Training in progress, epoch 4, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 37789864
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16aab112374e0637635192e631493e5cc9fe41a7e4e6e216c0bb99e95ae685a5
|
| 3 |
size 37789864
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2622266
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddbb2f973e41d59a5b9f8697998020f6be69cf0c9c3ecccf1446dc42581a6fef
|
| 3 |
size 2622266
|
last-checkpoint/rng_state.pth
CHANGED
|
Binary files a/last-checkpoint/rng_state.pth and b/last-checkpoint/rng_state.pth differ
|
|
|
last-checkpoint/scheduler.pt
CHANGED
|
Binary files a/last-checkpoint/scheduler.pt and b/last-checkpoint/scheduler.pt differ
|
|
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 3.
|
| 3 |
-
"best_model_checkpoint": "dq158/coqui/checkpoint-
|
| 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,
|
|
@@ -1516,13 +1516,512 @@
|
|
| 1516 |
"eval_steps_per_second": 0.552,
|
| 1517 |
"eval_translation_length": 4591104,
|
| 1518 |
"step": 121044
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1519 |
}
|
| 1520 |
],
|
| 1521 |
"logging_steps": 500,
|
| 1522 |
"max_steps": 1210440,
|
| 1523 |
"num_train_epochs": 30,
|
| 1524 |
"save_steps": 1000,
|
| 1525 |
-
"total_flos": 2.
|
| 1526 |
"trial_name": null,
|
| 1527 |
"trial_params": null
|
| 1528 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 3.0066018104553223,
|
| 3 |
+
"best_model_checkpoint": "dq158/coqui/checkpoint-161392",
|
| 4 |
+
"epoch": 4.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 161392,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1516 |
"eval_steps_per_second": 0.552,
|
| 1517 |
"eval_translation_length": 4591104,
|
| 1518 |
"step": 121044
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 3.01,
|
| 1522 |
+
"learning_rate": 7.80420680811743e-05,
|
| 1523 |
+
"loss": 3.0822,
|
| 1524 |
+
"step": 121500
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 3.02,
|
| 1528 |
+
"learning_rate": 7.80259880896051e-05,
|
| 1529 |
+
"loss": 3.0585,
|
| 1530 |
+
"step": 122000
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 3.04,
|
| 1534 |
+
"learning_rate": 7.800984400770236e-05,
|
| 1535 |
+
"loss": 3.1003,
|
| 1536 |
+
"step": 122500
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 3.05,
|
| 1540 |
+
"learning_rate": 7.799363586267584e-05,
|
| 1541 |
+
"loss": 2.9734,
|
| 1542 |
+
"step": 123000
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 3.06,
|
| 1546 |
+
"learning_rate": 7.797736368184335e-05,
|
| 1547 |
+
"loss": 3.1722,
|
| 1548 |
+
"step": 123500
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 3.07,
|
| 1552 |
+
"learning_rate": 7.79610274926306e-05,
|
| 1553 |
+
"loss": 3.1007,
|
| 1554 |
+
"step": 124000
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 3.09,
|
| 1558 |
+
"learning_rate": 7.794462732257115e-05,
|
| 1559 |
+
"loss": 3.0796,
|
| 1560 |
+
"step": 124500
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 3.1,
|
| 1564 |
+
"learning_rate": 7.792816319930645e-05,
|
| 1565 |
+
"loss": 3.0691,
|
| 1566 |
+
"step": 125000
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 3.11,
|
| 1570 |
+
"learning_rate": 7.791163515058568e-05,
|
| 1571 |
+
"loss": 3.1111,
|
| 1572 |
+
"step": 125500
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 3.12,
|
| 1576 |
+
"learning_rate": 7.78950432042658e-05,
|
| 1577 |
+
"loss": 3.0355,
|
| 1578 |
+
"step": 126000
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 3.14,
|
| 1582 |
+
"learning_rate": 7.787838738831148e-05,
|
| 1583 |
+
"loss": 3.0751,
|
| 1584 |
+
"step": 126500
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 3.15,
|
| 1588 |
+
"learning_rate": 7.786166773079499e-05,
|
| 1589 |
+
"loss": 3.1197,
|
| 1590 |
+
"step": 127000
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 3.16,
|
| 1594 |
+
"learning_rate": 7.784488425989624e-05,
|
| 1595 |
+
"loss": 3.0997,
|
| 1596 |
+
"step": 127500
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 3.17,
|
| 1600 |
+
"learning_rate": 7.782803700390268e-05,
|
| 1601 |
+
"loss": 3.2085,
|
| 1602 |
+
"step": 128000
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 3.18,
|
| 1606 |
+
"learning_rate": 7.781112599120928e-05,
|
| 1607 |
+
"loss": 3.1391,
|
| 1608 |
+
"step": 128500
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 3.2,
|
| 1612 |
+
"learning_rate": 7.779415125031842e-05,
|
| 1613 |
+
"loss": 3.1137,
|
| 1614 |
+
"step": 129000
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 3.21,
|
| 1618 |
+
"learning_rate": 7.777711280983994e-05,
|
| 1619 |
+
"loss": 3.1373,
|
| 1620 |
+
"step": 129500
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 3.22,
|
| 1624 |
+
"learning_rate": 7.776001069849104e-05,
|
| 1625 |
+
"loss": 3.1228,
|
| 1626 |
+
"step": 130000
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 3.23,
|
| 1630 |
+
"learning_rate": 7.774284494509619e-05,
|
| 1631 |
+
"loss": 3.0356,
|
| 1632 |
+
"step": 130500
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 3.25,
|
| 1636 |
+
"learning_rate": 7.772561557858717e-05,
|
| 1637 |
+
"loss": 3.074,
|
| 1638 |
+
"step": 131000
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 3.26,
|
| 1642 |
+
"learning_rate": 7.770832262800298e-05,
|
| 1643 |
+
"loss": 3.101,
|
| 1644 |
+
"step": 131500
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 3.27,
|
| 1648 |
+
"learning_rate": 7.769096612248972e-05,
|
| 1649 |
+
"loss": 3.0026,
|
| 1650 |
+
"step": 132000
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 3.28,
|
| 1654 |
+
"learning_rate": 7.767354609130067e-05,
|
| 1655 |
+
"loss": 2.9642,
|
| 1656 |
+
"step": 132500
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 3.3,
|
| 1660 |
+
"learning_rate": 7.765606256379617e-05,
|
| 1661 |
+
"loss": 3.1591,
|
| 1662 |
+
"step": 133000
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 3.31,
|
| 1666 |
+
"learning_rate": 7.763851556944357e-05,
|
| 1667 |
+
"loss": 3.1563,
|
| 1668 |
+
"step": 133500
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 3.32,
|
| 1672 |
+
"learning_rate": 7.762090513781717e-05,
|
| 1673 |
+
"loss": 3.0032,
|
| 1674 |
+
"step": 134000
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 3.33,
|
| 1678 |
+
"learning_rate": 7.760323129859824e-05,
|
| 1679 |
+
"loss": 3.0752,
|
| 1680 |
+
"step": 134500
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 3.35,
|
| 1684 |
+
"learning_rate": 7.758549408157487e-05,
|
| 1685 |
+
"loss": 3.1228,
|
| 1686 |
+
"step": 135000
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 3.36,
|
| 1690 |
+
"learning_rate": 7.7567693516642e-05,
|
| 1691 |
+
"loss": 3.1929,
|
| 1692 |
+
"step": 135500
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 3.37,
|
| 1696 |
+
"learning_rate": 7.754982963380129e-05,
|
| 1697 |
+
"loss": 3.0963,
|
| 1698 |
+
"step": 136000
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 3.38,
|
| 1702 |
+
"learning_rate": 7.75319024631612e-05,
|
| 1703 |
+
"loss": 3.1731,
|
| 1704 |
+
"step": 136500
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 3.4,
|
| 1708 |
+
"learning_rate": 7.751391203493678e-05,
|
| 1709 |
+
"loss": 3.1822,
|
| 1710 |
+
"step": 137000
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 3.41,
|
| 1714 |
+
"learning_rate": 7.749585837944974e-05,
|
| 1715 |
+
"loss": 3.0849,
|
| 1716 |
+
"step": 137500
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 3.42,
|
| 1720 |
+
"learning_rate": 7.747774152712836e-05,
|
| 1721 |
+
"loss": 3.1609,
|
| 1722 |
+
"step": 138000
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 3.43,
|
| 1726 |
+
"learning_rate": 7.745956150850738e-05,
|
| 1727 |
+
"loss": 3.0784,
|
| 1728 |
+
"step": 138500
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 3.45,
|
| 1732 |
+
"learning_rate": 7.744131835422808e-05,
|
| 1733 |
+
"loss": 3.1391,
|
| 1734 |
+
"step": 139000
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 3.46,
|
| 1738 |
+
"learning_rate": 7.74230120950381e-05,
|
| 1739 |
+
"loss": 3.1141,
|
| 1740 |
+
"step": 139500
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 3.47,
|
| 1744 |
+
"learning_rate": 7.740464276179143e-05,
|
| 1745 |
+
"loss": 3.092,
|
| 1746 |
+
"step": 140000
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 3.48,
|
| 1750 |
+
"learning_rate": 7.738621038544842e-05,
|
| 1751 |
+
"loss": 3.0634,
|
| 1752 |
+
"step": 140500
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 3.49,
|
| 1756 |
+
"learning_rate": 7.736771499707562e-05,
|
| 1757 |
+
"loss": 3.1131,
|
| 1758 |
+
"step": 141000
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 3.51,
|
| 1762 |
+
"learning_rate": 7.734915662784582e-05,
|
| 1763 |
+
"loss": 3.1014,
|
| 1764 |
+
"step": 141500
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 3.52,
|
| 1768 |
+
"learning_rate": 7.733053530903793e-05,
|
| 1769 |
+
"loss": 3.1118,
|
| 1770 |
+
"step": 142000
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 3.53,
|
| 1774 |
+
"learning_rate": 7.7311851072037e-05,
|
| 1775 |
+
"loss": 3.0762,
|
| 1776 |
+
"step": 142500
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 3.54,
|
| 1780 |
+
"learning_rate": 7.729310394833408e-05,
|
| 1781 |
+
"loss": 3.0613,
|
| 1782 |
+
"step": 143000
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 3.56,
|
| 1786 |
+
"learning_rate": 7.727429396952622e-05,
|
| 1787 |
+
"loss": 3.1007,
|
| 1788 |
+
"step": 143500
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 3.57,
|
| 1792 |
+
"learning_rate": 7.725542116731643e-05,
|
| 1793 |
+
"loss": 3.0766,
|
| 1794 |
+
"step": 144000
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 3.58,
|
| 1798 |
+
"learning_rate": 7.72364855735136e-05,
|
| 1799 |
+
"loss": 3.0842,
|
| 1800 |
+
"step": 144500
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 3.59,
|
| 1804 |
+
"learning_rate": 7.721748722003242e-05,
|
| 1805 |
+
"loss": 3.1643,
|
| 1806 |
+
"step": 145000
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 3.61,
|
| 1810 |
+
"learning_rate": 7.719842613889342e-05,
|
| 1811 |
+
"loss": 3.0702,
|
| 1812 |
+
"step": 145500
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 3.62,
|
| 1816 |
+
"learning_rate": 7.717930236222277e-05,
|
| 1817 |
+
"loss": 3.2394,
|
| 1818 |
+
"step": 146000
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 3.63,
|
| 1822 |
+
"learning_rate": 7.716011592225239e-05,
|
| 1823 |
+
"loss": 3.0847,
|
| 1824 |
+
"step": 146500
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 3.64,
|
| 1828 |
+
"learning_rate": 7.714086685131975e-05,
|
| 1829 |
+
"loss": 3.1536,
|
| 1830 |
+
"step": 147000
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 3.66,
|
| 1834 |
+
"learning_rate": 7.712155518186792e-05,
|
| 1835 |
+
"loss": 3.0908,
|
| 1836 |
+
"step": 147500
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 3.67,
|
| 1840 |
+
"learning_rate": 7.710218094644548e-05,
|
| 1841 |
+
"loss": 3.0379,
|
| 1842 |
+
"step": 148000
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 3.68,
|
| 1846 |
+
"learning_rate": 7.708274417770644e-05,
|
| 1847 |
+
"loss": 3.0513,
|
| 1848 |
+
"step": 148500
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 3.69,
|
| 1852 |
+
"learning_rate": 7.706324490841019e-05,
|
| 1853 |
+
"loss": 3.1075,
|
| 1854 |
+
"step": 149000
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 3.71,
|
| 1858 |
+
"learning_rate": 7.704368317142151e-05,
|
| 1859 |
+
"loss": 3.1261,
|
| 1860 |
+
"step": 149500
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 3.72,
|
| 1864 |
+
"learning_rate": 7.702405899971042e-05,
|
| 1865 |
+
"loss": 3.164,
|
| 1866 |
+
"step": 150000
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 3.73,
|
| 1870 |
+
"learning_rate": 7.700437242635218e-05,
|
| 1871 |
+
"loss": 3.0038,
|
| 1872 |
+
"step": 150500
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 3.74,
|
| 1876 |
+
"learning_rate": 7.698462348452724e-05,
|
| 1877 |
+
"loss": 3.1165,
|
| 1878 |
+
"step": 151000
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 3.75,
|
| 1882 |
+
"learning_rate": 7.696481220752119e-05,
|
| 1883 |
+
"loss": 3.048,
|
| 1884 |
+
"step": 151500
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 3.77,
|
| 1888 |
+
"learning_rate": 7.694493862872459e-05,
|
| 1889 |
+
"loss": 3.0922,
|
| 1890 |
+
"step": 152000
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 3.78,
|
| 1894 |
+
"learning_rate": 7.69250027816331e-05,
|
| 1895 |
+
"loss": 3.1324,
|
| 1896 |
+
"step": 152500
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 3.79,
|
| 1900 |
+
"learning_rate": 7.69050046998473e-05,
|
| 1901 |
+
"loss": 3.1076,
|
| 1902 |
+
"step": 153000
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 3.8,
|
| 1906 |
+
"learning_rate": 7.688494441707267e-05,
|
| 1907 |
+
"loss": 3.0944,
|
| 1908 |
+
"step": 153500
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 3.82,
|
| 1912 |
+
"learning_rate": 7.686482196711948e-05,
|
| 1913 |
+
"loss": 3.0502,
|
| 1914 |
+
"step": 154000
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 3.83,
|
| 1918 |
+
"learning_rate": 7.684463738390284e-05,
|
| 1919 |
+
"loss": 3.0757,
|
| 1920 |
+
"step": 154500
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 3.84,
|
| 1924 |
+
"learning_rate": 7.682439070144252e-05,
|
| 1925 |
+
"loss": 3.0951,
|
| 1926 |
+
"step": 155000
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 3.85,
|
| 1930 |
+
"learning_rate": 7.680408195386303e-05,
|
| 1931 |
+
"loss": 3.0857,
|
| 1932 |
+
"step": 155500
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 3.87,
|
| 1936 |
+
"learning_rate": 7.678371117539342e-05,
|
| 1937 |
+
"loss": 3.0341,
|
| 1938 |
+
"step": 156000
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 3.88,
|
| 1942 |
+
"learning_rate": 7.676327840036733e-05,
|
| 1943 |
+
"loss": 3.0872,
|
| 1944 |
+
"step": 156500
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 3.89,
|
| 1948 |
+
"learning_rate": 7.674278366322286e-05,
|
| 1949 |
+
"loss": 3.0744,
|
| 1950 |
+
"step": 157000
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 3.9,
|
| 1954 |
+
"learning_rate": 7.672222699850256e-05,
|
| 1955 |
+
"loss": 3.0927,
|
| 1956 |
+
"step": 157500
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 3.92,
|
| 1960 |
+
"learning_rate": 7.670160844085336e-05,
|
| 1961 |
+
"loss": 3.0717,
|
| 1962 |
+
"step": 158000
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 3.93,
|
| 1966 |
+
"learning_rate": 7.668092802502651e-05,
|
| 1967 |
+
"loss": 2.964,
|
| 1968 |
+
"step": 158500
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 3.94,
|
| 1972 |
+
"learning_rate": 7.666018578587749e-05,
|
| 1973 |
+
"loss": 2.9958,
|
| 1974 |
+
"step": 159000
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 3.95,
|
| 1978 |
+
"learning_rate": 7.663938175836599e-05,
|
| 1979 |
+
"loss": 3.1135,
|
| 1980 |
+
"step": 159500
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 3.97,
|
| 1984 |
+
"learning_rate": 7.661851597755588e-05,
|
| 1985 |
+
"loss": 3.0903,
|
| 1986 |
+
"step": 160000
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 3.98,
|
| 1990 |
+
"learning_rate": 7.659758847861505e-05,
|
| 1991 |
+
"loss": 3.1399,
|
| 1992 |
+
"step": 160500
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 3.99,
|
| 1996 |
+
"learning_rate": 7.657659929681545e-05,
|
| 1997 |
+
"loss": 3.0583,
|
| 1998 |
+
"step": 161000
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 4.0,
|
| 2002 |
+
"eval_bleu": 1.0,
|
| 2003 |
+
"eval_brevity_penalty": 1.0,
|
| 2004 |
+
"eval_length_ratio": 1.0,
|
| 2005 |
+
"eval_loss": 3.0066018104553223,
|
| 2006 |
+
"eval_precisions": [
|
| 2007 |
+
1.0,
|
| 2008 |
+
1.0,
|
| 2009 |
+
1.0,
|
| 2010 |
+
1.0
|
| 2011 |
+
],
|
| 2012 |
+
"eval_reference_length": 4591104,
|
| 2013 |
+
"eval_runtime": 8243.7716,
|
| 2014 |
+
"eval_samples_per_second": 1.088,
|
| 2015 |
+
"eval_steps_per_second": 0.544,
|
| 2016 |
+
"eval_translation_length": 4591104,
|
| 2017 |
+
"step": 161392
|
| 2018 |
}
|
| 2019 |
],
|
| 2020 |
"logging_steps": 500,
|
| 2021 |
"max_steps": 1210440,
|
| 2022 |
"num_train_epochs": 30,
|
| 2023 |
"save_steps": 1000,
|
| 2024 |
+
"total_flos": 2.769874243981148e+18,
|
| 2025 |
"trial_name": null,
|
| 2026 |
"trial_params": null
|
| 2027 |
}
|