Training in progress, step 3500, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 737580392
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c03dfe3ff98720b641d5b3253f189443475f90c5848bfce1ee42b4e25e9a06d9
|
| 3 |
size 737580392
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1475248442
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:781ca001e4eef0894d5dc0a043ec1d7414e5f687b44a3bb27578a66df794e142
|
| 3 |
size 1475248442
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8cd02421b7ec256714ec03c37d51589e92544068eeda4bae107d407e8dfd0cb9
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8785b8509dc9a197581e45af973f623b343ec6de3eb0eeab89b29a64ed0e10d5
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 1.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2587,6 +2587,436 @@
|
|
| 2587 |
"eval_spearman_manhattan": 0.8242741076199507,
|
| 2588 |
"eval_steps_per_second": 15.409,
|
| 2589 |
"step": 3000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2590 |
}
|
| 2591 |
],
|
| 2592 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.640112464854733,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 3500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2587 |
"eval_spearman_manhattan": 0.8242741076199507,
|
| 2588 |
"eval_steps_per_second": 15.409,
|
| 2589 |
"step": 3000
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 1.4104967197750704,
|
| 2593 |
+
"grad_norm": 1.681735873222351,
|
| 2594 |
+
"learning_rate": 4.559219775070291e-05,
|
| 2595 |
+
"loss": 0.2028,
|
| 2596 |
+
"step": 3010
|
| 2597 |
+
},
|
| 2598 |
+
{
|
| 2599 |
+
"epoch": 1.415182755388941,
|
| 2600 |
+
"grad_norm": 1.349747896194458,
|
| 2601 |
+
"learning_rate": 4.5577553889409565e-05,
|
| 2602 |
+
"loss": 0.1849,
|
| 2603 |
+
"step": 3020
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 1.4198687910028116,
|
| 2607 |
+
"grad_norm": 1.752896785736084,
|
| 2608 |
+
"learning_rate": 4.556291002811622e-05,
|
| 2609 |
+
"loss": 0.1939,
|
| 2610 |
+
"step": 3030
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 1.4245548266166823,
|
| 2614 |
+
"grad_norm": 1.538013219833374,
|
| 2615 |
+
"learning_rate": 4.5548266166822874e-05,
|
| 2616 |
+
"loss": 0.1711,
|
| 2617 |
+
"step": 3040
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 1.429240862230553,
|
| 2621 |
+
"grad_norm": 1.6929601430892944,
|
| 2622 |
+
"learning_rate": 4.553362230552952e-05,
|
| 2623 |
+
"loss": 0.1835,
|
| 2624 |
+
"step": 3050
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 1.4339268978444237,
|
| 2628 |
+
"grad_norm": 1.9702001810073853,
|
| 2629 |
+
"learning_rate": 4.5518978444236176e-05,
|
| 2630 |
+
"loss": 0.1968,
|
| 2631 |
+
"step": 3060
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 1.4386129334582942,
|
| 2635 |
+
"grad_norm": 1.5056127309799194,
|
| 2636 |
+
"learning_rate": 4.550433458294283e-05,
|
| 2637 |
+
"loss": 0.1856,
|
| 2638 |
+
"step": 3070
|
| 2639 |
+
},
|
| 2640 |
+
{
|
| 2641 |
+
"epoch": 1.443298969072165,
|
| 2642 |
+
"grad_norm": 1.8639978170394897,
|
| 2643 |
+
"learning_rate": 4.5489690721649484e-05,
|
| 2644 |
+
"loss": 0.1854,
|
| 2645 |
+
"step": 3080
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 1.4479850046860356,
|
| 2649 |
+
"grad_norm": 1.6516441106796265,
|
| 2650 |
+
"learning_rate": 4.547504686035614e-05,
|
| 2651 |
+
"loss": 0.1816,
|
| 2652 |
+
"step": 3090
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 1.4526710402999063,
|
| 2656 |
+
"grad_norm": 1.3151347637176514,
|
| 2657 |
+
"learning_rate": 4.546040299906279e-05,
|
| 2658 |
+
"loss": 0.1777,
|
| 2659 |
+
"step": 3100
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 1.4526710402999063,
|
| 2663 |
+
"eval_loss": 0.036029454320669174,
|
| 2664 |
+
"eval_pearson_cosine": 0.8219421674547647,
|
| 2665 |
+
"eval_pearson_dot": 0.7544857842745216,
|
| 2666 |
+
"eval_pearson_euclidean": 0.8154423230591377,
|
| 2667 |
+
"eval_pearson_manhattan": 0.8169259606030721,
|
| 2668 |
+
"eval_runtime": 5.9655,
|
| 2669 |
+
"eval_samples_per_second": 251.446,
|
| 2670 |
+
"eval_spearman_cosine": 0.8227245969216198,
|
| 2671 |
+
"eval_spearman_dot": 0.7556651515459966,
|
| 2672 |
+
"eval_spearman_euclidean": 0.8221360306356487,
|
| 2673 |
+
"eval_spearman_manhattan": 0.8232106684973721,
|
| 2674 |
+
"eval_steps_per_second": 15.757,
|
| 2675 |
+
"step": 3100
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 1.457357075913777,
|
| 2679 |
+
"grad_norm": 1.5906847715377808,
|
| 2680 |
+
"learning_rate": 4.544575913776945e-05,
|
| 2681 |
+
"loss": 0.2292,
|
| 2682 |
+
"step": 3110
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 1.4620431115276475,
|
| 2686 |
+
"grad_norm": 1.2340494394302368,
|
| 2687 |
+
"learning_rate": 4.543111527647611e-05,
|
| 2688 |
+
"loss": 0.1885,
|
| 2689 |
+
"step": 3120
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 1.4667291471415183,
|
| 2693 |
+
"grad_norm": 1.7530288696289062,
|
| 2694 |
+
"learning_rate": 4.541647141518276e-05,
|
| 2695 |
+
"loss": 0.2088,
|
| 2696 |
+
"step": 3130
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 1.471415182755389,
|
| 2700 |
+
"grad_norm": 1.7493581771850586,
|
| 2701 |
+
"learning_rate": 4.540182755388942e-05,
|
| 2702 |
+
"loss": 0.1704,
|
| 2703 |
+
"step": 3140
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 1.4761012183692597,
|
| 2707 |
+
"grad_norm": 1.5426673889160156,
|
| 2708 |
+
"learning_rate": 4.5387183692596064e-05,
|
| 2709 |
+
"loss": 0.1743,
|
| 2710 |
+
"step": 3150
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 1.4807872539831304,
|
| 2714 |
+
"grad_norm": 1.4274178743362427,
|
| 2715 |
+
"learning_rate": 4.537253983130272e-05,
|
| 2716 |
+
"loss": 0.1929,
|
| 2717 |
+
"step": 3160
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 1.4854732895970009,
|
| 2721 |
+
"grad_norm": 1.372902512550354,
|
| 2722 |
+
"learning_rate": 4.535789597000937e-05,
|
| 2723 |
+
"loss": 0.1913,
|
| 2724 |
+
"step": 3170
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 1.4901593252108716,
|
| 2728 |
+
"grad_norm": 2.073024272918701,
|
| 2729 |
+
"learning_rate": 4.534325210871603e-05,
|
| 2730 |
+
"loss": 0.2104,
|
| 2731 |
+
"step": 3180
|
| 2732 |
+
},
|
| 2733 |
+
{
|
| 2734 |
+
"epoch": 1.4948453608247423,
|
| 2735 |
+
"grad_norm": 1.5448635816574097,
|
| 2736 |
+
"learning_rate": 4.532860824742268e-05,
|
| 2737 |
+
"loss": 0.2018,
|
| 2738 |
+
"step": 3190
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 1.499531396438613,
|
| 2742 |
+
"grad_norm": 1.4171442985534668,
|
| 2743 |
+
"learning_rate": 4.5313964386129336e-05,
|
| 2744 |
+
"loss": 0.1816,
|
| 2745 |
+
"step": 3200
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 1.499531396438613,
|
| 2749 |
+
"eval_loss": 0.036359407007694244,
|
| 2750 |
+
"eval_pearson_cosine": 0.821300031874455,
|
| 2751 |
+
"eval_pearson_dot": 0.7616346380196859,
|
| 2752 |
+
"eval_pearson_euclidean": 0.8169013356741246,
|
| 2753 |
+
"eval_pearson_manhattan": 0.8184860654262422,
|
| 2754 |
+
"eval_runtime": 5.9202,
|
| 2755 |
+
"eval_samples_per_second": 253.369,
|
| 2756 |
+
"eval_spearman_cosine": 0.8228088882125945,
|
| 2757 |
+
"eval_spearman_dot": 0.7590379988222679,
|
| 2758 |
+
"eval_spearman_euclidean": 0.8237353804832064,
|
| 2759 |
+
"eval_spearman_manhattan": 0.82472345338271,
|
| 2760 |
+
"eval_steps_per_second": 15.878,
|
| 2761 |
+
"step": 3200
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 1.5042174320524837,
|
| 2765 |
+
"grad_norm": 1.7904757261276245,
|
| 2766 |
+
"learning_rate": 4.529932052483599e-05,
|
| 2767 |
+
"loss": 0.1955,
|
| 2768 |
+
"step": 3210
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 1.5089034676663542,
|
| 2772 |
+
"grad_norm": 2.021733522415161,
|
| 2773 |
+
"learning_rate": 4.5284676663542644e-05,
|
| 2774 |
+
"loss": 0.1776,
|
| 2775 |
+
"step": 3220
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 1.513589503280225,
|
| 2779 |
+
"grad_norm": 1.2433106899261475,
|
| 2780 |
+
"learning_rate": 4.52700328022493e-05,
|
| 2781 |
+
"loss": 0.1626,
|
| 2782 |
+
"step": 3230
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 1.5182755388940956,
|
| 2786 |
+
"grad_norm": 1.4370200634002686,
|
| 2787 |
+
"learning_rate": 4.525538894095596e-05,
|
| 2788 |
+
"loss": 0.1752,
|
| 2789 |
+
"step": 3240
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 1.522961574507966,
|
| 2793 |
+
"grad_norm": 1.9471467733383179,
|
| 2794 |
+
"learning_rate": 4.524074507966261e-05,
|
| 2795 |
+
"loss": 0.1782,
|
| 2796 |
+
"step": 3250
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 1.527647610121837,
|
| 2800 |
+
"grad_norm": 1.829440712928772,
|
| 2801 |
+
"learning_rate": 4.522610121836926e-05,
|
| 2802 |
+
"loss": 0.2014,
|
| 2803 |
+
"step": 3260
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 1.5323336457357075,
|
| 2807 |
+
"grad_norm": 1.703355073928833,
|
| 2808 |
+
"learning_rate": 4.5211457357075915e-05,
|
| 2809 |
+
"loss": 0.1762,
|
| 2810 |
+
"step": 3270
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 1.5370196813495782,
|
| 2814 |
+
"grad_norm": 1.6706669330596924,
|
| 2815 |
+
"learning_rate": 4.519681349578257e-05,
|
| 2816 |
+
"loss": 0.1937,
|
| 2817 |
+
"step": 3280
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 1.541705716963449,
|
| 2821 |
+
"grad_norm": 1.5066584348678589,
|
| 2822 |
+
"learning_rate": 4.5182169634489224e-05,
|
| 2823 |
+
"loss": 0.2008,
|
| 2824 |
+
"step": 3290
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 1.5463917525773194,
|
| 2828 |
+
"grad_norm": 1.4767428636550903,
|
| 2829 |
+
"learning_rate": 4.516752577319588e-05,
|
| 2830 |
+
"loss": 0.229,
|
| 2831 |
+
"step": 3300
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 1.5463917525773194,
|
| 2835 |
+
"eval_loss": 0.039627715945243835,
|
| 2836 |
+
"eval_pearson_cosine": 0.8169224788587428,
|
| 2837 |
+
"eval_pearson_dot": 0.7529223671879777,
|
| 2838 |
+
"eval_pearson_euclidean": 0.816545384611743,
|
| 2839 |
+
"eval_pearson_manhattan": 0.8176740235496034,
|
| 2840 |
+
"eval_runtime": 6.324,
|
| 2841 |
+
"eval_samples_per_second": 237.191,
|
| 2842 |
+
"eval_spearman_cosine": 0.8199004477023643,
|
| 2843 |
+
"eval_spearman_dot": 0.7498417362110426,
|
| 2844 |
+
"eval_spearman_euclidean": 0.8235117073588528,
|
| 2845 |
+
"eval_spearman_manhattan": 0.8240841209374519,
|
| 2846 |
+
"eval_steps_per_second": 14.864,
|
| 2847 |
+
"step": 3300
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 1.5510777881911904,
|
| 2851 |
+
"grad_norm": 1.3127154111862183,
|
| 2852 |
+
"learning_rate": 4.515288191190253e-05,
|
| 2853 |
+
"loss": 0.1849,
|
| 2854 |
+
"step": 3310
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 1.5557638238050608,
|
| 2858 |
+
"grad_norm": 0.9424723982810974,
|
| 2859 |
+
"learning_rate": 4.513823805060919e-05,
|
| 2860 |
+
"loss": 0.2011,
|
| 2861 |
+
"step": 3320
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 1.5604498594189316,
|
| 2865 |
+
"grad_norm": 1.585274577140808,
|
| 2866 |
+
"learning_rate": 4.512359418931584e-05,
|
| 2867 |
+
"loss": 0.1935,
|
| 2868 |
+
"step": 3330
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 1.5651358950328023,
|
| 2872 |
+
"grad_norm": 1.4503992795944214,
|
| 2873 |
+
"learning_rate": 4.5108950328022495e-05,
|
| 2874 |
+
"loss": 0.2189,
|
| 2875 |
+
"step": 3340
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 1.569821930646673,
|
| 2879 |
+
"grad_norm": 1.6958515644073486,
|
| 2880 |
+
"learning_rate": 4.509430646672915e-05,
|
| 2881 |
+
"loss": 0.1949,
|
| 2882 |
+
"step": 3350
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 1.5745079662605437,
|
| 2886 |
+
"grad_norm": 1.7165809869766235,
|
| 2887 |
+
"learning_rate": 4.5079662605435804e-05,
|
| 2888 |
+
"loss": 0.1897,
|
| 2889 |
+
"step": 3360
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 1.5791940018744142,
|
| 2893 |
+
"grad_norm": 1.4036378860473633,
|
| 2894 |
+
"learning_rate": 4.506501874414246e-05,
|
| 2895 |
+
"loss": 0.2173,
|
| 2896 |
+
"step": 3370
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 1.5838800374882849,
|
| 2900 |
+
"grad_norm": 1.4913054704666138,
|
| 2901 |
+
"learning_rate": 4.505037488284911e-05,
|
| 2902 |
+
"loss": 0.1931,
|
| 2903 |
+
"step": 3380
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 1.5885660731021556,
|
| 2907 |
+
"grad_norm": 1.7645376920700073,
|
| 2908 |
+
"learning_rate": 4.503573102155577e-05,
|
| 2909 |
+
"loss": 0.201,
|
| 2910 |
+
"step": 3390
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 1.5932521087160263,
|
| 2914 |
+
"grad_norm": 1.2109887599945068,
|
| 2915 |
+
"learning_rate": 4.502108716026242e-05,
|
| 2916 |
+
"loss": 0.1742,
|
| 2917 |
+
"step": 3400
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 1.5932521087160263,
|
| 2921 |
+
"eval_loss": 0.034527041018009186,
|
| 2922 |
+
"eval_pearson_cosine": 0.8244957125194645,
|
| 2923 |
+
"eval_pearson_dot": 0.7646698456086369,
|
| 2924 |
+
"eval_pearson_euclidean": 0.8169295762368449,
|
| 2925 |
+
"eval_pearson_manhattan": 0.8185445852885636,
|
| 2926 |
+
"eval_runtime": 5.9441,
|
| 2927 |
+
"eval_samples_per_second": 252.35,
|
| 2928 |
+
"eval_spearman_cosine": 0.8251998390698203,
|
| 2929 |
+
"eval_spearman_dot": 0.7633691785479392,
|
| 2930 |
+
"eval_spearman_euclidean": 0.8242535656619165,
|
| 2931 |
+
"eval_spearman_manhattan": 0.825256335833904,
|
| 2932 |
+
"eval_steps_per_second": 15.814,
|
| 2933 |
+
"step": 3400
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 1.597938144329897,
|
| 2937 |
+
"grad_norm": 0.9772526621818542,
|
| 2938 |
+
"learning_rate": 4.5006443298969075e-05,
|
| 2939 |
+
"loss": 0.1646,
|
| 2940 |
+
"step": 3410
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 1.6026241799437675,
|
| 2944 |
+
"grad_norm": 1.2620090246200562,
|
| 2945 |
+
"learning_rate": 4.499179943767573e-05,
|
| 2946 |
+
"loss": 0.1849,
|
| 2947 |
+
"step": 3420
|
| 2948 |
+
},
|
| 2949 |
+
{
|
| 2950 |
+
"epoch": 1.6073102155576382,
|
| 2951 |
+
"grad_norm": 1.5649354457855225,
|
| 2952 |
+
"learning_rate": 4.4977155576382384e-05,
|
| 2953 |
+
"loss": 0.2149,
|
| 2954 |
+
"step": 3430
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 1.611996251171509,
|
| 2958 |
+
"grad_norm": 1.3986328840255737,
|
| 2959 |
+
"learning_rate": 4.496251171508904e-05,
|
| 2960 |
+
"loss": 0.1706,
|
| 2961 |
+
"step": 3440
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 1.6166822867853796,
|
| 2965 |
+
"grad_norm": 1.0502641201019287,
|
| 2966 |
+
"learning_rate": 4.4947867853795686e-05,
|
| 2967 |
+
"loss": 0.18,
|
| 2968 |
+
"step": 3450
|
| 2969 |
+
},
|
| 2970 |
+
{
|
| 2971 |
+
"epoch": 1.6213683223992503,
|
| 2972 |
+
"grad_norm": 2.1642649173736572,
|
| 2973 |
+
"learning_rate": 4.4933223992502347e-05,
|
| 2974 |
+
"loss": 0.1903,
|
| 2975 |
+
"step": 3460
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 1.6260543580131208,
|
| 2979 |
+
"grad_norm": 2.1181936264038086,
|
| 2980 |
+
"learning_rate": 4.4918580131209e-05,
|
| 2981 |
+
"loss": 0.2061,
|
| 2982 |
+
"step": 3470
|
| 2983 |
+
},
|
| 2984 |
+
{
|
| 2985 |
+
"epoch": 1.6307403936269915,
|
| 2986 |
+
"grad_norm": 1.52034592628479,
|
| 2987 |
+
"learning_rate": 4.4903936269915655e-05,
|
| 2988 |
+
"loss": 0.1612,
|
| 2989 |
+
"step": 3480
|
| 2990 |
+
},
|
| 2991 |
+
{
|
| 2992 |
+
"epoch": 1.6354264292408622,
|
| 2993 |
+
"grad_norm": 1.6476225852966309,
|
| 2994 |
+
"learning_rate": 4.488929240862231e-05,
|
| 2995 |
+
"loss": 0.1825,
|
| 2996 |
+
"step": 3490
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 1.640112464854733,
|
| 3000 |
+
"grad_norm": 1.46713387966156,
|
| 3001 |
+
"learning_rate": 4.4874648547328964e-05,
|
| 3002 |
+
"loss": 0.1606,
|
| 3003 |
+
"step": 3500
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 1.640112464854733,
|
| 3007 |
+
"eval_loss": 0.03454529866576195,
|
| 3008 |
+
"eval_pearson_cosine": 0.8219067533222528,
|
| 3009 |
+
"eval_pearson_dot": 0.7628975832542579,
|
| 3010 |
+
"eval_pearson_euclidean": 0.8128236856339015,
|
| 3011 |
+
"eval_pearson_manhattan": 0.8145631898507872,
|
| 3012 |
+
"eval_runtime": 5.977,
|
| 3013 |
+
"eval_samples_per_second": 250.962,
|
| 3014 |
+
"eval_spearman_cosine": 0.8230006137618159,
|
| 3015 |
+
"eval_spearman_dot": 0.7622148554080955,
|
| 3016 |
+
"eval_spearman_euclidean": 0.8212688136914371,
|
| 3017 |
+
"eval_spearman_manhattan": 0.8222685344671697,
|
| 3018 |
+
"eval_steps_per_second": 15.727,
|
| 3019 |
+
"step": 3500
|
| 3020 |
}
|
| 3021 |
],
|
| 3022 |
"logging_steps": 10,
|