 0.39,
|
| 389 |
+
"eval_loss": 0.3323810398578644,
|
| 390 |
+
"eval_runtime": 618.7233,
|
| 391 |
+
"eval_samples_per_second": 3.232,
|
| 392 |
+
"eval_steps_per_second": 0.404,
|
| 393 |
+
"step": 600
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.39,
|
| 397 |
+
"learning_rate": 0.00028014332247557,
|
| 398 |
+
"loss": 0.3235,
|
| 399 |
+
"step": 610
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.4,
|
| 403 |
+
"learning_rate": 0.00027975244299674267,
|
| 404 |
+
"loss": 0.3252,
|
| 405 |
+
"step": 620
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.41,
|
| 409 |
+
"learning_rate": 0.00027936156351791527,
|
| 410 |
+
"loss": 0.3336,
|
| 411 |
+
"step": 630
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.41,
|
| 415 |
+
"learning_rate": 0.0002789706840390879,
|
| 416 |
+
"loss": 0.3467,
|
| 417 |
+
"step": 640
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.42,
|
| 421 |
+
"learning_rate": 0.0002785798045602606,
|
| 422 |
+
"loss": 0.3443,
|
| 423 |
+
"step": 650
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.42,
|
| 427 |
+
"learning_rate": 0.0002781889250814332,
|
| 428 |
+
"loss": 0.3341,
|
| 429 |
+
"step": 660
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.43,
|
| 433 |
+
"learning_rate": 0.00027783713355048856,
|
| 434 |
+
"loss": 0.3489,
|
| 435 |
+
"step": 670
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.44,
|
| 439 |
+
"learning_rate": 0.0002774462540716612,
|
| 440 |
+
"loss": 0.3488,
|
| 441 |
+
"step": 680
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.44,
|
| 445 |
+
"learning_rate": 0.00027705537459283387,
|
| 446 |
+
"loss": 0.3274,
|
| 447 |
+
"step": 690
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.45,
|
| 451 |
+
"learning_rate": 0.0002766644951140065,
|
| 452 |
+
"loss": 0.3366,
|
| 453 |
+
"step": 700
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.46,
|
| 457 |
+
"learning_rate": 0.0002762736156351791,
|
| 458 |
+
"loss": 0.3237,
|
| 459 |
+
"step": 710
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.46,
|
| 463 |
+
"learning_rate": 0.0002758827361563518,
|
| 464 |
+
"loss": 0.3276,
|
| 465 |
+
"step": 720
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.47,
|
| 469 |
+
"learning_rate": 0.0002754918566775244,
|
| 470 |
+
"loss": 0.3411,
|
| 471 |
+
"step": 730
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.48,
|
| 475 |
+
"learning_rate": 0.00027510097719869703,
|
| 476 |
+
"loss": 0.3207,
|
| 477 |
+
"step": 740
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.48,
|
| 481 |
+
"learning_rate": 0.0002747100977198697,
|
| 482 |
+
"loss": 0.3261,
|
| 483 |
+
"step": 750
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.49,
|
| 487 |
+
"learning_rate": 0.00027431921824104234,
|
| 488 |
+
"loss": 0.3291,
|
| 489 |
+
"step": 760
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.5,
|
| 493 |
+
"learning_rate": 0.00027392833876221495,
|
| 494 |
+
"loss": 0.3323,
|
| 495 |
+
"step": 770
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.5,
|
| 499 |
+
"learning_rate": 0.0002735374592833876,
|
| 500 |
+
"loss": 0.3461,
|
| 501 |
+
"step": 780
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.51,
|
| 505 |
+
"learning_rate": 0.00027314657980456026,
|
| 506 |
+
"loss": 0.3284,
|
| 507 |
+
"step": 790
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.51,
|
| 511 |
+
"learning_rate": 0.0002727557003257329,
|
| 512 |
+
"loss": 0.329,
|
| 513 |
+
"step": 800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.51,
|
| 517 |
+
"eval_loss": 0.32687681913375854,
|
| 518 |
+
"eval_runtime": 617.3312,
|
| 519 |
+
"eval_samples_per_second": 3.24,
|
| 520 |
+
"eval_steps_per_second": 0.405,
|
| 521 |
+
"step": 800
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.52,
|
| 525 |
+
"learning_rate": 0.0002723648208469055,
|
| 526 |
+
"loss": 0.3361,
|
| 527 |
+
"step": 810
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.53,
|
| 531 |
+
"learning_rate": 0.00027197394136807817,
|
| 532 |
+
"loss": 0.3192,
|
| 533 |
+
"step": 820
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.53,
|
| 537 |
+
"learning_rate": 0.00027158306188925077,
|
| 538 |
+
"loss": 0.3237,
|
| 539 |
+
"step": 830
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.54,
|
| 543 |
+
"learning_rate": 0.0002711921824104234,
|
| 544 |
+
"loss": 0.3351,
|
| 545 |
+
"step": 840
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.55,
|
| 549 |
+
"learning_rate": 0.0002708013029315961,
|
| 550 |
+
"loss": 0.3411,
|
| 551 |
+
"step": 850
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 0.55,
|
| 555 |
+
"learning_rate": 0.00027041042345276873,
|
| 556 |
+
"loss": 0.3353,
|
| 557 |
+
"step": 860
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.56,
|
| 561 |
+
"learning_rate": 0.00027001954397394134,
|
| 562 |
+
"loss": 0.3343,
|
| 563 |
+
"step": 870
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.57,
|
| 567 |
+
"learning_rate": 0.000269628664495114,
|
| 568 |
+
"loss": 0.323,
|
| 569 |
+
"step": 880
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.57,
|
| 573 |
+
"learning_rate": 0.00026923778501628665,
|
| 574 |
+
"loss": 0.3312,
|
| 575 |
+
"step": 890
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.58,
|
| 579 |
+
"learning_rate": 0.00026884690553745925,
|
| 580 |
+
"loss": 0.3212,
|
| 581 |
+
"step": 900
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.59,
|
| 585 |
+
"learning_rate": 0.0002684560260586319,
|
| 586 |
+
"loss": 0.3253,
|
| 587 |
+
"step": 910
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 0.59,
|
| 591 |
+
"learning_rate": 0.0002680651465798045,
|
| 592 |
+
"loss": 0.3245,
|
| 593 |
+
"step": 920
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 0.6,
|
| 597 |
+
"learning_rate": 0.00026767426710097716,
|
| 598 |
+
"loss": 0.3329,
|
| 599 |
+
"step": 930
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.6,
|
| 603 |
+
"learning_rate": 0.0002672833876221498,
|
| 604 |
+
"loss": 0.3207,
|
| 605 |
+
"step": 940
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.61,
|
| 609 |
+
"learning_rate": 0.00026689250814332247,
|
| 610 |
+
"loss": 0.3327,
|
| 611 |
+
"step": 950
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.62,
|
| 615 |
+
"learning_rate": 0.0002665016286644951,
|
| 616 |
+
"loss": 0.3132,
|
| 617 |
+
"step": 960
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.62,
|
| 621 |
+
"learning_rate": 0.0002661107491856677,
|
| 622 |
+
"loss": 0.306,
|
| 623 |
+
"step": 970
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 0.63,
|
| 627 |
+
"learning_rate": 0.0002657198697068404,
|
| 628 |
+
"loss": 0.3169,
|
| 629 |
+
"step": 980
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 0.64,
|
| 633 |
+
"learning_rate": 0.000265328990228013,
|
| 634 |
+
"loss": 0.3217,
|
| 635 |
+
"step": 990
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 0.64,
|
| 639 |
+
"learning_rate": 0.00026493811074918564,
|
| 640 |
+
"loss": 0.3409,
|
| 641 |
+
"step": 1000
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.64,
|
| 645 |
+
"eval_loss": 0.3231165111064911,
|
| 646 |
+
"eval_runtime": 607.1505,
|
| 647 |
+
"eval_samples_per_second": 3.294,
|
| 648 |
+
"eval_steps_per_second": 0.412,
|
| 649 |
+
"step": 1000
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 0.65,
|
| 653 |
+
"learning_rate": 0.0002645472312703583,
|
| 654 |
+
"loss": 0.322,
|
| 655 |
+
"step": 1010
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.66,
|
| 659 |
+
"learning_rate": 0.0002641563517915309,
|
| 660 |
+
"loss": 0.3225,
|
| 661 |
+
"step": 1020
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.66,
|
| 665 |
+
"learning_rate": 0.00026376547231270355,
|
| 666 |
+
"loss": 0.3292,
|
| 667 |
+
"step": 1030
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.67,
|
| 671 |
+
"learning_rate": 0.0002633745928338762,
|
| 672 |
+
"loss": 0.3245,
|
| 673 |
+
"step": 1040
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 0.68,
|
| 677 |
+
"learning_rate": 0.00026298371335504886,
|
| 678 |
+
"loss": 0.3423,
|
| 679 |
+
"step": 1050
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 0.68,
|
| 683 |
+
"learning_rate": 0.00026259283387622146,
|
| 684 |
+
"loss": 0.3206,
|
| 685 |
+
"step": 1060
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 0.69,
|
| 689 |
+
"learning_rate": 0.0002622019543973941,
|
| 690 |
+
"loss": 0.3352,
|
| 691 |
+
"step": 1070
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.69,
|
| 695 |
+
"learning_rate": 0.00026181107491856677,
|
| 696 |
+
"loss": 0.3293,
|
| 697 |
+
"step": 1080
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.7,
|
| 701 |
+
"learning_rate": 0.00026142019543973937,
|
| 702 |
+
"loss": 0.3337,
|
| 703 |
+
"step": 1090
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.71,
|
| 707 |
+
"learning_rate": 0.000261029315960912,
|
| 708 |
+
"loss": 0.3341,
|
| 709 |
+
"step": 1100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.71,
|
| 713 |
+
"learning_rate": 0.0002606384364820847,
|
| 714 |
+
"loss": 0.3269,
|
| 715 |
+
"step": 1110
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 0.72,
|
| 719 |
+
"learning_rate": 0.0002602475570032573,
|
| 720 |
+
"loss": 0.3296,
|
| 721 |
+
"step": 1120
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.73,
|
| 725 |
+
"learning_rate": 0.00025985667752442994,
|
| 726 |
+
"loss": 0.3218,
|
| 727 |
+
"step": 1130
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 0.73,
|
| 731 |
+
"learning_rate": 0.0002594657980456026,
|
| 732 |
+
"loss": 0.319,
|
| 733 |
+
"step": 1140
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.74,
|
| 737 |
+
"learning_rate": 0.00025907491856677525,
|
| 738 |
+
"loss": 0.3268,
|
| 739 |
+
"step": 1150
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.75,
|
| 743 |
+
"learning_rate": 0.00025868403908794785,
|
| 744 |
+
"loss": 0.3158,
|
| 745 |
+
"step": 1160
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.75,
|
| 749 |
+
"learning_rate": 0.0002582931596091205,
|
| 750 |
+
"loss": 0.3288,
|
| 751 |
+
"step": 1170
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.76,
|
| 755 |
+
"learning_rate": 0.0002579022801302931,
|
| 756 |
+
"loss": 0.3114,
|
| 757 |
+
"step": 1180
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 0.77,
|
| 761 |
+
"learning_rate": 0.00025751140065146576,
|
| 762 |
+
"loss": 0.3274,
|
| 763 |
+
"step": 1190
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 0.77,
|
| 767 |
+
"learning_rate": 0.0002571205211726384,
|
| 768 |
+
"loss": 0.3227,
|
| 769 |
+
"step": 1200
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 0.77,
|
| 773 |
+
"eval_loss": 0.320186048746109,
|
| 774 |
+
"eval_runtime": 613.0278,
|
| 775 |
+
"eval_samples_per_second": 3.262,
|
| 776 |
+
"eval_steps_per_second": 0.408,
|
| 777 |
+
"step": 1200
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.78,
|
| 781 |
+
"learning_rate": 0.00025672964169381107,
|
| 782 |
+
"loss": 0.3321,
|
| 783 |
+
"step": 1210
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.78,
|
| 787 |
+
"learning_rate": 0.0002563387622149837,
|
| 788 |
+
"loss": 0.33,
|
| 789 |
+
"step": 1220
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.79,
|
| 793 |
+
"learning_rate": 0.00025594788273615633,
|
| 794 |
+
"loss": 0.3287,
|
| 795 |
+
"step": 1230
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.8,
|
| 799 |
+
"learning_rate": 0.000255557003257329,
|
| 800 |
+
"loss": 0.337,
|
| 801 |
+
"step": 1240
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.8,
|
| 805 |
+
"learning_rate": 0.0002551661237785016,
|
| 806 |
+
"loss": 0.3281,
|
| 807 |
+
"step": 1250
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.81,
|
| 811 |
+
"learning_rate": 0.00025477524429967424,
|
| 812 |
+
"loss": 0.3275,
|
| 813 |
+
"step": 1260
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.82,
|
| 817 |
+
"learning_rate": 0.0002543843648208469,
|
| 818 |
+
"loss": 0.323,
|
| 819 |
+
"step": 1270
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.82,
|
| 823 |
+
"learning_rate": 0.0002539934853420195,
|
| 824 |
+
"loss": 0.3238,
|
| 825 |
+
"step": 1280
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.83,
|
| 829 |
+
"learning_rate": 0.00025360260586319215,
|
| 830 |
+
"loss": 0.3159,
|
| 831 |
+
"step": 1290
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.84,
|
| 835 |
+
"learning_rate": 0.0002532117263843648,
|
| 836 |
+
"loss": 0.3235,
|
| 837 |
+
"step": 1300
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.84,
|
| 841 |
+
"learning_rate": 0.00025282084690553746,
|
| 842 |
+
"loss": 0.3215,
|
| 843 |
+
"step": 1310
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.85,
|
| 847 |
+
"learning_rate": 0.00025242996742671006,
|
| 848 |
+
"loss": 0.3193,
|
| 849 |
+
"step": 1320
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.86,
|
| 853 |
+
"learning_rate": 0.0002520390879478827,
|
| 854 |
+
"loss": 0.3145,
|
| 855 |
+
"step": 1330
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.86,
|
| 859 |
+
"learning_rate": 0.0002516482084690553,
|
| 860 |
+
"loss": 0.3415,
|
| 861 |
+
"step": 1340
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.87,
|
| 865 |
+
"learning_rate": 0.000251257328990228,
|
| 866 |
+
"loss": 0.3307,
|
| 867 |
+
"step": 1350
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.87,
|
| 871 |
+
"learning_rate": 0.00025086644951140063,
|
| 872 |
+
"loss": 0.3262,
|
| 873 |
+
"step": 1360
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.88,
|
| 877 |
+
"learning_rate": 0.0002504755700325733,
|
| 878 |
+
"loss": 0.3197,
|
| 879 |
+
"step": 1370
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.89,
|
| 883 |
+
"learning_rate": 0.0002500846905537459,
|
| 884 |
+
"loss": 0.3242,
|
| 885 |
+
"step": 1380
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.89,
|
| 889 |
+
"learning_rate": 0.00024969381107491854,
|
| 890 |
+
"loss": 0.3225,
|
| 891 |
+
"step": 1390
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.9,
|
| 895 |
+
"learning_rate": 0.0002493029315960912,
|
| 896 |
+
"loss": 0.3296,
|
| 897 |
+
"step": 1400
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.9,
|
| 901 |
+
"eval_loss": 0.31777673959732056,
|
| 902 |
+
"eval_runtime": 620.9719,
|
| 903 |
+
"eval_samples_per_second": 3.221,
|
| 904 |
+
"eval_steps_per_second": 0.403,
|
| 905 |
+
"step": 1400
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.91,
|
| 909 |
+
"learning_rate": 0.00024891205211726385,
|
| 910 |
+
"loss": 0.3242,
|
| 911 |
+
"step": 1410
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 0.91,
|
| 915 |
+
"learning_rate": 0.00024852117263843645,
|
| 916 |
+
"loss": 0.3164,
|
| 917 |
+
"step": 1420
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 0.92,
|
| 921 |
+
"learning_rate": 0.0002481302931596091,
|
| 922 |
+
"loss": 0.3055,
|
| 923 |
+
"step": 1430
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 0.93,
|
| 927 |
+
"learning_rate": 0.0002477394136807817,
|
| 928 |
+
"loss": 0.3191,
|
| 929 |
+
"step": 1440
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.93,
|
| 933 |
+
"learning_rate": 0.00024734853420195436,
|
| 934 |
+
"loss": 0.3219,
|
| 935 |
+
"step": 1450
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.94,
|
| 939 |
+
"learning_rate": 0.000246957654723127,
|
| 940 |
+
"loss": 0.3225,
|
| 941 |
+
"step": 1460
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.95,
|
| 945 |
+
"learning_rate": 0.0002465667752442997,
|
| 946 |
+
"loss": 0.3121,
|
| 947 |
+
"step": 1470
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.95,
|
| 951 |
+
"learning_rate": 0.0002461758957654723,
|
| 952 |
+
"loss": 0.325,
|
| 953 |
+
"step": 1480
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 0.96,
|
| 957 |
+
"learning_rate": 0.00024578501628664493,
|
| 958 |
+
"loss": 0.3203,
|
| 959 |
+
"step": 1490
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.96,
|
| 963 |
+
"learning_rate": 0.0002453941368078176,
|
| 964 |
+
"loss": 0.3211,
|
| 965 |
+
"step": 1500
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 0.97,
|
| 969 |
+
"learning_rate": 0.0002450032573289902,
|
| 970 |
+
"loss": 0.3212,
|
| 971 |
+
"step": 1510
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.98,
|
| 975 |
+
"learning_rate": 0.00024461237785016284,
|
| 976 |
+
"loss": 0.3213,
|
| 977 |
+
"step": 1520
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 0.98,
|
| 981 |
+
"learning_rate": 0.0002442214983713355,
|
| 982 |
+
"loss": 0.3263,
|
| 983 |
+
"step": 1530
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.99,
|
| 987 |
+
"learning_rate": 0.0002438306188925081,
|
| 988 |
+
"loss": 0.3187,
|
| 989 |
+
"step": 1540
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.0,
|
| 993 |
+
"learning_rate": 0.00024343973941368075,
|
| 994 |
+
"loss": 0.3129,
|
| 995 |
+
"step": 1550
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.0,
|
| 999 |
+
"learning_rate": 0.0002430488599348534,
|
| 1000 |
+
"loss": 0.3109,
|
| 1001 |
+
"step": 1560
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.01,
|
| 1005 |
+
"learning_rate": 0.00024265798045602604,
|
| 1006 |
+
"loss": 0.3093,
|
| 1007 |
+
"step": 1570
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 1.02,
|
| 1011 |
+
"learning_rate": 0.0002422671009771987,
|
| 1012 |
+
"loss": 0.3044,
|
| 1013 |
+
"step": 1580
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 1.02,
|
| 1017 |
+
"learning_rate": 0.0002418762214983713,
|
| 1018 |
+
"loss": 0.3137,
|
| 1019 |
+
"step": 1590
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.03,
|
| 1023 |
+
"learning_rate": 0.00024148534201954395,
|
| 1024 |
+
"loss": 0.3057,
|
| 1025 |
+
"step": 1600
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.03,
|
| 1029 |
+
"eval_loss": 0.3158407509326935,
|
| 1030 |
+
"eval_runtime": 618.9877,
|
| 1031 |
+
"eval_samples_per_second": 3.231,
|
| 1032 |
+
"eval_steps_per_second": 0.404,
|
| 1033 |
+
"step": 1600
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.04,
|
| 1037 |
+
"learning_rate": 0.0002410944625407166,
|
| 1038 |
+
"loss": 0.3057,
|
| 1039 |
+
"step": 1610
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.04,
|
| 1043 |
+
"learning_rate": 0.00024070358306188923,
|
| 1044 |
+
"loss": 0.3123,
|
| 1045 |
+
"step": 1620
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.05,
|
| 1049 |
+
"learning_rate": 0.0002403127035830619,
|
| 1050 |
+
"loss": 0.3042,
|
| 1051 |
+
"step": 1630
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.05,
|
| 1055 |
+
"learning_rate": 0.0002399218241042345,
|
| 1056 |
+
"loss": 0.3028,
|
| 1057 |
+
"step": 1640
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.06,
|
| 1061 |
+
"learning_rate": 0.00023953094462540714,
|
| 1062 |
+
"loss": 0.2986,
|
| 1063 |
+
"step": 1650
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.07,
|
| 1067 |
+
"learning_rate": 0.00023914006514657977,
|
| 1068 |
+
"loss": 0.2956,
|
| 1069 |
+
"step": 1660
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.07,
|
| 1073 |
+
"learning_rate": 0.00023874918566775243,
|
| 1074 |
+
"loss": 0.3183,
|
| 1075 |
+
"step": 1670
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.08,
|
| 1079 |
+
"learning_rate": 0.00023835830618892508,
|
| 1080 |
+
"loss": 0.3081,
|
| 1081 |
+
"step": 1680
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.09,
|
| 1085 |
+
"learning_rate": 0.00023796742671009768,
|
| 1086 |
+
"loss": 0.3147,
|
| 1087 |
+
"step": 1690
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.09,
|
| 1091 |
+
"learning_rate": 0.00023757654723127034,
|
| 1092 |
+
"loss": 0.2932,
|
| 1093 |
+
"step": 1700
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.1,
|
| 1097 |
+
"learning_rate": 0.00023718566775244297,
|
| 1098 |
+
"loss": 0.3019,
|
| 1099 |
+
"step": 1710
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.11,
|
| 1103 |
+
"learning_rate": 0.00023679478827361562,
|
| 1104 |
+
"loss": 0.3149,
|
| 1105 |
+
"step": 1720
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.11,
|
| 1109 |
+
"learning_rate": 0.00023640390879478828,
|
| 1110 |
+
"loss": 0.314,
|
| 1111 |
+
"step": 1730
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 1.12,
|
| 1115 |
+
"learning_rate": 0.00023601302931596088,
|
| 1116 |
+
"loss": 0.3025,
|
| 1117 |
+
"step": 1740
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.13,
|
| 1121 |
+
"learning_rate": 0.00023562214983713353,
|
| 1122 |
+
"loss": 0.3045,
|
| 1123 |
+
"step": 1750
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.13,
|
| 1127 |
+
"learning_rate": 0.00023523127035830616,
|
| 1128 |
+
"loss": 0.3036,
|
| 1129 |
+
"step": 1760
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.14,
|
| 1133 |
+
"learning_rate": 0.00023484039087947882,
|
| 1134 |
+
"loss": 0.3088,
|
| 1135 |
+
"step": 1770
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.14,
|
| 1139 |
+
"learning_rate": 0.00023444951140065144,
|
| 1140 |
+
"loss": 0.3156,
|
| 1141 |
+
"step": 1780
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.15,
|
| 1145 |
+
"learning_rate": 0.00023405863192182407,
|
| 1146 |
+
"loss": 0.3019,
|
| 1147 |
+
"step": 1790
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.16,
|
| 1151 |
+
"learning_rate": 0.0002336677524429967,
|
| 1152 |
+
"loss": 0.3145,
|
| 1153 |
+
"step": 1800
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 1.16,
|
| 1157 |
+
"eval_loss": 0.3144530653953552,
|
| 1158 |
+
"eval_runtime": 627.6292,
|
| 1159 |
+
"eval_samples_per_second": 3.187,
|
| 1160 |
+
"eval_steps_per_second": 0.398,
|
| 1161 |
+
"step": 1800
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.16,
|
| 1165 |
+
"learning_rate": 0.00023327687296416936,
|
| 1166 |
+
"loss": 0.3062,
|
| 1167 |
+
"step": 1810
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 1.17,
|
| 1171 |
+
"learning_rate": 0.000232885993485342,
|
| 1172 |
+
"loss": 0.3147,
|
| 1173 |
+
"step": 1820
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 1.18,
|
| 1177 |
+
"learning_rate": 0.00023249511400651464,
|
| 1178 |
+
"loss": 0.314,
|
| 1179 |
+
"step": 1830
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.18,
|
| 1183 |
+
"learning_rate": 0.00023210423452768727,
|
| 1184 |
+
"loss": 0.3025,
|
| 1185 |
+
"step": 1840
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.19,
|
| 1189 |
+
"learning_rate": 0.0002317133550488599,
|
| 1190 |
+
"loss": 0.3099,
|
| 1191 |
+
"step": 1850
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 1.2,
|
| 1195 |
+
"learning_rate": 0.00023132247557003255,
|
| 1196 |
+
"loss": 0.2983,
|
| 1197 |
+
"step": 1860
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 1.2,
|
| 1201 |
+
"learning_rate": 0.0002309315960912052,
|
| 1202 |
+
"loss": 0.3092,
|
| 1203 |
+
"step": 1870
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 1.21,
|
| 1207 |
+
"learning_rate": 0.00023054071661237783,
|
| 1208 |
+
"loss": 0.3157,
|
| 1209 |
+
"step": 1880
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 1.22,
|
| 1213 |
+
"learning_rate": 0.00023014983713355046,
|
| 1214 |
+
"loss": 0.3083,
|
| 1215 |
+
"step": 1890
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.22,
|
| 1219 |
+
"learning_rate": 0.0002297589576547231,
|
| 1220 |
+
"loss": 0.3099,
|
| 1221 |
+
"step": 1900
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.23,
|
| 1225 |
+
"learning_rate": 0.00022936807817589575,
|
| 1226 |
+
"loss": 0.3033,
|
| 1227 |
+
"step": 1910
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.23,
|
| 1231 |
+
"learning_rate": 0.00022897719869706837,
|
| 1232 |
+
"loss": 0.3116,
|
| 1233 |
+
"step": 1920
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 1.24,
|
| 1237 |
+
"learning_rate": 0.00022858631921824103,
|
| 1238 |
+
"loss": 0.3063,
|
| 1239 |
+
"step": 1930
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 1.25,
|
| 1243 |
+
"learning_rate": 0.00022819543973941368,
|
| 1244 |
+
"loss": 0.2945,
|
| 1245 |
+
"step": 1940
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 1.25,
|
| 1249 |
+
"learning_rate": 0.00022780456026058629,
|
| 1250 |
+
"loss": 0.3154,
|
| 1251 |
+
"step": 1950
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 1.26,
|
| 1255 |
+
"learning_rate": 0.00022741368078175894,
|
| 1256 |
+
"loss": 0.3065,
|
| 1257 |
+
"step": 1960
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.27,
|
| 1261 |
+
"learning_rate": 0.00022702280130293157,
|
| 1262 |
+
"loss": 0.3054,
|
| 1263 |
+
"step": 1970
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.27,
|
| 1267 |
+
"learning_rate": 0.00022663192182410422,
|
| 1268 |
+
"loss": 0.3148,
|
| 1269 |
+
"step": 1980
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.28,
|
| 1273 |
+
"learning_rate": 0.00022624104234527688,
|
| 1274 |
+
"loss": 0.3009,
|
| 1275 |
+
"step": 1990
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 1.29,
|
| 1279 |
+
"learning_rate": 0.00022585016286644948,
|
| 1280 |
+
"loss": 0.3194,
|
| 1281 |
+
"step": 2000
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 1.29,
|
| 1285 |
+
"eval_loss": 0.3134351372718811,
|
| 1286 |
+
"eval_runtime": 638.1689,
|
| 1287 |
+
"eval_samples_per_second": 3.134,
|
| 1288 |
+
"eval_steps_per_second": 0.392,
|
| 1289 |
+
"step": 2000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.29,
|
| 1293 |
+
"learning_rate": 0.00022545928338762214,
|
| 1294 |
+
"loss": 0.307,
|
| 1295 |
+
"step": 2010
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 1.3,
|
| 1299 |
+
"learning_rate": 0.00022506840390879476,
|
| 1300 |
+
"loss": 0.2982,
|
| 1301 |
+
"step": 2020
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 1.31,
|
| 1305 |
+
"learning_rate": 0.00022467752442996742,
|
| 1306 |
+
"loss": 0.3063,
|
| 1307 |
+
"step": 2030
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 1.31,
|
| 1311 |
+
"learning_rate": 0.00022428664495114005,
|
| 1312 |
+
"loss": 0.3007,
|
| 1313 |
+
"step": 2040
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.32,
|
| 1317 |
+
"learning_rate": 0.00022389576547231268,
|
| 1318 |
+
"loss": 0.3071,
|
| 1319 |
+
"step": 2050
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.32,
|
| 1323 |
+
"learning_rate": 0.0002235048859934853,
|
| 1324 |
+
"loss": 0.3134,
|
| 1325 |
+
"step": 2060
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.33,
|
| 1329 |
+
"learning_rate": 0.00022311400651465796,
|
| 1330 |
+
"loss": 0.297,
|
| 1331 |
+
"step": 2070
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 1.34,
|
| 1335 |
+
"learning_rate": 0.00022272312703583061,
|
| 1336 |
+
"loss": 0.3114,
|
| 1337 |
+
"step": 2080
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 1.34,
|
| 1341 |
+
"learning_rate": 0.00022233224755700324,
|
| 1342 |
+
"loss": 0.301,
|
| 1343 |
+
"step": 2090
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.35,
|
| 1347 |
+
"learning_rate": 0.00022194136807817587,
|
| 1348 |
+
"loss": 0.3189,
|
| 1349 |
+
"step": 2100
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 1.36,
|
| 1353 |
+
"learning_rate": 0.0002215504885993485,
|
| 1354 |
+
"loss": 0.3021,
|
| 1355 |
+
"step": 2110
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 1.36,
|
| 1359 |
+
"learning_rate": 0.00022115960912052115,
|
| 1360 |
+
"loss": 0.2987,
|
| 1361 |
+
"step": 2120
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.37,
|
| 1365 |
+
"learning_rate": 0.0002207687296416938,
|
| 1366 |
+
"loss": 0.305,
|
| 1367 |
+
"step": 2130
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.38,
|
| 1371 |
+
"learning_rate": 0.00022037785016286644,
|
| 1372 |
+
"loss": 0.321,
|
| 1373 |
+
"step": 2140
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.38,
|
| 1377 |
+
"learning_rate": 0.00021998697068403907,
|
| 1378 |
+
"loss": 0.2944,
|
| 1379 |
+
"step": 2150
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 1.39,
|
| 1383 |
+
"learning_rate": 0.0002195960912052117,
|
| 1384 |
+
"loss": 0.3051,
|
| 1385 |
+
"step": 2160
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 1.4,
|
| 1389 |
+
"learning_rate": 0.00021920521172638435,
|
| 1390 |
+
"loss": 0.2972,
|
| 1391 |
+
"step": 2170
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 1.4,
|
| 1395 |
+
"learning_rate": 0.00021881433224755698,
|
| 1396 |
+
"loss": 0.3037,
|
| 1397 |
+
"step": 2180
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 1.41,
|
| 1401 |
+
"learning_rate": 0.00021842345276872963,
|
| 1402 |
+
"loss": 0.3093,
|
| 1403 |
+
"step": 2190
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.41,
|
| 1407 |
+
"learning_rate": 0.00021803257328990223,
|
| 1408 |
+
"loss": 0.3161,
|
| 1409 |
+
"step": 2200
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.41,
|
| 1413 |
+
"eval_loss": 0.31146764755249023,
|
| 1414 |
+
"eval_runtime": 612.5254,
|
| 1415 |
+
"eval_samples_per_second": 3.265,
|
| 1416 |
+
"eval_steps_per_second": 0.408,
|
| 1417 |
+
"step": 2200
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.42,
|
| 1421 |
+
"learning_rate": 0.0002176416938110749,
|
| 1422 |
+
"loss": 0.2985,
|
| 1423 |
+
"step": 2210
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.43,
|
| 1427 |
+
"learning_rate": 0.00021725081433224754,
|
| 1428 |
+
"loss": 0.3099,
|
| 1429 |
+
"step": 2220
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.43,
|
| 1433 |
+
"learning_rate": 0.00021685993485342017,
|
| 1434 |
+
"loss": 0.3027,
|
| 1435 |
+
"step": 2230
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.44,
|
| 1439 |
+
"learning_rate": 0.00021646905537459283,
|
| 1440 |
+
"loss": 0.304,
|
| 1441 |
+
"step": 2240
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.45,
|
| 1445 |
+
"learning_rate": 0.00021607817589576543,
|
| 1446 |
+
"loss": 0.3134,
|
| 1447 |
+
"step": 2250
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 1.45,
|
| 1451 |
+
"learning_rate": 0.00021568729641693808,
|
| 1452 |
+
"loss": 0.319,
|
| 1453 |
+
"step": 2260
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 1.46,
|
| 1457 |
+
"learning_rate": 0.00021529641693811074,
|
| 1458 |
+
"loss": 0.3048,
|
| 1459 |
+
"step": 2270
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.47,
|
| 1463 |
+
"learning_rate": 0.00021490553745928337,
|
| 1464 |
+
"loss": 0.3008,
|
| 1465 |
+
"step": 2280
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.47,
|
| 1469 |
+
"learning_rate": 0.00021451465798045602,
|
| 1470 |
+
"loss": 0.3187,
|
| 1471 |
+
"step": 2290
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.48,
|
| 1475 |
+
"learning_rate": 0.00021412377850162865,
|
| 1476 |
+
"loss": 0.3055,
|
| 1477 |
+
"step": 2300
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.49,
|
| 1481 |
+
"learning_rate": 0.00021373289902280128,
|
| 1482 |
+
"loss": 0.3072,
|
| 1483 |
+
"step": 2310
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.49,
|
| 1487 |
+
"learning_rate": 0.0002133420195439739,
|
| 1488 |
+
"loss": 0.3031,
|
| 1489 |
+
"step": 2320
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 1.5,
|
| 1493 |
+
"learning_rate": 0.00021295114006514656,
|
| 1494 |
+
"loss": 0.3098,
|
| 1495 |
+
"step": 2330
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.5,
|
| 1499 |
+
"learning_rate": 0.00021256026058631922,
|
| 1500 |
+
"loss": 0.3073,
|
| 1501 |
+
"step": 2340
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.51,
|
| 1505 |
+
"learning_rate": 0.00021216938110749185,
|
| 1506 |
+
"loss": 0.3049,
|
| 1507 |
+
"step": 2350
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.52,
|
| 1511 |
+
"learning_rate": 0.00021177850162866447,
|
| 1512 |
+
"loss": 0.3114,
|
| 1513 |
+
"step": 2360
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.52,
|
| 1517 |
+
"learning_rate": 0.0002113876221498371,
|
| 1518 |
+
"loss": 0.3074,
|
| 1519 |
+
"step": 2370
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.53,
|
| 1523 |
+
"learning_rate": 0.00021099674267100976,
|
| 1524 |
+
"loss": 0.3187,
|
| 1525 |
+
"step": 2380
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.54,
|
| 1529 |
+
"learning_rate": 0.0002106058631921824,
|
| 1530 |
+
"loss": 0.3034,
|
| 1531 |
+
"step": 2390
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 1.54,
|
| 1535 |
+
"learning_rate": 0.00021021498371335504,
|
| 1536 |
+
"loss": 0.3007,
|
| 1537 |
+
"step": 2400
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.54,
|
| 1541 |
+
"eval_loss": 0.3109865188598633,
|
| 1542 |
+
"eval_runtime": 608.1214,
|
| 1543 |
+
"eval_samples_per_second": 3.289,
|
| 1544 |
+
"eval_steps_per_second": 0.411,
|
| 1545 |
+
"step": 2400
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 1.55,
|
| 1549 |
+
"learning_rate": 0.00020982410423452767,
|
| 1550 |
+
"loss": 0.289,
|
| 1551 |
+
"step": 2410
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 1.56,
|
| 1555 |
+
"learning_rate": 0.0002094332247557003,
|
| 1556 |
+
"loss": 0.297,
|
| 1557 |
+
"step": 2420
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.56,
|
| 1561 |
+
"learning_rate": 0.00020904234527687295,
|
| 1562 |
+
"loss": 0.3101,
|
| 1563 |
+
"step": 2430
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.57,
|
| 1567 |
+
"learning_rate": 0.00020865146579804558,
|
| 1568 |
+
"loss": 0.2995,
|
| 1569 |
+
"step": 2440
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 1.58,
|
| 1573 |
+
"learning_rate": 0.00020826058631921824,
|
| 1574 |
+
"loss": 0.3033,
|
| 1575 |
+
"step": 2450
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 1.58,
|
| 1579 |
+
"learning_rate": 0.00020786970684039084,
|
| 1580 |
+
"loss": 0.3059,
|
| 1581 |
+
"step": 2460
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 1.59,
|
| 1585 |
+
"learning_rate": 0.0002074788273615635,
|
| 1586 |
+
"loss": 0.3056,
|
| 1587 |
+
"step": 2470
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 1.59,
|
| 1591 |
+
"learning_rate": 0.00020708794788273615,
|
| 1592 |
+
"loss": 0.3003,
|
| 1593 |
+
"step": 2480
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 1.6,
|
| 1597 |
+
"learning_rate": 0.00020669706840390877,
|
| 1598 |
+
"loss": 0.2931,
|
| 1599 |
+
"step": 2490
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.61,
|
| 1603 |
+
"learning_rate": 0.00020630618892508143,
|
| 1604 |
+
"loss": 0.3042,
|
| 1605 |
+
"step": 2500
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.61,
|
| 1609 |
+
"learning_rate": 0.00020591530944625403,
|
| 1610 |
+
"loss": 0.3049,
|
| 1611 |
+
"step": 2510
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.62,
|
| 1615 |
+
"learning_rate": 0.00020552442996742669,
|
| 1616 |
+
"loss": 0.3006,
|
| 1617 |
+
"step": 2520
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 1.63,
|
| 1621 |
+
"learning_rate": 0.00020513355048859934,
|
| 1622 |
+
"loss": 0.3085,
|
| 1623 |
+
"step": 2530
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 1.63,
|
| 1627 |
+
"learning_rate": 0.00020474267100977197,
|
| 1628 |
+
"loss": 0.3142,
|
| 1629 |
+
"step": 2540
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 1.64,
|
| 1633 |
+
"learning_rate": 0.00020435179153094462,
|
| 1634 |
+
"loss": 0.2969,
|
| 1635 |
+
"step": 2550
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 1.65,
|
| 1639 |
+
"learning_rate": 0.00020396091205211723,
|
| 1640 |
+
"loss": 0.2993,
|
| 1641 |
+
"step": 2560
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.65,
|
| 1645 |
+
"learning_rate": 0.00020357003257328988,
|
| 1646 |
+
"loss": 0.3145,
|
| 1647 |
+
"step": 2570
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.66,
|
| 1651 |
+
"learning_rate": 0.0002031791530944625,
|
| 1652 |
+
"loss": 0.3038,
|
| 1653 |
+
"step": 2580
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 1.67,
|
| 1657 |
+
"learning_rate": 0.00020278827361563516,
|
| 1658 |
+
"loss": 0.3103,
|
| 1659 |
+
"step": 2590
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 1.67,
|
| 1663 |
+
"learning_rate": 0.00020239739413680782,
|
| 1664 |
+
"loss": 0.3104,
|
| 1665 |
+
"step": 2600
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 1.67,
|
| 1669 |
+
"eval_loss": 0.3095039129257202,
|
| 1670 |
+
"eval_runtime": 619.0261,
|
| 1671 |
+
"eval_samples_per_second": 3.231,
|
| 1672 |
+
"eval_steps_per_second": 0.404,
|
| 1673 |
+
"step": 2600
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.68,
|
| 1677 |
+
"learning_rate": 0.00020200651465798042,
|
| 1678 |
+
"loss": 0.2998,
|
| 1679 |
+
"step": 2610
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 1.68,
|
| 1683 |
+
"learning_rate": 0.00020161563517915308,
|
| 1684 |
+
"loss": 0.3099,
|
| 1685 |
+
"step": 2620
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 1.69,
|
| 1689 |
+
"learning_rate": 0.0002012247557003257,
|
| 1690 |
+
"loss": 0.295,
|
| 1691 |
+
"step": 2630
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 1.7,
|
| 1695 |
+
"learning_rate": 0.00020083387622149836,
|
| 1696 |
+
"loss": 0.3149,
|
| 1697 |
+
"step": 2640
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.7,
|
| 1701 |
+
"learning_rate": 0.00020044299674267101,
|
| 1702 |
+
"loss": 0.2916,
|
| 1703 |
+
"step": 2650
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.71,
|
| 1707 |
+
"learning_rate": 0.00020005211726384362,
|
| 1708 |
+
"loss": 0.3077,
|
| 1709 |
+
"step": 2660
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.72,
|
| 1713 |
+
"learning_rate": 0.00019966123778501627,
|
| 1714 |
+
"loss": 0.3039,
|
| 1715 |
+
"step": 2670
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 1.72,
|
| 1719 |
+
"learning_rate": 0.0001992703583061889,
|
| 1720 |
+
"loss": 0.2973,
|
| 1721 |
+
"step": 2680
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 1.73,
|
| 1725 |
+
"learning_rate": 0.00019887947882736155,
|
| 1726 |
+
"loss": 0.3065,
|
| 1727 |
+
"step": 2690
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 1.74,
|
| 1731 |
+
"learning_rate": 0.00019848859934853418,
|
| 1732 |
+
"loss": 0.302,
|
| 1733 |
+
"step": 2700
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 1.74,
|
| 1737 |
+
"learning_rate": 0.00019809771986970684,
|
| 1738 |
+
"loss": 0.3038,
|
| 1739 |
+
"step": 2710
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.75,
|
| 1743 |
+
"learning_rate": 0.00019770684039087944,
|
| 1744 |
+
"loss": 0.3141,
|
| 1745 |
+
"step": 2720
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.76,
|
| 1749 |
+
"learning_rate": 0.0001973159609120521,
|
| 1750 |
+
"loss": 0.3057,
|
| 1751 |
+
"step": 2730
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 1.76,
|
| 1755 |
+
"learning_rate": 0.00019692508143322475,
|
| 1756 |
+
"loss": 0.3071,
|
| 1757 |
+
"step": 2740
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 1.77,
|
| 1761 |
+
"learning_rate": 0.00019653420195439738,
|
| 1762 |
+
"loss": 0.3104,
|
| 1763 |
+
"step": 2750
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 1.77,
|
| 1767 |
+
"learning_rate": 0.00019614332247557003,
|
| 1768 |
+
"loss": 0.2963,
|
| 1769 |
+
"step": 2760
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 1.78,
|
| 1773 |
+
"learning_rate": 0.00019575244299674263,
|
| 1774 |
+
"loss": 0.3005,
|
| 1775 |
+
"step": 2770
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 1.79,
|
| 1779 |
+
"learning_rate": 0.0001953615635179153,
|
| 1780 |
+
"loss": 0.2999,
|
| 1781 |
+
"step": 2780
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.79,
|
| 1785 |
+
"learning_rate": 0.00019497068403908794,
|
| 1786 |
+
"loss": 0.3054,
|
| 1787 |
+
"step": 2790
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.8,
|
| 1791 |
+
"learning_rate": 0.00019457980456026057,
|
| 1792 |
+
"loss": 0.3005,
|
| 1793 |
+
"step": 2800
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 1.8,
|
| 1797 |
+
"eval_loss": 0.308525413274765,
|
| 1798 |
+
"eval_runtime": 624.5619,
|
| 1799 |
+
"eval_samples_per_second": 3.202,
|
| 1800 |
+
"eval_steps_per_second": 0.4,
|
| 1801 |
+
"step": 2800
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.81,
|
| 1805 |
+
"learning_rate": 0.00019418892508143323,
|
| 1806 |
+
"loss": 0.3012,
|
| 1807 |
+
"step": 2810
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 1.81,
|
| 1811 |
+
"learning_rate": 0.00019379804560260583,
|
| 1812 |
+
"loss": 0.3038,
|
| 1813 |
+
"step": 2820
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.82,
|
| 1817 |
+
"learning_rate": 0.00019340716612377848,
|
| 1818 |
+
"loss": 0.2939,
|
| 1819 |
+
"step": 2830
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 1.83,
|
| 1823 |
+
"learning_rate": 0.0001930162866449511,
|
| 1824 |
+
"loss": 0.2947,
|
| 1825 |
+
"step": 2840
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 1.83,
|
| 1829 |
+
"learning_rate": 0.00019262540716612377,
|
| 1830 |
+
"loss": 0.3038,
|
| 1831 |
+
"step": 2850
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 1.84,
|
| 1835 |
+
"learning_rate": 0.00019223452768729642,
|
| 1836 |
+
"loss": 0.3043,
|
| 1837 |
+
"step": 2860
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.85,
|
| 1841 |
+
"learning_rate": 0.00019184364820846902,
|
| 1842 |
+
"loss": 0.3122,
|
| 1843 |
+
"step": 2870
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.85,
|
| 1847 |
+
"learning_rate": 0.00019145276872964168,
|
| 1848 |
+
"loss": 0.2974,
|
| 1849 |
+
"step": 2880
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 1.86,
|
| 1853 |
+
"learning_rate": 0.0001910618892508143,
|
| 1854 |
+
"loss": 0.2962,
|
| 1855 |
+
"step": 2890
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 1.86,
|
| 1859 |
+
"learning_rate": 0.00019067100977198696,
|
| 1860 |
+
"loss": 0.2995,
|
| 1861 |
+
"step": 2900
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 1.87,
|
| 1865 |
+
"learning_rate": 0.00019028013029315962,
|
| 1866 |
+
"loss": 0.3048,
|
| 1867 |
+
"step": 2910
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 1.88,
|
| 1871 |
+
"learning_rate": 0.00018988925081433222,
|
| 1872 |
+
"loss": 0.2995,
|
| 1873 |
+
"step": 2920
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 1.88,
|
| 1877 |
+
"learning_rate": 0.00018949837133550487,
|
| 1878 |
+
"loss": 0.3016,
|
| 1879 |
+
"step": 2930
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.89,
|
| 1883 |
+
"learning_rate": 0.0001891074918566775,
|
| 1884 |
+
"loss": 0.289,
|
| 1885 |
+
"step": 2940
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.9,
|
| 1889 |
+
"learning_rate": 0.00018871661237785016,
|
| 1890 |
+
"loss": 0.3013,
|
| 1891 |
+
"step": 2950
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 1.9,
|
| 1895 |
+
"learning_rate": 0.00018832573289902279,
|
| 1896 |
+
"loss": 0.2949,
|
| 1897 |
+
"step": 2960
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 1.91,
|
| 1901 |
+
"learning_rate": 0.0001879348534201954,
|
| 1902 |
+
"loss": 0.3114,
|
| 1903 |
+
"step": 2970
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 1.92,
|
| 1907 |
+
"learning_rate": 0.00018754397394136804,
|
| 1908 |
+
"loss": 0.3051,
|
| 1909 |
+
"step": 2980
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 1.92,
|
| 1913 |
+
"learning_rate": 0.0001871530944625407,
|
| 1914 |
+
"loss": 0.2993,
|
| 1915 |
+
"step": 2990
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 1.93,
|
| 1919 |
+
"learning_rate": 0.00018676221498371335,
|
| 1920 |
+
"loss": 0.3048,
|
| 1921 |
+
"step": 3000
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.93,
|
| 1925 |
+
"eval_loss": 0.30739831924438477,
|
| 1926 |
+
"eval_runtime": 634.2258,
|
| 1927 |
+
"eval_samples_per_second": 3.153,
|
| 1928 |
+
"eval_steps_per_second": 0.394,
|
| 1929 |
+
"step": 3000
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 1.94,
|
| 1933 |
+
"learning_rate": 0.00018637133550488598,
|
| 1934 |
+
"loss": 0.2921,
|
| 1935 |
+
"step": 3010
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.94,
|
| 1939 |
+
"learning_rate": 0.0001859804560260586,
|
| 1940 |
+
"loss": 0.2976,
|
| 1941 |
+
"step": 3020
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 1.95,
|
| 1945 |
+
"learning_rate": 0.00018558957654723124,
|
| 1946 |
+
"loss": 0.3066,
|
| 1947 |
+
"step": 3030
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 1.95,
|
| 1951 |
+
"learning_rate": 0.0001851986970684039,
|
| 1952 |
+
"loss": 0.2971,
|
| 1953 |
+
"step": 3040
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 1.96,
|
| 1957 |
+
"learning_rate": 0.00018480781758957655,
|
| 1958 |
+
"loss": 0.3051,
|
| 1959 |
+
"step": 3050
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 1.97,
|
| 1963 |
+
"learning_rate": 0.00018441693811074918,
|
| 1964 |
+
"loss": 0.2982,
|
| 1965 |
+
"step": 3060
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.97,
|
| 1969 |
+
"learning_rate": 0.00018402605863192183,
|
| 1970 |
+
"loss": 0.2979,
|
| 1971 |
+
"step": 3070
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 1.98,
|
| 1975 |
+
"learning_rate": 0.00018363517915309443,
|
| 1976 |
+
"loss": 0.3063,
|
| 1977 |
+
"step": 3080
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.99,
|
| 1981 |
+
"learning_rate": 0.0001832442996742671,
|
| 1982 |
+
"loss": 0.3027,
|
| 1983 |
+
"step": 3090
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 1.99,
|
| 1987 |
+
"learning_rate": 0.00018285342019543971,
|
| 1988 |
+
"loss": 0.3117,
|
| 1989 |
+
"step": 3100
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 2.0,
|
| 1993 |
+
"learning_rate": 0.00018246254071661237,
|
| 1994 |
+
"loss": 0.322,
|
| 1995 |
+
"step": 3110
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 2.01,
|
| 1999 |
+
"learning_rate": 0.00018207166123778503,
|
| 2000 |
+
"loss": 0.3016,
|
| 2001 |
+
"step": 3120
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 2.01,
|
| 2005 |
+
"learning_rate": 0.00018168078175895763,
|
| 2006 |
+
"loss": 0.2895,
|
| 2007 |
+
"step": 3130
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 2.02,
|
| 2011 |
+
"learning_rate": 0.00018128990228013028,
|
| 2012 |
+
"loss": 0.2866,
|
| 2013 |
+
"step": 3140
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 2.03,
|
| 2017 |
+
"learning_rate": 0.0001808990228013029,
|
| 2018 |
+
"loss": 0.2965,
|
| 2019 |
+
"step": 3150
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 2.03,
|
| 2023 |
+
"learning_rate": 0.00018050814332247556,
|
| 2024 |
+
"loss": 0.2852,
|
| 2025 |
+
"step": 3160
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 2.04,
|
| 2029 |
+
"learning_rate": 0.00018011726384364822,
|
| 2030 |
+
"loss": 0.2843,
|
| 2031 |
+
"step": 3170
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 2.05,
|
| 2035 |
+
"learning_rate": 0.00017972638436482082,
|
| 2036 |
+
"loss": 0.2762,
|
| 2037 |
+
"step": 3180
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 2.05,
|
| 2041 |
+
"learning_rate": 0.00017933550488599348,
|
| 2042 |
+
"loss": 0.2911,
|
| 2043 |
+
"step": 3190
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 2.06,
|
| 2047 |
+
"learning_rate": 0.0001789446254071661,
|
| 2048 |
+
"loss": 0.2918,
|
| 2049 |
+
"step": 3200
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 2.06,
|
| 2053 |
+
"eval_loss": 0.3086208999156952,
|
| 2054 |
+
"eval_runtime": 609.127,
|
| 2055 |
+
"eval_samples_per_second": 3.283,
|
| 2056 |
+
"eval_steps_per_second": 0.41,
|
| 2057 |
+
"step": 3200
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 2.06,
|
| 2061 |
+
"learning_rate": 0.00017855374592833876,
|
| 2062 |
+
"loss": 0.2897,
|
| 2063 |
+
"step": 3210
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 2.07,
|
| 2067 |
+
"learning_rate": 0.0001781628664495114,
|
| 2068 |
+
"loss": 0.2673,
|
| 2069 |
+
"step": 3220
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 2.08,
|
| 2073 |
+
"learning_rate": 0.00017777198697068402,
|
| 2074 |
+
"loss": 0.2912,
|
| 2075 |
+
"step": 3230
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.08,
|
| 2079 |
+
"learning_rate": 0.00017738110749185664,
|
| 2080 |
+
"loss": 0.2951,
|
| 2081 |
+
"step": 3240
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.09,
|
| 2085 |
+
"learning_rate": 0.0001769902280130293,
|
| 2086 |
+
"loss": 0.2823,
|
| 2087 |
+
"step": 3250
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 2.1,
|
| 2091 |
+
"learning_rate": 0.00017659934853420195,
|
| 2092 |
+
"loss": 0.2888,
|
| 2093 |
+
"step": 3260
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 2.1,
|
| 2097 |
+
"learning_rate": 0.00017620846905537458,
|
| 2098 |
+
"loss": 0.289,
|
| 2099 |
+
"step": 3270
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 2.11,
|
| 2103 |
+
"learning_rate": 0.0001758175895765472,
|
| 2104 |
+
"loss": 0.281,
|
| 2105 |
+
"step": 3280
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 2.12,
|
| 2109 |
+
"learning_rate": 0.00017542671009771984,
|
| 2110 |
+
"loss": 0.2901,
|
| 2111 |
+
"step": 3290
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 2.12,
|
| 2115 |
+
"learning_rate": 0.0001750358306188925,
|
| 2116 |
+
"loss": 0.2858,
|
| 2117 |
+
"step": 3300
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.13,
|
| 2121 |
+
"learning_rate": 0.00017464495114006515,
|
| 2122 |
+
"loss": 0.2945,
|
| 2123 |
+
"step": 3310
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.14,
|
| 2127 |
+
"learning_rate": 0.00017425407166123778,
|
| 2128 |
+
"loss": 0.2858,
|
| 2129 |
+
"step": 3320
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 2.14,
|
| 2133 |
+
"learning_rate": 0.0001738631921824104,
|
| 2134 |
+
"loss": 0.2894,
|
| 2135 |
+
"step": 3330
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 2.15,
|
| 2139 |
+
"learning_rate": 0.00017347231270358303,
|
| 2140 |
+
"loss": 0.294,
|
| 2141 |
+
"step": 3340
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 2.15,
|
| 2145 |
+
"learning_rate": 0.0001730814332247557,
|
| 2146 |
+
"loss": 0.2867,
|
| 2147 |
+
"step": 3350
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 2.16,
|
| 2151 |
+
"learning_rate": 0.00017269055374592832,
|
| 2152 |
+
"loss": 0.2789,
|
| 2153 |
+
"step": 3360
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.17,
|
| 2157 |
+
"learning_rate": 0.00017229967426710097,
|
| 2158 |
+
"loss": 0.2896,
|
| 2159 |
+
"step": 3370
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.17,
|
| 2163 |
+
"learning_rate": 0.00017190879478827357,
|
| 2164 |
+
"loss": 0.2824,
|
| 2165 |
+
"step": 3380
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.18,
|
| 2169 |
+
"learning_rate": 0.00017151791530944623,
|
| 2170 |
+
"loss": 0.2815,
|
| 2171 |
+
"step": 3390
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 2.19,
|
| 2175 |
+
"learning_rate": 0.00017112703583061888,
|
| 2176 |
+
"loss": 0.2894,
|
| 2177 |
+
"step": 3400
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.19,
|
| 2181 |
+
"eval_loss": 0.3088096082210541,
|
| 2182 |
+
"eval_runtime": 609.5049,
|
| 2183 |
+
"eval_samples_per_second": 3.281,
|
| 2184 |
+
"eval_steps_per_second": 0.41,
|
| 2185 |
+
"step": 3400
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 2.19,
|
| 2189 |
+
"learning_rate": 0.0001707361563517915,
|
| 2190 |
+
"loss": 0.2812,
|
| 2191 |
+
"step": 3410
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 2.2,
|
| 2195 |
+
"learning_rate": 0.00017034527687296417,
|
| 2196 |
+
"loss": 0.2806,
|
| 2197 |
+
"step": 3420
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 2.21,
|
| 2201 |
+
"learning_rate": 0.00016995439739413677,
|
| 2202 |
+
"loss": 0.2797,
|
| 2203 |
+
"step": 3430
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 2.21,
|
| 2207 |
+
"learning_rate": 0.00016956351791530942,
|
| 2208 |
+
"loss": 0.282,
|
| 2209 |
+
"step": 3440
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.22,
|
| 2213 |
+
"learning_rate": 0.00016917263843648208,
|
| 2214 |
+
"loss": 0.2969,
|
| 2215 |
+
"step": 3450
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 2.23,
|
| 2219 |
+
"learning_rate": 0.0001687817589576547,
|
| 2220 |
+
"loss": 0.2759,
|
| 2221 |
+
"step": 3460
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 2.23,
|
| 2225 |
+
"learning_rate": 0.00016839087947882736,
|
| 2226 |
+
"loss": 0.2969,
|
| 2227 |
+
"step": 3470
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 2.24,
|
| 2231 |
+
"learning_rate": 0.000168,
|
| 2232 |
+
"loss": 0.2982,
|
| 2233 |
+
"step": 3480
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 2.24,
|
| 2237 |
+
"learning_rate": 0.00016760912052117262,
|
| 2238 |
+
"loss": 0.2932,
|
| 2239 |
+
"step": 3490
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 2.25,
|
| 2243 |
+
"learning_rate": 0.00016721824104234525,
|
| 2244 |
+
"loss": 0.2807,
|
| 2245 |
+
"step": 3500
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 2.26,
|
| 2249 |
+
"learning_rate": 0.0001668273615635179,
|
| 2250 |
+
"loss": 0.2866,
|
| 2251 |
+
"step": 3510
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.26,
|
| 2255 |
+
"learning_rate": 0.00016643648208469056,
|
| 2256 |
+
"loss": 0.2855,
|
| 2257 |
+
"step": 3520
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 2.27,
|
| 2261 |
+
"learning_rate": 0.00016604560260586319,
|
| 2262 |
+
"loss": 0.2839,
|
| 2263 |
+
"step": 3530
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 2.28,
|
| 2267 |
+
"learning_rate": 0.00016565472312703581,
|
| 2268 |
+
"loss": 0.2819,
|
| 2269 |
+
"step": 3540
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 2.28,
|
| 2273 |
+
"learning_rate": 0.00016526384364820844,
|
| 2274 |
+
"loss": 0.2771,
|
| 2275 |
+
"step": 3550
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 2.29,
|
| 2279 |
+
"learning_rate": 0.0001648729641693811,
|
| 2280 |
+
"loss": 0.2833,
|
| 2281 |
+
"step": 3560
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 2.3,
|
| 2285 |
+
"learning_rate": 0.00016448208469055375,
|
| 2286 |
+
"loss": 0.2842,
|
| 2287 |
+
"step": 3570
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 2.3,
|
| 2291 |
+
"learning_rate": 0.00016409120521172638,
|
| 2292 |
+
"loss": 0.2916,
|
| 2293 |
+
"step": 3580
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.31,
|
| 2297 |
+
"learning_rate": 0.000163700325732899,
|
| 2298 |
+
"loss": 0.2942,
|
| 2299 |
+
"step": 3590
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 2.32,
|
| 2303 |
+
"learning_rate": 0.00016330944625407164,
|
| 2304 |
+
"loss": 0.2856,
|
| 2305 |
+
"step": 3600
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 2.32,
|
| 2309 |
+
"eval_loss": 0.3076852560043335,
|
| 2310 |
+
"eval_runtime": 610.2962,
|
| 2311 |
+
"eval_samples_per_second": 3.277,
|
| 2312 |
+
"eval_steps_per_second": 0.41,
|
| 2313 |
+
"step": 3600
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 2.32,
|
| 2317 |
+
"learning_rate": 0.0001629185667752443,
|
| 2318 |
+
"loss": 0.2952,
|
| 2319 |
+
"step": 3610
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 2.33,
|
| 2323 |
+
"learning_rate": 0.00016252768729641692,
|
| 2324 |
+
"loss": 0.2816,
|
| 2325 |
+
"step": 3620
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 2.33,
|
| 2329 |
+
"learning_rate": 0.00016213680781758958,
|
| 2330 |
+
"loss": 0.2958,
|
| 2331 |
+
"step": 3630
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 2.34,
|
| 2335 |
+
"learning_rate": 0.00016174592833876218,
|
| 2336 |
+
"loss": 0.2893,
|
| 2337 |
+
"step": 3640
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 2.35,
|
| 2341 |
+
"learning_rate": 0.00016135504885993483,
|
| 2342 |
+
"loss": 0.2934,
|
| 2343 |
+
"step": 3650
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 2.35,
|
| 2347 |
+
"learning_rate": 0.0001609641693811075,
|
| 2348 |
+
"loss": 0.2885,
|
| 2349 |
+
"step": 3660
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.36,
|
| 2353 |
+
"learning_rate": 0.00016057328990228012,
|
| 2354 |
+
"loss": 0.2789,
|
| 2355 |
+
"step": 3670
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 2.37,
|
| 2359 |
+
"learning_rate": 0.00016018241042345277,
|
| 2360 |
+
"loss": 0.2929,
|
| 2361 |
+
"step": 3680
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 2.37,
|
| 2365 |
+
"learning_rate": 0.00015979153094462537,
|
| 2366 |
+
"loss": 0.2839,
|
| 2367 |
+
"step": 3690
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 2.38,
|
| 2371 |
+
"learning_rate": 0.00015940065146579803,
|
| 2372 |
+
"loss": 0.2901,
|
| 2373 |
+
"step": 3700
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 2.39,
|
| 2377 |
+
"learning_rate": 0.00015900977198697068,
|
| 2378 |
+
"loss": 0.291,
|
| 2379 |
+
"step": 3710
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 2.39,
|
| 2383 |
+
"learning_rate": 0.0001586188925081433,
|
| 2384 |
+
"loss": 0.2944,
|
| 2385 |
+
"step": 3720
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 2.4,
|
| 2389 |
+
"learning_rate": 0.00015822801302931597,
|
| 2390 |
+
"loss": 0.2922,
|
| 2391 |
+
"step": 3730
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.41,
|
| 2395 |
+
"learning_rate": 0.00015783713355048857,
|
| 2396 |
+
"loss": 0.2836,
|
| 2397 |
+
"step": 3740
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 2.41,
|
| 2401 |
+
"learning_rate": 0.00015744625407166122,
|
| 2402 |
+
"loss": 0.2897,
|
| 2403 |
+
"step": 3750
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 2.42,
|
| 2407 |
+
"learning_rate": 0.00015705537459283385,
|
| 2408 |
+
"loss": 0.2896,
|
| 2409 |
+
"step": 3760
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 2.42,
|
| 2413 |
+
"learning_rate": 0.0001566644951140065,
|
| 2414 |
+
"loss": 0.2858,
|
| 2415 |
+
"step": 3770
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 2.43,
|
| 2419 |
+
"learning_rate": 0.00015627361563517916,
|
| 2420 |
+
"loss": 0.2922,
|
| 2421 |
+
"step": 3780
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 2.44,
|
| 2425 |
+
"learning_rate": 0.00015588273615635176,
|
| 2426 |
+
"loss": 0.2838,
|
| 2427 |
+
"step": 3790
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 2.44,
|
| 2431 |
+
"learning_rate": 0.00015549185667752442,
|
| 2432 |
+
"loss": 0.2812,
|
| 2433 |
+
"step": 3800
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.44,
|
| 2437 |
+
"eval_loss": 0.30719852447509766,
|
| 2438 |
+
"eval_runtime": 610.1393,
|
| 2439 |
+
"eval_samples_per_second": 3.278,
|
| 2440 |
+
"eval_steps_per_second": 0.41,
|
| 2441 |
+
"step": 3800
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.45,
|
| 2445 |
+
"learning_rate": 0.00015510097719869704,
|
| 2446 |
+
"loss": 0.2878,
|
| 2447 |
+
"step": 3810
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.46,
|
| 2451 |
+
"learning_rate": 0.0001547100977198697,
|
| 2452 |
+
"loss": 0.2857,
|
| 2453 |
+
"step": 3820
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.46,
|
| 2457 |
+
"learning_rate": 0.00015431921824104236,
|
| 2458 |
+
"loss": 0.2892,
|
| 2459 |
+
"step": 3830
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.47,
|
| 2463 |
+
"learning_rate": 0.00015392833876221498,
|
| 2464 |
+
"loss": 0.2982,
|
| 2465 |
+
"step": 3840
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.48,
|
| 2469 |
+
"learning_rate": 0.0001535374592833876,
|
| 2470 |
+
"loss": 0.2944,
|
| 2471 |
+
"step": 3850
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.48,
|
| 2475 |
+
"learning_rate": 0.00015314657980456024,
|
| 2476 |
+
"loss": 0.2853,
|
| 2477 |
+
"step": 3860
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 2.49,
|
| 2481 |
+
"learning_rate": 0.0001527557003257329,
|
| 2482 |
+
"loss": 0.2739,
|
| 2483 |
+
"step": 3870
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 2.5,
|
| 2487 |
+
"learning_rate": 0.00015236482084690552,
|
| 2488 |
+
"loss": 0.2928,
|
| 2489 |
+
"step": 3880
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.5,
|
| 2493 |
+
"learning_rate": 0.00015197394136807818,
|
| 2494 |
+
"loss": 0.2797,
|
| 2495 |
+
"step": 3890
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 2.51,
|
| 2499 |
+
"learning_rate": 0.00015158306188925078,
|
| 2500 |
+
"loss": 0.2787,
|
| 2501 |
+
"step": 3900
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 2.51,
|
| 2505 |
+
"learning_rate": 0.00015119218241042343,
|
| 2506 |
+
"loss": 0.2877,
|
| 2507 |
+
"step": 3910
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 2.52,
|
| 2511 |
+
"learning_rate": 0.0001508013029315961,
|
| 2512 |
+
"loss": 0.2879,
|
| 2513 |
+
"step": 3920
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 2.53,
|
| 2517 |
+
"learning_rate": 0.00015041042345276872,
|
| 2518 |
+
"loss": 0.2836,
|
| 2519 |
+
"step": 3930
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 2.53,
|
| 2523 |
+
"learning_rate": 0.00015001954397394137,
|
| 2524 |
+
"loss": 0.283,
|
| 2525 |
+
"step": 3940
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 2.54,
|
| 2529 |
+
"learning_rate": 0.000149628664495114,
|
| 2530 |
+
"loss": 0.2966,
|
| 2531 |
+
"step": 3950
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.55,
|
| 2535 |
+
"learning_rate": 0.00014923778501628663,
|
| 2536 |
+
"loss": 0.2896,
|
| 2537 |
+
"step": 3960
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 2.55,
|
| 2541 |
+
"learning_rate": 0.00014884690553745928,
|
| 2542 |
+
"loss": 0.2852,
|
| 2543 |
+
"step": 3970
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 2.56,
|
| 2547 |
+
"learning_rate": 0.0001484560260586319,
|
| 2548 |
+
"loss": 0.2818,
|
| 2549 |
+
"step": 3980
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 2.57,
|
| 2553 |
+
"learning_rate": 0.00014806514657980454,
|
| 2554 |
+
"loss": 0.2874,
|
| 2555 |
+
"step": 3990
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 2.57,
|
| 2559 |
+
"learning_rate": 0.0001476742671009772,
|
| 2560 |
+
"loss": 0.2811,
|
| 2561 |
+
"step": 4000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 2.57,
|
| 2565 |
+
"eval_loss": 0.3067513406276703,
|
| 2566 |
+
"eval_runtime": 607.4662,
|
| 2567 |
+
"eval_samples_per_second": 3.292,
|
| 2568 |
+
"eval_steps_per_second": 0.412,
|
| 2569 |
+
"step": 4000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 2.58,
|
| 2573 |
+
"learning_rate": 0.00014728338762214982,
|
| 2574 |
+
"loss": 0.2942,
|
| 2575 |
+
"step": 4010
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 2.59,
|
| 2579 |
+
"learning_rate": 0.00014689250814332245,
|
| 2580 |
+
"loss": 0.2932,
|
| 2581 |
+
"step": 4020
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 2.59,
|
| 2585 |
+
"learning_rate": 0.0001465016286644951,
|
| 2586 |
+
"loss": 0.2794,
|
| 2587 |
+
"step": 4030
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.6,
|
| 2591 |
+
"learning_rate": 0.00014611074918566774,
|
| 2592 |
+
"loss": 0.2891,
|
| 2593 |
+
"step": 4040
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 2.6,
|
| 2597 |
+
"learning_rate": 0.0001457198697068404,
|
| 2598 |
+
"loss": 0.2772,
|
| 2599 |
+
"step": 4050
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 2.61,
|
| 2603 |
+
"learning_rate": 0.00014532899022801302,
|
| 2604 |
+
"loss": 0.2996,
|
| 2605 |
+
"step": 4060
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 2.62,
|
| 2609 |
+
"learning_rate": 0.00014493811074918565,
|
| 2610 |
+
"loss": 0.2849,
|
| 2611 |
+
"step": 4070
|
| 2612 |
+
},
|
| 2613 |
+
{
|
| 2614 |
+
"epoch": 2.62,
|
| 2615 |
+
"learning_rate": 0.0001445472312703583,
|
| 2616 |
+
"loss": 0.2769,
|
| 2617 |
+
"step": 4080
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 2.63,
|
| 2621 |
+
"learning_rate": 0.00014415635179153093,
|
| 2622 |
+
"loss": 0.2926,
|
| 2623 |
+
"step": 4090
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 2.64,
|
| 2627 |
+
"learning_rate": 0.00014376547231270356,
|
| 2628 |
+
"loss": 0.2892,
|
| 2629 |
+
"step": 4100
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.64,
|
| 2633 |
+
"learning_rate": 0.00014337459283387621,
|
| 2634 |
+
"loss": 0.2776,
|
| 2635 |
+
"step": 4110
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 2.65,
|
| 2639 |
+
"learning_rate": 0.00014298371335504884,
|
| 2640 |
+
"loss": 0.2938,
|
| 2641 |
+
"step": 4120
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 2.66,
|
| 2645 |
+
"learning_rate": 0.0001425928338762215,
|
| 2646 |
+
"loss": 0.2866,
|
| 2647 |
+
"step": 4130
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 2.66,
|
| 2651 |
+
"learning_rate": 0.00014220195439739413,
|
| 2652 |
+
"loss": 0.2863,
|
| 2653 |
+
"step": 4140
|
| 2654 |
+
},
|
| 2655 |
+
{
|
| 2656 |
+
"epoch": 2.67,
|
| 2657 |
+
"learning_rate": 0.00014181107491856675,
|
| 2658 |
+
"loss": 0.2883,
|
| 2659 |
+
"step": 4150
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 2.68,
|
| 2663 |
+
"learning_rate": 0.0001414201954397394,
|
| 2664 |
+
"loss": 0.2875,
|
| 2665 |
+
"step": 4160
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 2.68,
|
| 2669 |
+
"learning_rate": 0.00014102931596091204,
|
| 2670 |
+
"loss": 0.2879,
|
| 2671 |
+
"step": 4170
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.69,
|
| 2675 |
+
"learning_rate": 0.0001406384364820847,
|
| 2676 |
+
"loss": 0.2956,
|
| 2677 |
+
"step": 4180
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 2.69,
|
| 2681 |
+
"learning_rate": 0.00014024755700325732,
|
| 2682 |
+
"loss": 0.286,
|
| 2683 |
+
"step": 4190
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 2.7,
|
| 2687 |
+
"learning_rate": 0.00013985667752442995,
|
| 2688 |
+
"loss": 0.2856,
|
| 2689 |
+
"step": 4200
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 2.7,
|
| 2693 |
+
"eval_loss": 0.3060224652290344,
|
| 2694 |
+
"eval_runtime": 607.0774,
|
| 2695 |
+
"eval_samples_per_second": 3.294,
|
| 2696 |
+
"eval_steps_per_second": 0.412,
|
| 2697 |
+
"step": 4200
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 2.71,
|
| 2701 |
+
"learning_rate": 0.0001394657980456026,
|
| 2702 |
+
"loss": 0.2849,
|
| 2703 |
+
"step": 4210
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 2.71,
|
| 2707 |
+
"learning_rate": 0.00013907491856677523,
|
| 2708 |
+
"loss": 0.2956,
|
| 2709 |
+
"step": 4220
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 2.72,
|
| 2713 |
+
"learning_rate": 0.00013868403908794786,
|
| 2714 |
+
"loss": 0.2815,
|
| 2715 |
+
"step": 4230
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 2.73,
|
| 2719 |
+
"learning_rate": 0.0001382931596091205,
|
| 2720 |
+
"loss": 0.2909,
|
| 2721 |
+
"step": 4240
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 2.73,
|
| 2725 |
+
"learning_rate": 0.00013790228013029314,
|
| 2726 |
+
"loss": 0.2781,
|
| 2727 |
+
"step": 4250
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.74,
|
| 2731 |
+
"learning_rate": 0.0001375114006514658,
|
| 2732 |
+
"loss": 0.2818,
|
| 2733 |
+
"step": 4260
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 2.75,
|
| 2737 |
+
"learning_rate": 0.00013712052117263843,
|
| 2738 |
+
"loss": 0.2777,
|
| 2739 |
+
"step": 4270
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 2.75,
|
| 2743 |
+
"learning_rate": 0.00013672964169381106,
|
| 2744 |
+
"loss": 0.2858,
|
| 2745 |
+
"step": 4280
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 2.76,
|
| 2749 |
+
"learning_rate": 0.00013633876221498368,
|
| 2750 |
+
"loss": 0.2786,
|
| 2751 |
+
"step": 4290
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 2.77,
|
| 2755 |
+
"learning_rate": 0.00013594788273615634,
|
| 2756 |
+
"loss": 0.2876,
|
| 2757 |
+
"step": 4300
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 2.77,
|
| 2761 |
+
"learning_rate": 0.000135557003257329,
|
| 2762 |
+
"loss": 0.2877,
|
| 2763 |
+
"step": 4310
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 2.78,
|
| 2767 |
+
"learning_rate": 0.00013516612377850162,
|
| 2768 |
+
"loss": 0.2906,
|
| 2769 |
+
"step": 4320
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.78,
|
| 2773 |
+
"learning_rate": 0.00013477524429967425,
|
| 2774 |
+
"loss": 0.2836,
|
| 2775 |
+
"step": 4330
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 2.79,
|
| 2779 |
+
"learning_rate": 0.0001343843648208469,
|
| 2780 |
+
"loss": 0.2777,
|
| 2781 |
+
"step": 4340
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.8,
|
| 2785 |
+
"learning_rate": 0.00013399348534201953,
|
| 2786 |
+
"loss": 0.2878,
|
| 2787 |
+
"step": 4350
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 2.8,
|
| 2791 |
+
"learning_rate": 0.00013360260586319216,
|
| 2792 |
+
"loss": 0.2831,
|
| 2793 |
+
"step": 4360
|
| 2794 |
+
},
|
| 2795 |
+
{
|
| 2796 |
+
"epoch": 2.81,
|
| 2797 |
+
"learning_rate": 0.0001332117263843648,
|
| 2798 |
+
"loss": 0.2877,
|
| 2799 |
+
"step": 4370
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 2.82,
|
| 2803 |
+
"learning_rate": 0.00013282084690553744,
|
| 2804 |
+
"loss": 0.2818,
|
| 2805 |
+
"step": 4380
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 2.82,
|
| 2809 |
+
"learning_rate": 0.0001324299674267101,
|
| 2810 |
+
"loss": 0.2796,
|
| 2811 |
+
"step": 4390
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.83,
|
| 2815 |
+
"learning_rate": 0.00013203908794788273,
|
| 2816 |
+
"loss": 0.2887,
|
| 2817 |
+
"step": 4400
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 2.83,
|
| 2821 |
+
"eval_loss": 0.30531054735183716,
|
| 2822 |
+
"eval_runtime": 612.1078,
|
| 2823 |
+
"eval_samples_per_second": 3.267,
|
| 2824 |
+
"eval_steps_per_second": 0.408,
|
| 2825 |
+
"step": 4400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.84,
|
| 2829 |
+
"learning_rate": 0.00013164820846905536,
|
| 2830 |
+
"loss": 0.2897,
|
| 2831 |
+
"step": 4410
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 2.84,
|
| 2835 |
+
"learning_rate": 0.00013125732899022798,
|
| 2836 |
+
"loss": 0.2839,
|
| 2837 |
+
"step": 4420
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 2.85,
|
| 2841 |
+
"learning_rate": 0.00013086644951140064,
|
| 2842 |
+
"loss": 0.2922,
|
| 2843 |
+
"step": 4430
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 2.86,
|
| 2847 |
+
"learning_rate": 0.0001304755700325733,
|
| 2848 |
+
"loss": 0.2794,
|
| 2849 |
+
"step": 4440
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 2.86,
|
| 2853 |
+
"learning_rate": 0.00013008469055374592,
|
| 2854 |
+
"loss": 0.2863,
|
| 2855 |
+
"step": 4450
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 2.87,
|
| 2859 |
+
"learning_rate": 0.00012969381107491855,
|
| 2860 |
+
"loss": 0.286,
|
| 2861 |
+
"step": 4460
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 2.87,
|
| 2865 |
+
"learning_rate": 0.00012930293159609118,
|
| 2866 |
+
"loss": 0.2956,
|
| 2867 |
+
"step": 4470
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.88,
|
| 2871 |
+
"learning_rate": 0.00012891205211726383,
|
| 2872 |
+
"loss": 0.2838,
|
| 2873 |
+
"step": 4480
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 2.89,
|
| 2877 |
+
"learning_rate": 0.00012852117263843646,
|
| 2878 |
+
"loss": 0.2791,
|
| 2879 |
+
"step": 4490
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 2.89,
|
| 2883 |
+
"learning_rate": 0.0001281302931596091,
|
| 2884 |
+
"loss": 0.2778,
|
| 2885 |
+
"step": 4500
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 2.9,
|
| 2889 |
+
"learning_rate": 0.00012773941368078175,
|
| 2890 |
+
"loss": 0.2813,
|
| 2891 |
+
"step": 4510
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 2.91,
|
| 2895 |
+
"learning_rate": 0.0001273485342019544,
|
| 2896 |
+
"loss": 0.2824,
|
| 2897 |
+
"step": 4520
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 2.91,
|
| 2901 |
+
"learning_rate": 0.00012695765472312703,
|
| 2902 |
+
"loss": 0.2735,
|
| 2903 |
+
"step": 4530
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 2.92,
|
| 2907 |
+
"learning_rate": 0.00012656677524429966,
|
| 2908 |
+
"loss": 0.2796,
|
| 2909 |
+
"step": 4540
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.93,
|
| 2913 |
+
"learning_rate": 0.00012617589576547229,
|
| 2914 |
+
"loss": 0.2813,
|
| 2915 |
+
"step": 4550
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 2.93,
|
| 2919 |
+
"learning_rate": 0.00012578501628664494,
|
| 2920 |
+
"loss": 0.2847,
|
| 2921 |
+
"step": 4560
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.94,
|
| 2925 |
+
"learning_rate": 0.0001253941368078176,
|
| 2926 |
+
"loss": 0.2927,
|
| 2927 |
+
"step": 4570
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 2.95,
|
| 2931 |
+
"learning_rate": 0.00012500325732899022,
|
| 2932 |
+
"loss": 0.2903,
|
| 2933 |
+
"step": 4580
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 2.95,
|
| 2937 |
+
"learning_rate": 0.00012461237785016285,
|
| 2938 |
+
"loss": 0.2872,
|
| 2939 |
+
"step": 4590
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 2.96,
|
| 2943 |
+
"learning_rate": 0.00012422149837133548,
|
| 2944 |
+
"loss": 0.2761,
|
| 2945 |
+
"step": 4600
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 2.96,
|
| 2949 |
+
"eval_loss": 0.3047821521759033,
|
| 2950 |
+
"eval_runtime": 608.7677,
|
| 2951 |
+
"eval_samples_per_second": 3.285,
|
| 2952 |
+
"eval_steps_per_second": 0.411,
|
| 2953 |
+
"step": 4600
|
| 2954 |
+
}
|
| 2955 |
+
],
|
| 2956 |
+
"logging_steps": 10,
|
| 2957 |
+
"max_steps": 7775,
|
| 2958 |
+
"num_train_epochs": 5,
|
| 2959 |
+
"save_steps": 200,
|
| 2960 |
+
"total_flos": 4.1682312131139994e+18,
|
| 2961 |
+
"trial_name": null,
|
| 2962 |
+
"trial_params": null
|
| 2963 |
+
}
|
with_input/decomp_code_with_intermediates/llama-2-7b/best_model/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab07d50472e198dde14f28aa7ad4f294a27b19676460e69c8353014486305e3f
|
| 3 |
+
size 4091
|