Training in progress, step 2400, 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 2433024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:697f816d0621c0d00e8e0ca9fe72e8fdd524ca12acf84d2e3bd927115c5c70e9
|
| 3 |
size 2433024
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2498406
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:82e1d534960590164150b98a4eca7e5d48f9e8072cfbd3c56be0575735e0c5b5
|
| 3 |
size 2498406
|
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 sha256:97474a95cf2d0e6166f036d8937e33ebebb2adb23cf1177f88edc10dc549c905
|
| 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:bf1f53caaa12767db3c6df563992bbf88f4b84dc57ec5080b22deb9c2c56ec6e
|
| 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": 0.
|
| 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,
|
|
@@ -1495,6 +1495,302 @@
|
|
| 1495 |
"eval_samples_per_second": 40.529,
|
| 1496 |
"eval_steps_per_second": 10.146,
|
| 1497 |
"step": 2000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1498 |
}
|
| 1499 |
],
|
| 1500 |
"logging_steps": 10,
|
|
@@ -1514,7 +1810,7 @@
|
|
| 1514 |
"attributes": {}
|
| 1515 |
}
|
| 1516 |
},
|
| 1517 |
-
"total_flos":
|
| 1518 |
"train_batch_size": 4,
|
| 1519 |
"trial_name": null,
|
| 1520 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7102693104468778,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 2400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1495 |
"eval_samples_per_second": 40.529,
|
| 1496 |
"eval_steps_per_second": 10.146,
|
| 1497 |
"step": 2000
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 0.5948505474992601,
|
| 1501 |
+
"grad_norm": 0.52734375,
|
| 1502 |
+
"learning_rate": 2.0280410844810428e-05,
|
| 1503 |
+
"loss": 10.5051,
|
| 1504 |
+
"step": 2010
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 0.5978100029594554,
|
| 1508 |
+
"grad_norm": 0.51953125,
|
| 1509 |
+
"learning_rate": 1.9488821249060297e-05,
|
| 1510 |
+
"loss": 10.497,
|
| 1511 |
+
"step": 2020
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 0.6007694584196508,
|
| 1515 |
+
"grad_norm": 0.58984375,
|
| 1516 |
+
"learning_rate": 1.871131877836879e-05,
|
| 1517 |
+
"loss": 10.5321,
|
| 1518 |
+
"step": 2030
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 0.6037289138798461,
|
| 1522 |
+
"grad_norm": 0.609375,
|
| 1523 |
+
"learning_rate": 1.7948039473155554e-05,
|
| 1524 |
+
"loss": 10.5627,
|
| 1525 |
+
"step": 2040
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 0.6066883693400414,
|
| 1529 |
+
"grad_norm": 1.2734375,
|
| 1530 |
+
"learning_rate": 1.7199116885197995e-05,
|
| 1531 |
+
"loss": 10.6749,
|
| 1532 |
+
"step": 2050
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 0.6096478248002367,
|
| 1536 |
+
"grad_norm": 0.46875,
|
| 1537 |
+
"learning_rate": 1.646468205426377e-05,
|
| 1538 |
+
"loss": 10.5196,
|
| 1539 |
+
"step": 2060
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 0.6126072802604321,
|
| 1543 |
+
"grad_norm": 0.51953125,
|
| 1544 |
+
"learning_rate": 1.5744863485182537e-05,
|
| 1545 |
+
"loss": 10.5294,
|
| 1546 |
+
"step": 2070
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 0.6155667357206274,
|
| 1550 |
+
"grad_norm": 0.5546875,
|
| 1551 |
+
"learning_rate": 1.5039787125361326e-05,
|
| 1552 |
+
"loss": 10.5178,
|
| 1553 |
+
"step": 2080
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 0.6185261911808227,
|
| 1557 |
+
"grad_norm": 0.703125,
|
| 1558 |
+
"learning_rate": 1.4349576342747462e-05,
|
| 1559 |
+
"loss": 10.6195,
|
| 1560 |
+
"step": 2090
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 0.621485646641018,
|
| 1564 |
+
"grad_norm": 0.94921875,
|
| 1565 |
+
"learning_rate": 1.3674351904242611e-05,
|
| 1566 |
+
"loss": 10.5954,
|
| 1567 |
+
"step": 2100
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 0.6244451021012134,
|
| 1571 |
+
"grad_norm": 0.458984375,
|
| 1572 |
+
"learning_rate": 1.3014231954572287e-05,
|
| 1573 |
+
"loss": 10.5128,
|
| 1574 |
+
"step": 2110
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 0.6274045575614087,
|
| 1578 |
+
"grad_norm": 0.52734375,
|
| 1579 |
+
"learning_rate": 1.2369331995613665e-05,
|
| 1580 |
+
"loss": 10.4853,
|
| 1581 |
+
"step": 2120
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 0.630364013021604,
|
| 1585 |
+
"grad_norm": 0.58984375,
|
| 1586 |
+
"learning_rate": 1.173976486618631e-05,
|
| 1587 |
+
"loss": 10.5693,
|
| 1588 |
+
"step": 2130
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 0.6333234684817993,
|
| 1592 |
+
"grad_norm": 0.5859375,
|
| 1593 |
+
"learning_rate": 1.1125640722308628e-05,
|
| 1594 |
+
"loss": 10.6354,
|
| 1595 |
+
"step": 2140
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 0.6362829239419947,
|
| 1599 |
+
"grad_norm": 1.7734375,
|
| 1600 |
+
"learning_rate": 1.0527067017923654e-05,
|
| 1601 |
+
"loss": 10.6234,
|
| 1602 |
+
"step": 2150
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 0.63924237940219,
|
| 1606 |
+
"grad_norm": 0.486328125,
|
| 1607 |
+
"learning_rate": 9.944148486097793e-06,
|
| 1608 |
+
"loss": 10.5555,
|
| 1609 |
+
"step": 2160
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 0.6422018348623854,
|
| 1613 |
+
"grad_norm": 0.515625,
|
| 1614 |
+
"learning_rate": 9.376987120695545e-06,
|
| 1615 |
+
"loss": 10.5214,
|
| 1616 |
+
"step": 2170
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.6451612903225806,
|
| 1620 |
+
"grad_norm": 0.62109375,
|
| 1621 |
+
"learning_rate": 8.825682158533554e-06,
|
| 1622 |
+
"loss": 10.5274,
|
| 1623 |
+
"step": 2180
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 0.648120745782776,
|
| 1627 |
+
"grad_norm": 0.79296875,
|
| 1628 |
+
"learning_rate": 8.290330062017016e-06,
|
| 1629 |
+
"loss": 10.6728,
|
| 1630 |
+
"step": 2190
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 0.6510802012429713,
|
| 1634 |
+
"grad_norm": 1.109375,
|
| 1635 |
+
"learning_rate": 7.771024502261526e-06,
|
| 1636 |
+
"loss": 10.6148,
|
| 1637 |
+
"step": 2200
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 0.6510802012429713,
|
| 1641 |
+
"eval_loss": 10.584245681762695,
|
| 1642 |
+
"eval_runtime": 37.0238,
|
| 1643 |
+
"eval_samples_per_second": 40.569,
|
| 1644 |
+
"eval_steps_per_second": 10.156,
|
| 1645 |
+
"step": 2200
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 0.6540396567031667,
|
| 1649 |
+
"grad_norm": 0.486328125,
|
| 1650 |
+
"learning_rate": 7.267856342703461e-06,
|
| 1651 |
+
"loss": 10.5768,
|
| 1652 |
+
"step": 2210
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.6569991121633619,
|
| 1656 |
+
"grad_norm": 0.51171875,
|
| 1657 |
+
"learning_rate": 6.780913623201346e-06,
|
| 1658 |
+
"loss": 10.4856,
|
| 1659 |
+
"step": 2220
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.6599585676235573,
|
| 1663 |
+
"grad_norm": 0.859375,
|
| 1664 |
+
"learning_rate": 6.310281544631546e-06,
|
| 1665 |
+
"loss": 10.6694,
|
| 1666 |
+
"step": 2230
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 0.6629180230837526,
|
| 1670 |
+
"grad_norm": 0.57421875,
|
| 1671 |
+
"learning_rate": 5.856042453980526e-06,
|
| 1672 |
+
"loss": 10.5306,
|
| 1673 |
+
"step": 2240
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 0.665877478543948,
|
| 1677 |
+
"grad_norm": 0.80078125,
|
| 1678 |
+
"learning_rate": 5.418275829936537e-06,
|
| 1679 |
+
"loss": 10.525,
|
| 1680 |
+
"step": 2250
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 0.6688369340041432,
|
| 1684 |
+
"grad_norm": 0.55859375,
|
| 1685 |
+
"learning_rate": 4.997058268983135e-06,
|
| 1686 |
+
"loss": 10.6271,
|
| 1687 |
+
"step": 2260
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 0.6717963894643386,
|
| 1691 |
+
"grad_norm": 0.53125,
|
| 1692 |
+
"learning_rate": 4.592463471997022e-06,
|
| 1693 |
+
"loss": 10.4891,
|
| 1694 |
+
"step": 2270
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.6747558449245339,
|
| 1698 |
+
"grad_norm": 0.546875,
|
| 1699 |
+
"learning_rate": 4.204562231352516e-06,
|
| 1700 |
+
"loss": 10.5647,
|
| 1701 |
+
"step": 2280
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.6777153003847292,
|
| 1705 |
+
"grad_norm": 0.63671875,
|
| 1706 |
+
"learning_rate": 3.83342241853496e-06,
|
| 1707 |
+
"loss": 10.506,
|
| 1708 |
+
"step": 2290
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 0.6806747558449245,
|
| 1712 |
+
"grad_norm": 0.87109375,
|
| 1713 |
+
"learning_rate": 3.4791089722651436e-06,
|
| 1714 |
+
"loss": 10.6414,
|
| 1715 |
+
"step": 2300
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 0.6836342113051198,
|
| 1719 |
+
"grad_norm": 0.478515625,
|
| 1720 |
+
"learning_rate": 3.1416838871368924e-06,
|
| 1721 |
+
"loss": 10.4905,
|
| 1722 |
+
"step": 2310
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.6865936667653152,
|
| 1726 |
+
"grad_norm": 0.55078125,
|
| 1727 |
+
"learning_rate": 2.821206202769899e-06,
|
| 1728 |
+
"loss": 10.5434,
|
| 1729 |
+
"step": 2320
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 0.6895531222255105,
|
| 1733 |
+
"grad_norm": 0.796875,
|
| 1734 |
+
"learning_rate": 2.5177319934794e-06,
|
| 1735 |
+
"loss": 10.6355,
|
| 1736 |
+
"step": 2330
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.6925125776857058,
|
| 1740 |
+
"grad_norm": 0.8984375,
|
| 1741 |
+
"learning_rate": 2.2313143584648423e-06,
|
| 1742 |
+
"loss": 10.5647,
|
| 1743 |
+
"step": 2340
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 0.6954720331459011,
|
| 1747 |
+
"grad_norm": 0.84765625,
|
| 1748 |
+
"learning_rate": 1.9620034125190644e-06,
|
| 1749 |
+
"loss": 10.7043,
|
| 1750 |
+
"step": 2350
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 0.6984314886060965,
|
| 1754 |
+
"grad_norm": 0.5,
|
| 1755 |
+
"learning_rate": 1.7098462772596302e-06,
|
| 1756 |
+
"loss": 10.4649,
|
| 1757 |
+
"step": 2360
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 0.7013909440662918,
|
| 1761 |
+
"grad_norm": 0.5078125,
|
| 1762 |
+
"learning_rate": 1.4748870728839347e-06,
|
| 1763 |
+
"loss": 10.4574,
|
| 1764 |
+
"step": 2370
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 0.7043503995264871,
|
| 1768 |
+
"grad_norm": 0.578125,
|
| 1769 |
+
"learning_rate": 1.2571669104494256e-06,
|
| 1770 |
+
"loss": 10.4982,
|
| 1771 |
+
"step": 2380
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 0.7073098549866824,
|
| 1775 |
+
"grad_norm": 0.9453125,
|
| 1776 |
+
"learning_rate": 1.0567238846803996e-06,
|
| 1777 |
+
"loss": 10.5488,
|
| 1778 |
+
"step": 2390
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.7102693104468778,
|
| 1782 |
+
"grad_norm": 0.796875,
|
| 1783 |
+
"learning_rate": 8.735930673024806e-07,
|
| 1784 |
+
"loss": 10.4931,
|
| 1785 |
+
"step": 2400
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 0.7102693104468778,
|
| 1789 |
+
"eval_loss": 10.585838317871094,
|
| 1790 |
+
"eval_runtime": 36.7691,
|
| 1791 |
+
"eval_samples_per_second": 40.85,
|
| 1792 |
+
"eval_steps_per_second": 10.226,
|
| 1793 |
+
"step": 2400
|
| 1794 |
}
|
| 1795 |
],
|
| 1796 |
"logging_steps": 10,
|
|
|
|
| 1810 |
"attributes": {}
|
| 1811 |
}
|
| 1812 |
},
|
| 1813 |
+
"total_flos": 235181703168.0,
|
| 1814 |
"train_batch_size": 4,
|
| 1815 |
"trial_name": null,
|
| 1816 |
"trial_params": null
|