Training in progress, epoch 3, 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 3132668808
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50fc49030561fa37359d41c9429af925790f6ff15c62bda88dbf4bf5e674d823
|
| 3 |
size 3132668808
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6265677800
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c6cfb494ef985708c25be871e5f5af87233127e2bd5707124b7933dcd1f91f1
|
| 3 |
size 6265677800
|
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:c2c8ba8be617b22bef6eaefcd9296941e895893aba5d37eef6005f01420e156e
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:206e95878ece1a626a5fd7465b45e403240308793212a55f6ccb557776f398f1
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1677,13 +1677,854 @@
|
|
| 1677 |
"eval_steps_per_second": 0.423,
|
| 1678 |
"eval_translation_length": 7761920,
|
| 1679 |
"step": 136438
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1680 |
}
|
| 1681 |
],
|
| 1682 |
"logging_steps": 500,
|
| 1683 |
"max_steps": 341095,
|
| 1684 |
"num_train_epochs": 5,
|
| 1685 |
"save_steps": 5000,
|
| 1686 |
-
"total_flos":
|
| 1687 |
"trial_name": null,
|
| 1688 |
"trial_params": null
|
| 1689 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 204657,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1677 |
"eval_steps_per_second": 0.423,
|
| 1678 |
"eval_translation_length": 7761920,
|
| 1679 |
"step": 136438
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 2.0,
|
| 1683 |
+
"learning_rate": 5.2444163076084484e-05,
|
| 1684 |
+
"loss": 2.1805,
|
| 1685 |
+
"step": 136500
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 2.01,
|
| 1689 |
+
"learning_rate": 5.226870912326529e-05,
|
| 1690 |
+
"loss": 1.8318,
|
| 1691 |
+
"step": 137000
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 2.02,
|
| 1695 |
+
"learning_rate": 5.209299421803256e-05,
|
| 1696 |
+
"loss": 1.7868,
|
| 1697 |
+
"step": 137500
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 2.02,
|
| 1701 |
+
"learning_rate": 5.1917022097798964e-05,
|
| 1702 |
+
"loss": 1.8616,
|
| 1703 |
+
"step": 138000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 2.03,
|
| 1707 |
+
"learning_rate": 5.174079650544804e-05,
|
| 1708 |
+
"loss": 1.862,
|
| 1709 |
+
"step": 138500
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 2.04,
|
| 1713 |
+
"learning_rate": 5.1564321189254605e-05,
|
| 1714 |
+
"loss": 1.7721,
|
| 1715 |
+
"step": 139000
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 2.04,
|
| 1719 |
+
"learning_rate": 5.138759990280507e-05,
|
| 1720 |
+
"loss": 1.8478,
|
| 1721 |
+
"step": 139500
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 2.05,
|
| 1725 |
+
"learning_rate": 5.1210636404917554e-05,
|
| 1726 |
+
"loss": 1.935,
|
| 1727 |
+
"step": 140000
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 2.06,
|
| 1731 |
+
"learning_rate": 5.1033434459561965e-05,
|
| 1732 |
+
"loss": 1.8431,
|
| 1733 |
+
"step": 140500
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 2.07,
|
| 1737 |
+
"learning_rate": 5.0855997835779926e-05,
|
| 1738 |
+
"loss": 1.8654,
|
| 1739 |
+
"step": 141000
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.07,
|
| 1743 |
+
"learning_rate": 5.067833030760462e-05,
|
| 1744 |
+
"loss": 1.7853,
|
| 1745 |
+
"step": 141500
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 2.08,
|
| 1749 |
+
"learning_rate": 5.0500435653980455e-05,
|
| 1750 |
+
"loss": 1.8631,
|
| 1751 |
+
"step": 142000
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 2.09,
|
| 1755 |
+
"learning_rate": 5.032231765868284e-05,
|
| 1756 |
+
"loss": 1.8427,
|
| 1757 |
+
"step": 142500
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 2.1,
|
| 1761 |
+
"learning_rate": 5.014398011023752e-05,
|
| 1762 |
+
"loss": 1.8528,
|
| 1763 |
+
"step": 143000
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 2.1,
|
| 1767 |
+
"learning_rate": 4.9965426801840127e-05,
|
| 1768 |
+
"loss": 1.8048,
|
| 1769 |
+
"step": 143500
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 2.11,
|
| 1773 |
+
"learning_rate": 4.9786661531275425e-05,
|
| 1774 |
+
"loss": 1.8064,
|
| 1775 |
+
"step": 144000
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 2.12,
|
| 1779 |
+
"learning_rate": 4.96076881008366e-05,
|
| 1780 |
+
"loss": 1.9335,
|
| 1781 |
+
"step": 144500
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.13,
|
| 1785 |
+
"learning_rate": 4.9428510317244295e-05,
|
| 1786 |
+
"loss": 1.9127,
|
| 1787 |
+
"step": 145000
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 2.13,
|
| 1791 |
+
"learning_rate": 4.924913199156575e-05,
|
| 1792 |
+
"loss": 1.8177,
|
| 1793 |
+
"step": 145500
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 2.14,
|
| 1797 |
+
"learning_rate": 4.906955693913363e-05,
|
| 1798 |
+
"loss": 1.8531,
|
| 1799 |
+
"step": 146000
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.15,
|
| 1803 |
+
"learning_rate": 4.8889788979465e-05,
|
| 1804 |
+
"loss": 1.8327,
|
| 1805 |
+
"step": 146500
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.15,
|
| 1809 |
+
"learning_rate": 4.870983193617993e-05,
|
| 1810 |
+
"loss": 1.8751,
|
| 1811 |
+
"step": 147000
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 2.16,
|
| 1815 |
+
"learning_rate": 4.8529689636920306e-05,
|
| 1816 |
+
"loss": 1.865,
|
| 1817 |
+
"step": 147500
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 2.17,
|
| 1821 |
+
"learning_rate": 4.834936591326834e-05,
|
| 1822 |
+
"loss": 1.9178,
|
| 1823 |
+
"step": 148000
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.18,
|
| 1827 |
+
"learning_rate": 4.8168864600665134e-05,
|
| 1828 |
+
"loss": 1.8717,
|
| 1829 |
+
"step": 148500
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.18,
|
| 1833 |
+
"learning_rate": 4.798818953832902e-05,
|
| 1834 |
+
"loss": 1.8991,
|
| 1835 |
+
"step": 149000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 2.19,
|
| 1839 |
+
"learning_rate": 4.780734456917396e-05,
|
| 1840 |
+
"loss": 1.8914,
|
| 1841 |
+
"step": 149500
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.2,
|
| 1845 |
+
"learning_rate": 4.7626333539727793e-05,
|
| 1846 |
+
"loss": 1.9122,
|
| 1847 |
+
"step": 150000
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.21,
|
| 1851 |
+
"learning_rate": 4.744516030005041e-05,
|
| 1852 |
+
"loss": 1.8922,
|
| 1853 |
+
"step": 150500
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 2.21,
|
| 1857 |
+
"learning_rate": 4.72638287036519e-05,
|
| 1858 |
+
"loss": 1.929,
|
| 1859 |
+
"step": 151000
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 2.22,
|
| 1863 |
+
"learning_rate": 4.7082342607410514e-05,
|
| 1864 |
+
"loss": 1.8535,
|
| 1865 |
+
"step": 151500
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 2.23,
|
| 1869 |
+
"learning_rate": 4.6900705871490736e-05,
|
| 1870 |
+
"loss": 1.8546,
|
| 1871 |
+
"step": 152000
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 2.24,
|
| 1875 |
+
"learning_rate": 4.671892235926106e-05,
|
| 1876 |
+
"loss": 1.8847,
|
| 1877 |
+
"step": 152500
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 2.24,
|
| 1881 |
+
"learning_rate": 4.653699593721192e-05,
|
| 1882 |
+
"loss": 1.8957,
|
| 1883 |
+
"step": 153000
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 2.25,
|
| 1887 |
+
"learning_rate": 4.635493047487339e-05,
|
| 1888 |
+
"loss": 1.909,
|
| 1889 |
+
"step": 153500
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 2.26,
|
| 1893 |
+
"learning_rate": 4.61727298447329e-05,
|
| 1894 |
+
"loss": 1.8491,
|
| 1895 |
+
"step": 154000
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 2.26,
|
| 1899 |
+
"learning_rate": 4.5990397922152834e-05,
|
| 1900 |
+
"loss": 1.9239,
|
| 1901 |
+
"step": 154500
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 2.27,
|
| 1905 |
+
"learning_rate": 4.580793858528819e-05,
|
| 1906 |
+
"loss": 1.893,
|
| 1907 |
+
"step": 155000
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 2.28,
|
| 1911 |
+
"learning_rate": 4.562535571500399e-05,
|
| 1912 |
+
"loss": 1.8956,
|
| 1913 |
+
"step": 155500
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 2.29,
|
| 1917 |
+
"learning_rate": 4.5442653194792806e-05,
|
| 1918 |
+
"loss": 1.9018,
|
| 1919 |
+
"step": 156000
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 2.29,
|
| 1923 |
+
"learning_rate": 4.525983491069215e-05,
|
| 1924 |
+
"loss": 1.8923,
|
| 1925 |
+
"step": 156500
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 2.3,
|
| 1929 |
+
"learning_rate": 4.507690475120176e-05,
|
| 1930 |
+
"loss": 1.8883,
|
| 1931 |
+
"step": 157000
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 2.31,
|
| 1935 |
+
"learning_rate": 4.489386660720095e-05,
|
| 1936 |
+
"loss": 1.8547,
|
| 1937 |
+
"step": 157500
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 2.32,
|
| 1941 |
+
"learning_rate": 4.471072437186585e-05,
|
| 1942 |
+
"loss": 1.8442,
|
| 1943 |
+
"step": 158000
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 2.32,
|
| 1947 |
+
"learning_rate": 4.452748194058659e-05,
|
| 1948 |
+
"loss": 1.8919,
|
| 1949 |
+
"step": 158500
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 2.33,
|
| 1953 |
+
"learning_rate": 4.4344143210884436e-05,
|
| 1954 |
+
"loss": 1.8636,
|
| 1955 |
+
"step": 159000
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 2.34,
|
| 1959 |
+
"learning_rate": 4.416071208232888e-05,
|
| 1960 |
+
"loss": 1.8752,
|
| 1961 |
+
"step": 159500
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 2.35,
|
| 1965 |
+
"learning_rate": 4.397719245645475e-05,
|
| 1966 |
+
"loss": 1.8668,
|
| 1967 |
+
"step": 160000
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 2.35,
|
| 1971 |
+
"learning_rate": 4.3793588236679173e-05,
|
| 1972 |
+
"loss": 1.9294,
|
| 1973 |
+
"step": 160500
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 2.36,
|
| 1977 |
+
"learning_rate": 4.360990332821855e-05,
|
| 1978 |
+
"loss": 1.8352,
|
| 1979 |
+
"step": 161000
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 2.37,
|
| 1983 |
+
"learning_rate": 4.342614163800553e-05,
|
| 1984 |
+
"loss": 1.9387,
|
| 1985 |
+
"step": 161500
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 2.37,
|
| 1989 |
+
"learning_rate": 4.32423070746059e-05,
|
| 1990 |
+
"loss": 1.8752,
|
| 1991 |
+
"step": 162000
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 2.38,
|
| 1995 |
+
"learning_rate": 4.3058403548135426e-05,
|
| 1996 |
+
"loss": 1.9911,
|
| 1997 |
+
"step": 162500
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 2.39,
|
| 2001 |
+
"learning_rate": 4.287443497017667e-05,
|
| 2002 |
+
"loss": 1.8686,
|
| 2003 |
+
"step": 163000
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 2.4,
|
| 2007 |
+
"learning_rate": 4.269040525369591e-05,
|
| 2008 |
+
"loss": 1.8261,
|
| 2009 |
+
"step": 163500
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 2.4,
|
| 2013 |
+
"learning_rate": 4.250631831295975e-05,
|
| 2014 |
+
"loss": 1.8804,
|
| 2015 |
+
"step": 164000
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 2.41,
|
| 2019 |
+
"learning_rate": 4.2322178063451956e-05,
|
| 2020 |
+
"loss": 1.8813,
|
| 2021 |
+
"step": 164500
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 2.42,
|
| 2025 |
+
"learning_rate": 4.2137988421790176e-05,
|
| 2026 |
+
"loss": 1.8637,
|
| 2027 |
+
"step": 165000
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 2.43,
|
| 2031 |
+
"learning_rate": 4.195375330564259e-05,
|
| 2032 |
+
"loss": 1.8864,
|
| 2033 |
+
"step": 165500
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 2.43,
|
| 2037 |
+
"learning_rate": 4.176947663364464e-05,
|
| 2038 |
+
"loss": 1.8915,
|
| 2039 |
+
"step": 166000
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 2.44,
|
| 2043 |
+
"learning_rate": 4.158516232531565e-05,
|
| 2044 |
+
"loss": 1.8492,
|
| 2045 |
+
"step": 166500
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 2.45,
|
| 2049 |
+
"learning_rate": 4.140081430097541e-05,
|
| 2050 |
+
"loss": 1.8539,
|
| 2051 |
+
"step": 167000
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 2.46,
|
| 2055 |
+
"learning_rate": 4.121643648166088e-05,
|
| 2056 |
+
"loss": 1.8811,
|
| 2057 |
+
"step": 167500
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 2.46,
|
| 2061 |
+
"learning_rate": 4.1032032789042785e-05,
|
| 2062 |
+
"loss": 1.8897,
|
| 2063 |
+
"step": 168000
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 2.47,
|
| 2067 |
+
"learning_rate": 4.084760714534209e-05,
|
| 2068 |
+
"loss": 1.8895,
|
| 2069 |
+
"step": 168500
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 2.48,
|
| 2073 |
+
"learning_rate": 4.066316347324673e-05,
|
| 2074 |
+
"loss": 1.9344,
|
| 2075 |
+
"step": 169000
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.48,
|
| 2079 |
+
"learning_rate": 4.047870569582805e-05,
|
| 2080 |
+
"loss": 1.8943,
|
| 2081 |
+
"step": 169500
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.49,
|
| 2085 |
+
"learning_rate": 4.029423773645742e-05,
|
| 2086 |
+
"loss": 1.8904,
|
| 2087 |
+
"step": 170000
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 2.5,
|
| 2091 |
+
"learning_rate": 4.0109763518722786e-05,
|
| 2092 |
+
"loss": 1.8938,
|
| 2093 |
+
"step": 170500
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 2.51,
|
| 2097 |
+
"learning_rate": 3.992528696634522e-05,
|
| 2098 |
+
"loss": 1.937,
|
| 2099 |
+
"step": 171000
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 2.51,
|
| 2103 |
+
"learning_rate": 3.9740812003095435e-05,
|
| 2104 |
+
"loss": 1.9084,
|
| 2105 |
+
"step": 171500
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 2.52,
|
| 2109 |
+
"learning_rate": 3.9556342552710344e-05,
|
| 2110 |
+
"loss": 1.8989,
|
| 2111 |
+
"step": 172000
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 2.53,
|
| 2115 |
+
"learning_rate": 3.937188253880958e-05,
|
| 2116 |
+
"loss": 1.9068,
|
| 2117 |
+
"step": 172500
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.54,
|
| 2121 |
+
"learning_rate": 3.9187435884812134e-05,
|
| 2122 |
+
"loss": 1.8497,
|
| 2123 |
+
"step": 173000
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.54,
|
| 2127 |
+
"learning_rate": 3.900300651385277e-05,
|
| 2128 |
+
"loss": 1.8392,
|
| 2129 |
+
"step": 173500
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 2.55,
|
| 2133 |
+
"learning_rate": 3.8818598348698666e-05,
|
| 2134 |
+
"loss": 1.9431,
|
| 2135 |
+
"step": 174000
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 2.56,
|
| 2139 |
+
"learning_rate": 3.863421531166596e-05,
|
| 2140 |
+
"loss": 1.9361,
|
| 2141 |
+
"step": 174500
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 2.57,
|
| 2145 |
+
"learning_rate": 3.844986132453632e-05,
|
| 2146 |
+
"loss": 1.8738,
|
| 2147 |
+
"step": 175000
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 2.57,
|
| 2151 |
+
"learning_rate": 3.826554030847353e-05,
|
| 2152 |
+
"loss": 1.9457,
|
| 2153 |
+
"step": 175500
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.58,
|
| 2157 |
+
"learning_rate": 3.8081256183940086e-05,
|
| 2158 |
+
"loss": 1.9293,
|
| 2159 |
+
"step": 176000
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.59,
|
| 2163 |
+
"learning_rate": 3.7897012870613815e-05,
|
| 2164 |
+
"loss": 1.9107,
|
| 2165 |
+
"step": 176500
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.59,
|
| 2169 |
+
"learning_rate": 3.771281428730448e-05,
|
| 2170 |
+
"loss": 1.8691,
|
| 2171 |
+
"step": 177000
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 2.6,
|
| 2175 |
+
"learning_rate": 3.75286643518705e-05,
|
| 2176 |
+
"loss": 1.856,
|
| 2177 |
+
"step": 177500
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.61,
|
| 2181 |
+
"learning_rate": 3.7344566981135476e-05,
|
| 2182 |
+
"loss": 1.9258,
|
| 2183 |
+
"step": 178000
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 2.62,
|
| 2187 |
+
"learning_rate": 3.716052609080507e-05,
|
| 2188 |
+
"loss": 1.8476,
|
| 2189 |
+
"step": 178500
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 2.62,
|
| 2193 |
+
"learning_rate": 3.697654559538356e-05,
|
| 2194 |
+
"loss": 1.8763,
|
| 2195 |
+
"step": 179000
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.63,
|
| 2199 |
+
"learning_rate": 3.6792629408090645e-05,
|
| 2200 |
+
"loss": 1.903,
|
| 2201 |
+
"step": 179500
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.64,
|
| 2205 |
+
"learning_rate": 3.660878144077822e-05,
|
| 2206 |
+
"loss": 1.8962,
|
| 2207 |
+
"step": 180000
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 2.65,
|
| 2211 |
+
"learning_rate": 3.642500560384714e-05,
|
| 2212 |
+
"loss": 1.965,
|
| 2213 |
+
"step": 180500
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 2.65,
|
| 2217 |
+
"learning_rate": 3.62413058061641e-05,
|
| 2218 |
+
"loss": 1.9176,
|
| 2219 |
+
"step": 181000
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 2.66,
|
| 2223 |
+
"learning_rate": 3.605768595497842e-05,
|
| 2224 |
+
"loss": 1.9391,
|
| 2225 |
+
"step": 181500
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 2.67,
|
| 2229 |
+
"learning_rate": 3.5874149955839e-05,
|
| 2230 |
+
"loss": 1.9248,
|
| 2231 |
+
"step": 182000
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 2.68,
|
| 2235 |
+
"learning_rate": 3.569070171251121e-05,
|
| 2236 |
+
"loss": 1.9606,
|
| 2237 |
+
"step": 182500
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.68,
|
| 2241 |
+
"learning_rate": 3.5507345126893905e-05,
|
| 2242 |
+
"loss": 1.9195,
|
| 2243 |
+
"step": 183000
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.69,
|
| 2247 |
+
"learning_rate": 3.5324084098936375e-05,
|
| 2248 |
+
"loss": 1.8899,
|
| 2249 |
+
"step": 183500
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 2.7,
|
| 2253 |
+
"learning_rate": 3.514092252655542e-05,
|
| 2254 |
+
"loss": 1.9041,
|
| 2255 |
+
"step": 184000
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 2.7,
|
| 2259 |
+
"learning_rate": 3.495786430555247e-05,
|
| 2260 |
+
"loss": 1.8918,
|
| 2261 |
+
"step": 184500
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 2.71,
|
| 2265 |
+
"learning_rate": 3.477491332953069e-05,
|
| 2266 |
+
"loss": 1.8615,
|
| 2267 |
+
"step": 185000
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 2.72,
|
| 2271 |
+
"learning_rate": 3.459207348981214e-05,
|
| 2272 |
+
"loss": 1.9503,
|
| 2273 |
+
"step": 185500
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 2.73,
|
| 2277 |
+
"learning_rate": 3.4409348675355084e-05,
|
| 2278 |
+
"loss": 1.8883,
|
| 2279 |
+
"step": 186000
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.73,
|
| 2283 |
+
"learning_rate": 3.422674277267118e-05,
|
| 2284 |
+
"loss": 1.8797,
|
| 2285 |
+
"step": 186500
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.74,
|
| 2289 |
+
"learning_rate": 3.404425966574292e-05,
|
| 2290 |
+
"loss": 1.9118,
|
| 2291 |
+
"step": 187000
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 2.75,
|
| 2295 |
+
"learning_rate": 3.386190323594091e-05,
|
| 2296 |
+
"loss": 1.8573,
|
| 2297 |
+
"step": 187500
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 2.76,
|
| 2301 |
+
"learning_rate": 3.3679677361941365e-05,
|
| 2302 |
+
"loss": 1.9455,
|
| 2303 |
+
"step": 188000
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 2.76,
|
| 2307 |
+
"learning_rate": 3.349758591964365e-05,
|
| 2308 |
+
"loss": 1.9058,
|
| 2309 |
+
"step": 188500
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 2.77,
|
| 2313 |
+
"learning_rate": 3.3315632782087784e-05,
|
| 2314 |
+
"loss": 1.9369,
|
| 2315 |
+
"step": 189000
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.78,
|
| 2319 |
+
"learning_rate": 3.3133821819372065e-05,
|
| 2320 |
+
"loss": 1.9005,
|
| 2321 |
+
"step": 189500
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.79,
|
| 2325 |
+
"learning_rate": 3.295215689857077e-05,
|
| 2326 |
+
"loss": 1.9293,
|
| 2327 |
+
"step": 190000
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.79,
|
| 2331 |
+
"learning_rate": 3.277064188365191e-05,
|
| 2332 |
+
"loss": 1.9078,
|
| 2333 |
+
"step": 190500
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 2.8,
|
| 2337 |
+
"learning_rate": 3.258928063539502e-05,
|
| 2338 |
+
"loss": 1.928,
|
| 2339 |
+
"step": 191000
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 2.81,
|
| 2343 |
+
"learning_rate": 3.240807701130909e-05,
|
| 2344 |
+
"loss": 1.9753,
|
| 2345 |
+
"step": 191500
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 2.81,
|
| 2349 |
+
"learning_rate": 3.2227034865550444e-05,
|
| 2350 |
+
"loss": 1.9069,
|
| 2351 |
+
"step": 192000
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 2.82,
|
| 2355 |
+
"learning_rate": 3.204615804884081e-05,
|
| 2356 |
+
"loss": 1.9923,
|
| 2357 |
+
"step": 192500
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.83,
|
| 2361 |
+
"learning_rate": 3.1865450408385434e-05,
|
| 2362 |
+
"loss": 1.8337,
|
| 2363 |
+
"step": 193000
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.84,
|
| 2367 |
+
"learning_rate": 3.1684915787791185e-05,
|
| 2368 |
+
"loss": 1.9092,
|
| 2369 |
+
"step": 193500
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.84,
|
| 2373 |
+
"learning_rate": 3.150455802698488e-05,
|
| 2374 |
+
"loss": 1.8451,
|
| 2375 |
+
"step": 194000
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.85,
|
| 2379 |
+
"learning_rate": 3.132438096213156e-05,
|
| 2380 |
+
"loss": 1.8952,
|
| 2381 |
+
"step": 194500
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 2.86,
|
| 2385 |
+
"learning_rate": 3.114438842555287e-05,
|
| 2386 |
+
"loss": 1.9238,
|
| 2387 |
+
"step": 195000
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 2.87,
|
| 2391 |
+
"learning_rate": 3.096458424564564e-05,
|
| 2392 |
+
"loss": 1.8765,
|
| 2393 |
+
"step": 195500
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 2.87,
|
| 2397 |
+
"learning_rate": 3.07849722468004e-05,
|
| 2398 |
+
"loss": 1.8946,
|
| 2399 |
+
"step": 196000
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.88,
|
| 2403 |
+
"learning_rate": 3.060555624931998e-05,
|
| 2404 |
+
"loss": 1.8671,
|
| 2405 |
+
"step": 196500
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.89,
|
| 2409 |
+
"learning_rate": 3.042634006933835e-05,
|
| 2410 |
+
"loss": 1.9126,
|
| 2411 |
+
"step": 197000
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.9,
|
| 2415 |
+
"learning_rate": 3.024732751873942e-05,
|
| 2416 |
+
"loss": 1.9464,
|
| 2417 |
+
"step": 197500
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 2.9,
|
| 2421 |
+
"learning_rate": 3.006852240507591e-05,
|
| 2422 |
+
"loss": 1.8723,
|
| 2423 |
+
"step": 198000
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 2.91,
|
| 2427 |
+
"learning_rate": 2.988992853148847e-05,
|
| 2428 |
+
"loss": 1.892,
|
| 2429 |
+
"step": 198500
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 2.92,
|
| 2433 |
+
"learning_rate": 2.9711549696624666e-05,
|
| 2434 |
+
"loss": 1.8947,
|
| 2435 |
+
"step": 199000
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 2.92,
|
| 2439 |
+
"learning_rate": 2.9533389694558266e-05,
|
| 2440 |
+
"loss": 1.9721,
|
| 2441 |
+
"step": 199500
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.93,
|
| 2445 |
+
"learning_rate": 2.9355452314708538e-05,
|
| 2446 |
+
"loss": 1.9074,
|
| 2447 |
+
"step": 200000
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.94,
|
| 2451 |
+
"learning_rate": 2.9177741341759592e-05,
|
| 2452 |
+
"loss": 1.9363,
|
| 2453 |
+
"step": 200500
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.95,
|
| 2457 |
+
"learning_rate": 2.9000260555579957e-05,
|
| 2458 |
+
"loss": 1.8815,
|
| 2459 |
+
"step": 201000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.95,
|
| 2463 |
+
"learning_rate": 2.8823013731142127e-05,
|
| 2464 |
+
"loss": 1.9579,
|
| 2465 |
+
"step": 201500
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.96,
|
| 2469 |
+
"learning_rate": 2.8646004638442282e-05,
|
| 2470 |
+
"loss": 1.9756,
|
| 2471 |
+
"step": 202000
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.97,
|
| 2475 |
+
"learning_rate": 2.8469237042420128e-05,
|
| 2476 |
+
"loss": 1.8637,
|
| 2477 |
+
"step": 202500
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 2.98,
|
| 2481 |
+
"learning_rate": 2.8292714702878754e-05,
|
| 2482 |
+
"loss": 1.8707,
|
| 2483 |
+
"step": 203000
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 2.98,
|
| 2487 |
+
"learning_rate": 2.811644137440477e-05,
|
| 2488 |
+
"loss": 1.9256,
|
| 2489 |
+
"step": 203500
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.99,
|
| 2493 |
+
"learning_rate": 2.7940420806288327e-05,
|
| 2494 |
+
"loss": 1.9739,
|
| 2495 |
+
"step": 204000
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 3.0,
|
| 2499 |
+
"learning_rate": 2.776465674244345e-05,
|
| 2500 |
+
"loss": 1.8675,
|
| 2501 |
+
"step": 204500
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 3.0,
|
| 2505 |
+
"eval_bleu": 1.0,
|
| 2506 |
+
"eval_brevity_penalty": 1.0,
|
| 2507 |
+
"eval_length_ratio": 1.0,
|
| 2508 |
+
"eval_loss": 2.297696352005005,
|
| 2509 |
+
"eval_precisions": [
|
| 2510 |
+
1.0,
|
| 2511 |
+
1.0,
|
| 2512 |
+
1.0,
|
| 2513 |
+
1.0
|
| 2514 |
+
],
|
| 2515 |
+
"eval_reference_length": 7761920,
|
| 2516 |
+
"eval_runtime": 17697.7556,
|
| 2517 |
+
"eval_samples_per_second": 0.857,
|
| 2518 |
+
"eval_steps_per_second": 0.428,
|
| 2519 |
+
"eval_translation_length": 7761920,
|
| 2520 |
+
"step": 204657
|
| 2521 |
}
|
| 2522 |
],
|
| 2523 |
"logging_steps": 500,
|
| 2524 |
"max_steps": 341095,
|
| 2525 |
"num_train_epochs": 5,
|
| 2526 |
"save_steps": 5000,
|
| 2527 |
+
"total_flos": 9.433737890333983e+17,
|
| 2528 |
"trial_name": null,
|
| 2529 |
"trial_params": null
|
| 2530 |
}
|