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 389074464
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58acdbaf2a95685c6722b6405b19372d905ef1fba6996d7695325050f3d72ea5
|
| 3 |
size 389074464
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 198011252
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a446132d7c31b956eaefc7388b38251641542ce8b95d74805e64fc306de2f523
|
| 3 |
size 198011252
|
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 sha256:48d730db741b952a021078a26f33d6e0bf86090f9b5ed7a694df4d03784c8b62
|
| 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 sha256:7d5e1448282b30a66cf7be83aef18a251fdb6205c0184b42e99ae724602144bf
|
| 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": 0.
|
| 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": 16.361,
|
| 1448 |
"eval_steps_per_second": 4.09,
|
| 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": 1.1618001461029053,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-250",
|
| 4 |
+
"epoch": 0.3395585738539898,
|
| 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": 16.361,
|
| 1448 |
"eval_steps_per_second": 4.09,
|
| 1449 |
"step": 200
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.2730050933786078,
|
| 1453 |
+
"grad_norm": 0.22767877578735352,
|
| 1454 |
+
"learning_rate": 5.161079439470866e-05,
|
| 1455 |
+
"loss": 1.2705,
|
| 1456 |
+
"step": 201
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 0.27436332767402377,
|
| 1460 |
+
"grad_norm": 0.23202009499073029,
|
| 1461 |
+
"learning_rate": 5.1208187261806615e-05,
|
| 1462 |
+
"loss": 1.1301,
|
| 1463 |
+
"step": 202
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 0.27572156196943975,
|
| 1467 |
+
"grad_norm": 0.2392009049654007,
|
| 1468 |
+
"learning_rate": 5.080550173136457e-05,
|
| 1469 |
+
"loss": 1.2333,
|
| 1470 |
+
"step": 203
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 0.2770797962648557,
|
| 1474 |
+
"grad_norm": 0.2345076948404312,
|
| 1475 |
+
"learning_rate": 5.0402763933069496e-05,
|
| 1476 |
+
"loss": 1.1608,
|
| 1477 |
+
"step": 204
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 0.27843803056027167,
|
| 1481 |
+
"grad_norm": 0.2247873842716217,
|
| 1482 |
+
"learning_rate": 5e-05,
|
| 1483 |
+
"loss": 1.1777,
|
| 1484 |
+
"step": 205
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.2797962648556876,
|
| 1488 |
+
"grad_norm": 0.22494539618492126,
|
| 1489 |
+
"learning_rate": 4.9597236066930516e-05,
|
| 1490 |
+
"loss": 1.2278,
|
| 1491 |
+
"step": 206
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 0.2811544991511036,
|
| 1495 |
+
"grad_norm": 0.20997026562690735,
|
| 1496 |
+
"learning_rate": 4.919449826863544e-05,
|
| 1497 |
+
"loss": 1.1539,
|
| 1498 |
+
"step": 207
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 0.2825127334465195,
|
| 1502 |
+
"grad_norm": 0.20957647264003754,
|
| 1503 |
+
"learning_rate": 4.87918127381934e-05,
|
| 1504 |
+
"loss": 1.1532,
|
| 1505 |
+
"step": 208
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 0.2838709677419355,
|
| 1509 |
+
"grad_norm": 0.21663248538970947,
|
| 1510 |
+
"learning_rate": 4.8389205605291365e-05,
|
| 1511 |
+
"loss": 1.2097,
|
| 1512 |
+
"step": 209
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 0.28522920203735147,
|
| 1516 |
+
"grad_norm": 0.2205830067396164,
|
| 1517 |
+
"learning_rate": 4.798670299452926e-05,
|
| 1518 |
+
"loss": 1.1235,
|
| 1519 |
+
"step": 210
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 0.2865874363327674,
|
| 1523 |
+
"grad_norm": 0.21743406355381012,
|
| 1524 |
+
"learning_rate": 4.758433102372466e-05,
|
| 1525 |
+
"loss": 1.0935,
|
| 1526 |
+
"step": 211
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.2879456706281834,
|
| 1530 |
+
"grad_norm": 0.21099945902824402,
|
| 1531 |
+
"learning_rate": 4.7182115802218126e-05,
|
| 1532 |
+
"loss": 1.2179,
|
| 1533 |
+
"step": 212
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.2893039049235993,
|
| 1537 |
+
"grad_norm": 0.2079070806503296,
|
| 1538 |
+
"learning_rate": 4.678008342917903e-05,
|
| 1539 |
+
"loss": 1.1754,
|
| 1540 |
+
"step": 213
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 0.2906621392190153,
|
| 1544 |
+
"grad_norm": 0.2137642800807953,
|
| 1545 |
+
"learning_rate": 4.6378259991911886e-05,
|
| 1546 |
+
"loss": 1.2339,
|
| 1547 |
+
"step": 214
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 0.2920203735144312,
|
| 1551 |
+
"grad_norm": 0.22139067947864532,
|
| 1552 |
+
"learning_rate": 4.597667156416371e-05,
|
| 1553 |
+
"loss": 1.0929,
|
| 1554 |
+
"step": 215
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 0.2933786078098472,
|
| 1558 |
+
"grad_norm": 0.21825869381427765,
|
| 1559 |
+
"learning_rate": 4.5575344204432084e-05,
|
| 1560 |
+
"loss": 1.1509,
|
| 1561 |
+
"step": 216
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 0.29473684210526313,
|
| 1565 |
+
"grad_norm": 0.217268168926239,
|
| 1566 |
+
"learning_rate": 4.5174303954274244e-05,
|
| 1567 |
+
"loss": 1.1935,
|
| 1568 |
+
"step": 217
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.2960950764006791,
|
| 1572 |
+
"grad_norm": 0.21918080747127533,
|
| 1573 |
+
"learning_rate": 4.477357683661734e-05,
|
| 1574 |
+
"loss": 1.1205,
|
| 1575 |
+
"step": 218
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.2974533106960951,
|
| 1579 |
+
"grad_norm": 0.21873408555984497,
|
| 1580 |
+
"learning_rate": 4.437318885406973e-05,
|
| 1581 |
+
"loss": 1.0912,
|
| 1582 |
+
"step": 219
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 0.298811544991511,
|
| 1586 |
+
"grad_norm": 0.22883340716362,
|
| 1587 |
+
"learning_rate": 4.397316598723385e-05,
|
| 1588 |
+
"loss": 1.1726,
|
| 1589 |
+
"step": 220
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 0.300169779286927,
|
| 1593 |
+
"grad_norm": 0.22588761150836945,
|
| 1594 |
+
"learning_rate": 4.3573534193020274e-05,
|
| 1595 |
+
"loss": 1.1836,
|
| 1596 |
+
"step": 221
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 0.30152801358234294,
|
| 1600 |
+
"grad_norm": 0.22106902301311493,
|
| 1601 |
+
"learning_rate": 4.317431940296343e-05,
|
| 1602 |
+
"loss": 1.1351,
|
| 1603 |
+
"step": 222
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 0.3028862478777589,
|
| 1607 |
+
"grad_norm": 0.22283874452114105,
|
| 1608 |
+
"learning_rate": 4.277554752153895e-05,
|
| 1609 |
+
"loss": 1.1228,
|
| 1610 |
+
"step": 223
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.30424448217317485,
|
| 1614 |
+
"grad_norm": 0.21989330649375916,
|
| 1615 |
+
"learning_rate": 4.237724442448273e-05,
|
| 1616 |
+
"loss": 1.1015,
|
| 1617 |
+
"step": 224
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.30560271646859083,
|
| 1621 |
+
"grad_norm": 0.2293674349784851,
|
| 1622 |
+
"learning_rate": 4.197943595711198e-05,
|
| 1623 |
+
"loss": 1.1147,
|
| 1624 |
+
"step": 225
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 0.3069609507640068,
|
| 1628 |
+
"grad_norm": 0.2231401950120926,
|
| 1629 |
+
"learning_rate": 4.1582147932648074e-05,
|
| 1630 |
+
"loss": 1.1674,
|
| 1631 |
+
"step": 226
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 0.30831918505942274,
|
| 1635 |
+
"grad_norm": 0.2178943008184433,
|
| 1636 |
+
"learning_rate": 4.118540613054156e-05,
|
| 1637 |
+
"loss": 1.1364,
|
| 1638 |
+
"step": 227
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 0.3096774193548387,
|
| 1642 |
+
"grad_norm": 0.22939598560333252,
|
| 1643 |
+
"learning_rate": 4.078923629479943e-05,
|
| 1644 |
+
"loss": 1.0993,
|
| 1645 |
+
"step": 228
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 0.31103565365025465,
|
| 1649 |
+
"grad_norm": 0.22620512545108795,
|
| 1650 |
+
"learning_rate": 4.039366413231458e-05,
|
| 1651 |
+
"loss": 1.1105,
|
| 1652 |
+
"step": 229
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.31239388794567063,
|
| 1656 |
+
"grad_norm": 0.23713254928588867,
|
| 1657 |
+
"learning_rate": 3.9998715311197785e-05,
|
| 1658 |
+
"loss": 1.1572,
|
| 1659 |
+
"step": 230
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.31375212224108656,
|
| 1663 |
+
"grad_norm": 0.23715730011463165,
|
| 1664 |
+
"learning_rate": 3.960441545911204e-05,
|
| 1665 |
+
"loss": 1.1551,
|
| 1666 |
+
"step": 231
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 0.31511035653650254,
|
| 1670 |
+
"grad_norm": 0.2342255711555481,
|
| 1671 |
+
"learning_rate": 3.92107901616097e-05,
|
| 1672 |
+
"loss": 1.1296,
|
| 1673 |
+
"step": 232
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 0.31646859083191853,
|
| 1677 |
+
"grad_norm": 0.2432532161474228,
|
| 1678 |
+
"learning_rate": 3.8817864960472236e-05,
|
| 1679 |
+
"loss": 1.1381,
|
| 1680 |
+
"step": 233
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 0.31782682512733446,
|
| 1684 |
+
"grad_norm": 0.23796603083610535,
|
| 1685 |
+
"learning_rate": 3.842566535205286e-05,
|
| 1686 |
+
"loss": 1.1609,
|
| 1687 |
+
"step": 234
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 0.31918505942275044,
|
| 1691 |
+
"grad_norm": 0.23435711860656738,
|
| 1692 |
+
"learning_rate": 3.803421678562213e-05,
|
| 1693 |
+
"loss": 1.2078,
|
| 1694 |
+
"step": 235
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.32054329371816637,
|
| 1698 |
+
"grad_norm": 0.23656964302062988,
|
| 1699 |
+
"learning_rate": 3.764354466171652e-05,
|
| 1700 |
+
"loss": 1.1346,
|
| 1701 |
+
"step": 236
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.32190152801358235,
|
| 1705 |
+
"grad_norm": 0.24362805485725403,
|
| 1706 |
+
"learning_rate": 3.725367433049033e-05,
|
| 1707 |
+
"loss": 1.1695,
|
| 1708 |
+
"step": 237
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 0.3232597623089983,
|
| 1712 |
+
"grad_norm": 0.2416103184223175,
|
| 1713 |
+
"learning_rate": 3.6864631090070655e-05,
|
| 1714 |
+
"loss": 1.1603,
|
| 1715 |
+
"step": 238
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 0.32461799660441426,
|
| 1719 |
+
"grad_norm": 0.2516671121120453,
|
| 1720 |
+
"learning_rate": 3.6476440184915815e-05,
|
| 1721 |
+
"loss": 1.1832,
|
| 1722 |
+
"step": 239
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.32597623089983024,
|
| 1726 |
+
"grad_norm": 0.25407183170318604,
|
| 1727 |
+
"learning_rate": 3.608912680417737e-05,
|
| 1728 |
+
"loss": 1.1619,
|
| 1729 |
+
"step": 240
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 0.32733446519524617,
|
| 1733 |
+
"grad_norm": 0.24797658622264862,
|
| 1734 |
+
"learning_rate": 3.570271608006555e-05,
|
| 1735 |
+
"loss": 1.1321,
|
| 1736 |
+
"step": 241
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.32869269949066215,
|
| 1740 |
+
"grad_norm": 0.26799654960632324,
|
| 1741 |
+
"learning_rate": 3.531723308621847e-05,
|
| 1742 |
+
"loss": 1.2035,
|
| 1743 |
+
"step": 242
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 0.3300509337860781,
|
| 1747 |
+
"grad_norm": 0.27492526173591614,
|
| 1748 |
+
"learning_rate": 3.493270283607522e-05,
|
| 1749 |
+
"loss": 1.171,
|
| 1750 |
+
"step": 243
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 0.33140916808149407,
|
| 1754 |
+
"grad_norm": 0.2613578736782074,
|
| 1755 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 1756 |
+
"loss": 1.1238,
|
| 1757 |
+
"step": 244
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 0.33276740237691,
|
| 1761 |
+
"grad_norm": 0.28807345032691956,
|
| 1762 |
+
"learning_rate": 3.4166600309926387e-05,
|
| 1763 |
+
"loss": 1.1153,
|
| 1764 |
+
"step": 245
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 0.334125636672326,
|
| 1768 |
+
"grad_norm": 0.28468626737594604,
|
| 1769 |
+
"learning_rate": 3.3785077745215873e-05,
|
| 1770 |
+
"loss": 1.1984,
|
| 1771 |
+
"step": 246
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 0.33548387096774196,
|
| 1775 |
+
"grad_norm": 0.28014445304870605,
|
| 1776 |
+
"learning_rate": 3.340460734357359e-05,
|
| 1777 |
+
"loss": 1.1793,
|
| 1778 |
+
"step": 247
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.3368421052631579,
|
| 1782 |
+
"grad_norm": 0.2991635203361511,
|
| 1783 |
+
"learning_rate": 3.3025213793178646e-05,
|
| 1784 |
+
"loss": 1.1491,
|
| 1785 |
+
"step": 248
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 0.33820033955857387,
|
| 1789 |
+
"grad_norm": 0.34155353903770447,
|
| 1790 |
+
"learning_rate": 3.264692171233485e-05,
|
| 1791 |
+
"loss": 1.1472,
|
| 1792 |
+
"step": 249
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 0.3395585738539898,
|
| 1796 |
+
"grad_norm": 0.39328497648239136,
|
| 1797 |
+
"learning_rate": 3.226975564787322e-05,
|
| 1798 |
+
"loss": 1.1768,
|
| 1799 |
+
"step": 250
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 0.3395585738539898,
|
| 1803 |
+
"eval_loss": 1.1618001461029053,
|
| 1804 |
+
"eval_runtime": 75.9188,
|
| 1805 |
+
"eval_samples_per_second": 16.333,
|
| 1806 |
+
"eval_steps_per_second": 4.083,
|
| 1807 |
+
"step": 250
|
| 1808 |
}
|
| 1809 |
],
|
| 1810 |
"logging_steps": 1,
|
|
|
|
| 1833 |
"attributes": {}
|
| 1834 |
}
|
| 1835 |
},
|
| 1836 |
+
"total_flos": 2.0123162805849293e+17,
|
| 1837 |
"train_batch_size": 8,
|
| 1838 |
"trial_name": null,
|
| 1839 |
"trial_params": null
|