Training in progress, step 250, 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 201892112
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fef308b8921bdaa4b3462e88f1c7c0ade05615f1ab91ad36b5610d3f5b3c17d
|
| 3 |
size 201892112
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 403961210
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a90254e302ec8b6d14e289c29594d11ebf0d484538ff3c10b69180beee9fbee6
|
| 3 |
size 403961210
|
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 sha256:0b3a0b8ae4b1ef3d273ffeb9e44259fa1cdfe737176b54fe581e2c5bb5beb35e
|
| 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:a0f2082f8ae2500626171b5f3c174135f50005132f4fb0d89617e1b5f611c23e
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 2.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1447,6 +1447,364 @@
|
|
| 1447 |
"eval_samples_per_second": 42.613,
|
| 1448 |
"eval_steps_per_second": 21.448,
|
| 1449 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1450 |
}
|
| 1451 |
],
|
| 1452 |
"logging_steps": 1,
|
|
@@ -1475,7 +1833,7 @@
|
|
| 1475 |
"attributes": {}
|
| 1476 |
}
|
| 1477 |
},
|
| 1478 |
-
"total_flos":
|
| 1479 |
"train_batch_size": 8,
|
| 1480 |
"trial_name": null,
|
| 1481 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.2524751424789429,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-250",
|
| 4 |
+
"epoch": 2.793296089385475,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 250,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1447 |
"eval_samples_per_second": 42.613,
|
| 1448 |
"eval_steps_per_second": 21.448,
|
| 1449 |
"step": 200
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 2.2458100558659218,
|
| 1453 |
+
"grad_norm": 2.0551750659942627,
|
| 1454 |
+
"learning_rate": 1.8678916237581522e-05,
|
| 1455 |
+
"loss": 1.0606,
|
| 1456 |
+
"step": 201
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 2.2569832402234637,
|
| 1460 |
+
"grad_norm": 1.3848985433578491,
|
| 1461 |
+
"learning_rate": 1.816933090929055e-05,
|
| 1462 |
+
"loss": 1.2715,
|
| 1463 |
+
"step": 202
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 2.2681564245810057,
|
| 1467 |
+
"grad_norm": 1.1580418348312378,
|
| 1468 |
+
"learning_rate": 1.7665245337452368e-05,
|
| 1469 |
+
"loss": 1.1072,
|
| 1470 |
+
"step": 203
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 2.2793296089385473,
|
| 1474 |
+
"grad_norm": 1.1287815570831299,
|
| 1475 |
+
"learning_rate": 1.716674661881848e-05,
|
| 1476 |
+
"loss": 1.1187,
|
| 1477 |
+
"step": 204
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 2.2905027932960893,
|
| 1481 |
+
"grad_norm": 1.1857539415359497,
|
| 1482 |
+
"learning_rate": 1.667392088483456e-05,
|
| 1483 |
+
"loss": 1.3102,
|
| 1484 |
+
"step": 205
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 2.3016759776536313,
|
| 1488 |
+
"grad_norm": 1.0973467826843262,
|
| 1489 |
+
"learning_rate": 1.6186853286758397e-05,
|
| 1490 |
+
"loss": 1.2696,
|
| 1491 |
+
"step": 206
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 2.3128491620111733,
|
| 1495 |
+
"grad_norm": 1.0390197038650513,
|
| 1496 |
+
"learning_rate": 1.570562798094747e-05,
|
| 1497 |
+
"loss": 1.2218,
|
| 1498 |
+
"step": 207
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 2.3240223463687153,
|
| 1502 |
+
"grad_norm": 1.2470310926437378,
|
| 1503 |
+
"learning_rate": 1.5230328114318127e-05,
|
| 1504 |
+
"loss": 1.3041,
|
| 1505 |
+
"step": 208
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 2.335195530726257,
|
| 1509 |
+
"grad_norm": 1.1841509342193604,
|
| 1510 |
+
"learning_rate": 1.4761035809979395e-05,
|
| 1511 |
+
"loss": 1.2008,
|
| 1512 |
+
"step": 209
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 2.346368715083799,
|
| 1516 |
+
"grad_norm": 1.027030110359192,
|
| 1517 |
+
"learning_rate": 1.4297832153043656e-05,
|
| 1518 |
+
"loss": 1.0652,
|
| 1519 |
+
"step": 210
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 2.357541899441341,
|
| 1523 |
+
"grad_norm": 1.1887884140014648,
|
| 1524 |
+
"learning_rate": 1.3840797176616466e-05,
|
| 1525 |
+
"loss": 1.1927,
|
| 1526 |
+
"step": 211
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 2.368715083798883,
|
| 1530 |
+
"grad_norm": 1.1091079711914062,
|
| 1531 |
+
"learning_rate": 1.3390009847968504e-05,
|
| 1532 |
+
"loss": 1.157,
|
| 1533 |
+
"step": 212
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 2.3798882681564244,
|
| 1537 |
+
"grad_norm": 1.1894493103027344,
|
| 1538 |
+
"learning_rate": 1.2945548054891321e-05,
|
| 1539 |
+
"loss": 1.1287,
|
| 1540 |
+
"step": 213
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 2.3910614525139664,
|
| 1544 |
+
"grad_norm": 1.2934393882751465,
|
| 1545 |
+
"learning_rate": 1.2507488592239847e-05,
|
| 1546 |
+
"loss": 1.127,
|
| 1547 |
+
"step": 214
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.4022346368715084,
|
| 1551 |
+
"grad_norm": 1.1570240259170532,
|
| 1552 |
+
"learning_rate": 1.2075907148663579e-05,
|
| 1553 |
+
"loss": 1.0553,
|
| 1554 |
+
"step": 215
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 2.4134078212290504,
|
| 1558 |
+
"grad_norm": 1.172692060470581,
|
| 1559 |
+
"learning_rate": 1.1650878293528994e-05,
|
| 1560 |
+
"loss": 1.0155,
|
| 1561 |
+
"step": 216
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 2.4245810055865924,
|
| 1565 |
+
"grad_norm": 1.197474718093872,
|
| 1566 |
+
"learning_rate": 1.1232475464035385e-05,
|
| 1567 |
+
"loss": 1.135,
|
| 1568 |
+
"step": 217
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 2.435754189944134,
|
| 1572 |
+
"grad_norm": 1.1875431537628174,
|
| 1573 |
+
"learning_rate": 1.0820770952526155e-05,
|
| 1574 |
+
"loss": 1.1129,
|
| 1575 |
+
"step": 218
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 2.446927374301676,
|
| 1579 |
+
"grad_norm": 1.4438897371292114,
|
| 1580 |
+
"learning_rate": 1.0415835893998116e-05,
|
| 1581 |
+
"loss": 1.2491,
|
| 1582 |
+
"step": 219
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 2.458100558659218,
|
| 1586 |
+
"grad_norm": 1.3826245069503784,
|
| 1587 |
+
"learning_rate": 1.0017740253810609e-05,
|
| 1588 |
+
"loss": 1.1198,
|
| 1589 |
+
"step": 220
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 2.46927374301676,
|
| 1593 |
+
"grad_norm": 1.3730559349060059,
|
| 1594 |
+
"learning_rate": 9.62655281559679e-06,
|
| 1595 |
+
"loss": 1.0396,
|
| 1596 |
+
"step": 221
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 2.4804469273743015,
|
| 1600 |
+
"grad_norm": 1.6272450685501099,
|
| 1601 |
+
"learning_rate": 9.242341169379076e-06,
|
| 1602 |
+
"loss": 1.1171,
|
| 1603 |
+
"step": 222
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 2.4916201117318435,
|
| 1607 |
+
"grad_norm": 1.9734584093093872,
|
| 1608 |
+
"learning_rate": 8.865171699890834e-06,
|
| 1609 |
+
"loss": 1.1374,
|
| 1610 |
+
"step": 223
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 2.5027932960893855,
|
| 1614 |
+
"grad_norm": 1.3264243602752686,
|
| 1615 |
+
"learning_rate": 8.49510957510633e-06,
|
| 1616 |
+
"loss": 1.3275,
|
| 1617 |
+
"step": 224
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 2.5139664804469275,
|
| 1621 |
+
"grad_norm": 1.3713715076446533,
|
| 1622 |
+
"learning_rate": 8.132218734980852e-06,
|
| 1623 |
+
"loss": 1.2514,
|
| 1624 |
+
"step": 225
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 2.5251396648044695,
|
| 1628 |
+
"grad_norm": 1.2266292572021484,
|
| 1629 |
+
"learning_rate": 7.776561880403072e-06,
|
| 1630 |
+
"loss": 1.1443,
|
| 1631 |
+
"step": 226
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 2.536312849162011,
|
| 1635 |
+
"grad_norm": 1.2601666450500488,
|
| 1636 |
+
"learning_rate": 7.4282004623615396e-06,
|
| 1637 |
+
"loss": 1.2684,
|
| 1638 |
+
"step": 227
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 2.547486033519553,
|
| 1642 |
+
"grad_norm": 1.222358226776123,
|
| 1643 |
+
"learning_rate": 7.0871946713269856e-06,
|
| 1644 |
+
"loss": 1.1953,
|
| 1645 |
+
"step": 228
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 2.558659217877095,
|
| 1649 |
+
"grad_norm": 1.1266566514968872,
|
| 1650 |
+
"learning_rate": 6.753603426852589e-06,
|
| 1651 |
+
"loss": 1.1542,
|
| 1652 |
+
"step": 229
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 2.5698324022346366,
|
| 1656 |
+
"grad_norm": 1.035483479499817,
|
| 1657 |
+
"learning_rate": 6.427484367393699e-06,
|
| 1658 |
+
"loss": 1.1903,
|
| 1659 |
+
"step": 230
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 2.5810055865921786,
|
| 1663 |
+
"grad_norm": 1.0014451742172241,
|
| 1664 |
+
"learning_rate": 6.108893840348995e-06,
|
| 1665 |
+
"loss": 1.0823,
|
| 1666 |
+
"step": 231
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 2.5921787709497206,
|
| 1670 |
+
"grad_norm": 1.1314361095428467,
|
| 1671 |
+
"learning_rate": 5.797886892324694e-06,
|
| 1672 |
+
"loss": 1.2311,
|
| 1673 |
+
"step": 232
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 2.6033519553072626,
|
| 1677 |
+
"grad_norm": 1.0532621145248413,
|
| 1678 |
+
"learning_rate": 5.494517259623477e-06,
|
| 1679 |
+
"loss": 1.1181,
|
| 1680 |
+
"step": 233
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 2.6145251396648046,
|
| 1684 |
+
"grad_norm": 1.0458093881607056,
|
| 1685 |
+
"learning_rate": 5.198837358959901e-06,
|
| 1686 |
+
"loss": 1.1474,
|
| 1687 |
+
"step": 234
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 2.6256983240223466,
|
| 1691 |
+
"grad_norm": 1.1469138860702515,
|
| 1692 |
+
"learning_rate": 4.910898278403669e-06,
|
| 1693 |
+
"loss": 1.1677,
|
| 1694 |
+
"step": 235
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 2.636871508379888,
|
| 1698 |
+
"grad_norm": 1.135291337966919,
|
| 1699 |
+
"learning_rate": 4.630749768552589e-06,
|
| 1700 |
+
"loss": 1.1351,
|
| 1701 |
+
"step": 236
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 2.64804469273743,
|
| 1705 |
+
"grad_norm": 1.09344482421875,
|
| 1706 |
+
"learning_rate": 4.358440233936617e-06,
|
| 1707 |
+
"loss": 1.1221,
|
| 1708 |
+
"step": 237
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 2.659217877094972,
|
| 1712 |
+
"grad_norm": 1.190507411956787,
|
| 1713 |
+
"learning_rate": 4.094016724654359e-06,
|
| 1714 |
+
"loss": 1.0893,
|
| 1715 |
+
"step": 238
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 2.6703910614525137,
|
| 1719 |
+
"grad_norm": 1.1962109804153442,
|
| 1720 |
+
"learning_rate": 3.837524928243774e-06,
|
| 1721 |
+
"loss": 1.0827,
|
| 1722 |
+
"step": 239
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 2.6815642458100557,
|
| 1726 |
+
"grad_norm": 1.3094158172607422,
|
| 1727 |
+
"learning_rate": 3.589009161788104e-06,
|
| 1728 |
+
"loss": 1.0457,
|
| 1729 |
+
"step": 240
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 2.6927374301675977,
|
| 1733 |
+
"grad_norm": 1.3989758491516113,
|
| 1734 |
+
"learning_rate": 3.3485123642587658e-06,
|
| 1735 |
+
"loss": 1.0812,
|
| 1736 |
+
"step": 241
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 2.7039106145251397,
|
| 1740 |
+
"grad_norm": 1.4350848197937012,
|
| 1741 |
+
"learning_rate": 3.116076089096265e-06,
|
| 1742 |
+
"loss": 1.0437,
|
| 1743 |
+
"step": 242
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.7150837988826817,
|
| 1747 |
+
"grad_norm": 1.4926385879516602,
|
| 1748 |
+
"learning_rate": 2.8917404970305097e-06,
|
| 1749 |
+
"loss": 1.0832,
|
| 1750 |
+
"step": 243
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 2.7262569832402237,
|
| 1754 |
+
"grad_norm": 1.7393081188201904,
|
| 1755 |
+
"learning_rate": 2.675544349141779e-06,
|
| 1756 |
+
"loss": 1.0668,
|
| 1757 |
+
"step": 244
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 2.7374301675977653,
|
| 1761 |
+
"grad_norm": 1.9382210969924927,
|
| 1762 |
+
"learning_rate": 2.4675250001635232e-06,
|
| 1763 |
+
"loss": 1.0107,
|
| 1764 |
+
"step": 245
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 2.7486033519553073,
|
| 1768 |
+
"grad_norm": 0.7568252086639404,
|
| 1769 |
+
"learning_rate": 2.2677183920281343e-06,
|
| 1770 |
+
"loss": 1.0928,
|
| 1771 |
+
"step": 246
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 2.7597765363128492,
|
| 1775 |
+
"grad_norm": 0.9040629267692566,
|
| 1776 |
+
"learning_rate": 2.076159047656889e-06,
|
| 1777 |
+
"loss": 1.3023,
|
| 1778 |
+
"step": 247
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 2.770949720670391,
|
| 1782 |
+
"grad_norm": 0.9287111163139343,
|
| 1783 |
+
"learning_rate": 1.892880064994934e-06,
|
| 1784 |
+
"loss": 1.1539,
|
| 1785 |
+
"step": 248
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 2.782122905027933,
|
| 1789 |
+
"grad_norm": 0.9362537860870361,
|
| 1790 |
+
"learning_rate": 1.7179131112926627e-06,
|
| 1791 |
+
"loss": 1.1163,
|
| 1792 |
+
"step": 249
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 2.793296089385475,
|
| 1796 |
+
"grad_norm": 0.9989909529685974,
|
| 1797 |
+
"learning_rate": 1.551288417634106e-06,
|
| 1798 |
+
"loss": 1.1343,
|
| 1799 |
+
"step": 250
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.793296089385475,
|
| 1803 |
+
"eval_loss": 1.2524751424789429,
|
| 1804 |
+
"eval_runtime": 3.5444,
|
| 1805 |
+
"eval_samples_per_second": 42.603,
|
| 1806 |
+
"eval_steps_per_second": 21.442,
|
| 1807 |
+
"step": 250
|
| 1808 |
}
|
| 1809 |
],
|
| 1810 |
"logging_steps": 1,
|
|
|
|
| 1833 |
"attributes": {}
|
| 1834 |
}
|
| 1835 |
},
|
| 1836 |
+
"total_flos": 5.3328696311808e+16,
|
| 1837 |
"train_batch_size": 8,
|
| 1838 |
"trial_name": null,
|
| 1839 |
"trial_params": null
|