Training in progress, step 3000, 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 598434152
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71d0eabfbdd37b2fae13fa797c3f0436a08906b6795d79ec3027697d278db38e
|
| 3 |
size 598434152
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1196955962
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65be9700ffb2ead35027d047544f87600617dedade6f5c3f6c069f12bb951ba1
|
| 3 |
size 1196955962
|
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 sha256:59efe62e4ca0647678855566a69eaafc20fb9e01c9af7b6b454bf0717d7bf5f7
|
| 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 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f7bf369a63d501e73812630c6dd5a11aa21e67b93b5316e7295ec988e3218d6
|
| 3 |
size 1000
|
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": 250,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1535,6 +1535,770 @@
|
|
| 1535 |
"eval_spearman_manhattan": 0.7766391726420421,
|
| 1536 |
"eval_steps_per_second": 32.826,
|
| 1537 |
"step": 2000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1538 |
}
|
| 1539 |
],
|
| 1540 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.4058106841611997,
|
| 5 |
"eval_steps": 250,
|
| 6 |
+
"global_step": 3000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1535 |
"eval_spearman_manhattan": 0.7766391726420421,
|
| 1536 |
"eval_steps_per_second": 32.826,
|
| 1537 |
"step": 2000
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 0.9418931583880038,
|
| 1541 |
+
"grad_norm": 5.383081912994385,
|
| 1542 |
+
"learning_rate": 9.411316776007498e-06,
|
| 1543 |
+
"loss": 0.5343,
|
| 1544 |
+
"step": 2010
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.9465791940018744,
|
| 1548 |
+
"grad_norm": 5.533719539642334,
|
| 1549 |
+
"learning_rate": 9.408388003748829e-06,
|
| 1550 |
+
"loss": 0.5313,
|
| 1551 |
+
"step": 2020
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.9512652296157451,
|
| 1555 |
+
"grad_norm": 4.267172336578369,
|
| 1556 |
+
"learning_rate": 9.40545923149016e-06,
|
| 1557 |
+
"loss": 0.5172,
|
| 1558 |
+
"step": 2030
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 0.9559512652296157,
|
| 1562 |
+
"grad_norm": 4.8553009033203125,
|
| 1563 |
+
"learning_rate": 9.402530459231491e-06,
|
| 1564 |
+
"loss": 0.5104,
|
| 1565 |
+
"step": 2040
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 0.9606373008434864,
|
| 1569 |
+
"grad_norm": 6.460834503173828,
|
| 1570 |
+
"learning_rate": 9.399601686972821e-06,
|
| 1571 |
+
"loss": 0.5225,
|
| 1572 |
+
"step": 2050
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.9653233364573571,
|
| 1576 |
+
"grad_norm": 27.46290397644043,
|
| 1577 |
+
"learning_rate": 9.396672914714152e-06,
|
| 1578 |
+
"loss": 0.544,
|
| 1579 |
+
"step": 2060
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 0.9700093720712277,
|
| 1583 |
+
"grad_norm": 4.89717435836792,
|
| 1584 |
+
"learning_rate": 9.393744142455483e-06,
|
| 1585 |
+
"loss": 0.5653,
|
| 1586 |
+
"step": 2070
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.9746954076850984,
|
| 1590 |
+
"grad_norm": 4.803583145141602,
|
| 1591 |
+
"learning_rate": 9.390815370196814e-06,
|
| 1592 |
+
"loss": 0.5739,
|
| 1593 |
+
"step": 2080
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.979381443298969,
|
| 1597 |
+
"grad_norm": 4.121029853820801,
|
| 1598 |
+
"learning_rate": 9.387886597938146e-06,
|
| 1599 |
+
"loss": 0.5192,
|
| 1600 |
+
"step": 2090
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 0.9840674789128397,
|
| 1604 |
+
"grad_norm": 4.464984893798828,
|
| 1605 |
+
"learning_rate": 9.384957825679475e-06,
|
| 1606 |
+
"loss": 0.5393,
|
| 1607 |
+
"step": 2100
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 0.9887535145267105,
|
| 1611 |
+
"grad_norm": 6.364498615264893,
|
| 1612 |
+
"learning_rate": 9.382029053420806e-06,
|
| 1613 |
+
"loss": 0.5764,
|
| 1614 |
+
"step": 2110
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.993439550140581,
|
| 1618 |
+
"grad_norm": 3.743790864944458,
|
| 1619 |
+
"learning_rate": 9.379100281162138e-06,
|
| 1620 |
+
"loss": 0.5276,
|
| 1621 |
+
"step": 2120
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 0.9981255857544518,
|
| 1625 |
+
"grad_norm": 4.737389087677002,
|
| 1626 |
+
"learning_rate": 9.376171508903469e-06,
|
| 1627 |
+
"loss": 0.5211,
|
| 1628 |
+
"step": 2130
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 1.0028116213683225,
|
| 1632 |
+
"grad_norm": 3.622758626937866,
|
| 1633 |
+
"learning_rate": 9.3732427366448e-06,
|
| 1634 |
+
"loss": 0.5329,
|
| 1635 |
+
"step": 2140
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 1.007497656982193,
|
| 1639 |
+
"grad_norm": 3.5359978675842285,
|
| 1640 |
+
"learning_rate": 9.37031396438613e-06,
|
| 1641 |
+
"loss": 0.4941,
|
| 1642 |
+
"step": 2150
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 1.0121836925960637,
|
| 1646 |
+
"grad_norm": 4.669582843780518,
|
| 1647 |
+
"learning_rate": 9.36738519212746e-06,
|
| 1648 |
+
"loss": 0.4821,
|
| 1649 |
+
"step": 2160
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 1.0168697282099344,
|
| 1653 |
+
"grad_norm": 3.767122507095337,
|
| 1654 |
+
"learning_rate": 9.364456419868792e-06,
|
| 1655 |
+
"loss": 0.4886,
|
| 1656 |
+
"step": 2170
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 1.021555763823805,
|
| 1660 |
+
"grad_norm": 3.9681687355041504,
|
| 1661 |
+
"learning_rate": 9.361527647610123e-06,
|
| 1662 |
+
"loss": 0.493,
|
| 1663 |
+
"step": 2180
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 1.0262417994376758,
|
| 1667 |
+
"grad_norm": 3.389897108078003,
|
| 1668 |
+
"learning_rate": 9.358598875351454e-06,
|
| 1669 |
+
"loss": 0.4688,
|
| 1670 |
+
"step": 2190
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 1.0309278350515463,
|
| 1674 |
+
"grad_norm": 3.5152347087860107,
|
| 1675 |
+
"learning_rate": 9.355670103092784e-06,
|
| 1676 |
+
"loss": 0.4625,
|
| 1677 |
+
"step": 2200
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 1.035613870665417,
|
| 1681 |
+
"grad_norm": 3.23901629447937,
|
| 1682 |
+
"learning_rate": 9.352741330834115e-06,
|
| 1683 |
+
"loss": 0.5143,
|
| 1684 |
+
"step": 2210
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 1.0402999062792877,
|
| 1688 |
+
"grad_norm": 4.617633819580078,
|
| 1689 |
+
"learning_rate": 9.349812558575446e-06,
|
| 1690 |
+
"loss": 0.4732,
|
| 1691 |
+
"step": 2220
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 1.0449859418931584,
|
| 1695 |
+
"grad_norm": 5.245469570159912,
|
| 1696 |
+
"learning_rate": 9.346883786316777e-06,
|
| 1697 |
+
"loss": 0.5213,
|
| 1698 |
+
"step": 2230
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 1.0496719775070291,
|
| 1702 |
+
"grad_norm": 4.20419454574585,
|
| 1703 |
+
"learning_rate": 9.343955014058108e-06,
|
| 1704 |
+
"loss": 0.5042,
|
| 1705 |
+
"step": 2240
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.0543580131208996,
|
| 1709 |
+
"grad_norm": 4.6322102546691895,
|
| 1710 |
+
"learning_rate": 9.341026241799438e-06,
|
| 1711 |
+
"loss": 0.4982,
|
| 1712 |
+
"step": 2250
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 1.0543580131208996,
|
| 1716 |
+
"eval_loss": 0.05779802054166794,
|
| 1717 |
+
"eval_pearson_cosine": 0.7770314842083366,
|
| 1718 |
+
"eval_pearson_dot": 0.6498110843024136,
|
| 1719 |
+
"eval_pearson_euclidean": 0.7709013065859232,
|
| 1720 |
+
"eval_pearson_manhattan": 0.7695278239114174,
|
| 1721 |
+
"eval_runtime": 48.4856,
|
| 1722 |
+
"eval_samples_per_second": 30.937,
|
| 1723 |
+
"eval_spearman_cosine": 0.7783328375480574,
|
| 1724 |
+
"eval_spearman_dot": 0.6551905692522538,
|
| 1725 |
+
"eval_spearman_euclidean": 0.7802862933680744,
|
| 1726 |
+
"eval_spearman_manhattan": 0.7790525675974715,
|
| 1727 |
+
"eval_steps_per_second": 30.937,
|
| 1728 |
+
"step": 2250
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 1.0590440487347703,
|
| 1732 |
+
"grad_norm": 4.474431991577148,
|
| 1733 |
+
"learning_rate": 9.33809746954077e-06,
|
| 1734 |
+
"loss": 0.5227,
|
| 1735 |
+
"step": 2260
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"epoch": 1.063730084348641,
|
| 1739 |
+
"grad_norm": 4.538947105407715,
|
| 1740 |
+
"learning_rate": 9.3351686972821e-06,
|
| 1741 |
+
"loss": 0.5158,
|
| 1742 |
+
"step": 2270
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 1.0684161199625117,
|
| 1746 |
+
"grad_norm": 6.6143693923950195,
|
| 1747 |
+
"learning_rate": 9.332239925023432e-06,
|
| 1748 |
+
"loss": 0.461,
|
| 1749 |
+
"step": 2280
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 1.0731021555763824,
|
| 1753 |
+
"grad_norm": 4.316189765930176,
|
| 1754 |
+
"learning_rate": 9.329311152764761e-06,
|
| 1755 |
+
"loss": 0.5079,
|
| 1756 |
+
"step": 2290
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 1.077788191190253,
|
| 1760 |
+
"grad_norm": 4.054687976837158,
|
| 1761 |
+
"learning_rate": 9.326382380506092e-06,
|
| 1762 |
+
"loss": 0.5022,
|
| 1763 |
+
"step": 2300
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 1.0824742268041236,
|
| 1767 |
+
"grad_norm": 4.232051849365234,
|
| 1768 |
+
"learning_rate": 9.323453608247423e-06,
|
| 1769 |
+
"loss": 0.5096,
|
| 1770 |
+
"step": 2310
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 1.0871602624179943,
|
| 1774 |
+
"grad_norm": 3.7785236835479736,
|
| 1775 |
+
"learning_rate": 9.320524835988755e-06,
|
| 1776 |
+
"loss": 0.4614,
|
| 1777 |
+
"step": 2320
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 1.091846298031865,
|
| 1781 |
+
"grad_norm": 4.865905284881592,
|
| 1782 |
+
"learning_rate": 9.317596063730086e-06,
|
| 1783 |
+
"loss": 0.5135,
|
| 1784 |
+
"step": 2330
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 1.0965323336457358,
|
| 1788 |
+
"grad_norm": 4.681485176086426,
|
| 1789 |
+
"learning_rate": 9.314667291471417e-06,
|
| 1790 |
+
"loss": 0.5061,
|
| 1791 |
+
"step": 2340
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 1.1012183692596063,
|
| 1795 |
+
"grad_norm": 4.256619453430176,
|
| 1796 |
+
"learning_rate": 9.311738519212747e-06,
|
| 1797 |
+
"loss": 0.4627,
|
| 1798 |
+
"step": 2350
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 1.105904404873477,
|
| 1802 |
+
"grad_norm": 4.459606170654297,
|
| 1803 |
+
"learning_rate": 9.308809746954078e-06,
|
| 1804 |
+
"loss": 0.5171,
|
| 1805 |
+
"step": 2360
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 1.1105904404873477,
|
| 1809 |
+
"grad_norm": 4.008665084838867,
|
| 1810 |
+
"learning_rate": 9.305880974695409e-06,
|
| 1811 |
+
"loss": 0.4422,
|
| 1812 |
+
"step": 2370
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 1.1152764761012184,
|
| 1816 |
+
"grad_norm": 3.674177885055542,
|
| 1817 |
+
"learning_rate": 9.302952202436738e-06,
|
| 1818 |
+
"loss": 0.5233,
|
| 1819 |
+
"step": 2380
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 1.119962511715089,
|
| 1823 |
+
"grad_norm": 4.463940620422363,
|
| 1824 |
+
"learning_rate": 9.30002343017807e-06,
|
| 1825 |
+
"loss": 0.4731,
|
| 1826 |
+
"step": 2390
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 1.1246485473289598,
|
| 1830 |
+
"grad_norm": 3.9289097785949707,
|
| 1831 |
+
"learning_rate": 9.2970946579194e-06,
|
| 1832 |
+
"loss": 0.4869,
|
| 1833 |
+
"step": 2400
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 1.1293345829428303,
|
| 1837 |
+
"grad_norm": 4.097565174102783,
|
| 1838 |
+
"learning_rate": 9.294165885660732e-06,
|
| 1839 |
+
"loss": 0.4594,
|
| 1840 |
+
"step": 2410
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 1.134020618556701,
|
| 1844 |
+
"grad_norm": 4.55318546295166,
|
| 1845 |
+
"learning_rate": 9.291237113402063e-06,
|
| 1846 |
+
"loss": 0.494,
|
| 1847 |
+
"step": 2420
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 1.1387066541705717,
|
| 1851 |
+
"grad_norm": 4.425617694854736,
|
| 1852 |
+
"learning_rate": 9.288308341143394e-06,
|
| 1853 |
+
"loss": 0.4829,
|
| 1854 |
+
"step": 2430
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 1.1433926897844424,
|
| 1858 |
+
"grad_norm": 3.908015489578247,
|
| 1859 |
+
"learning_rate": 9.285379568884726e-06,
|
| 1860 |
+
"loss": 0.4793,
|
| 1861 |
+
"step": 2440
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 1.148078725398313,
|
| 1865 |
+
"grad_norm": 3.7293996810913086,
|
| 1866 |
+
"learning_rate": 9.282450796626055e-06,
|
| 1867 |
+
"loss": 0.5399,
|
| 1868 |
+
"step": 2450
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 1.1527647610121836,
|
| 1872 |
+
"grad_norm": 4.584681034088135,
|
| 1873 |
+
"learning_rate": 9.279522024367386e-06,
|
| 1874 |
+
"loss": 0.4479,
|
| 1875 |
+
"step": 2460
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 1.1574507966260543,
|
| 1879 |
+
"grad_norm": 4.109914302825928,
|
| 1880 |
+
"learning_rate": 9.276593252108716e-06,
|
| 1881 |
+
"loss": 0.4599,
|
| 1882 |
+
"step": 2470
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 1.162136832239925,
|
| 1886 |
+
"grad_norm": 4.446422100067139,
|
| 1887 |
+
"learning_rate": 9.273664479850047e-06,
|
| 1888 |
+
"loss": 0.4727,
|
| 1889 |
+
"step": 2480
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 1.1668228678537957,
|
| 1893 |
+
"grad_norm": 5.975160598754883,
|
| 1894 |
+
"learning_rate": 9.270735707591378e-06,
|
| 1895 |
+
"loss": 0.4509,
|
| 1896 |
+
"step": 2490
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 1.1715089034676662,
|
| 1900 |
+
"grad_norm": 4.379275321960449,
|
| 1901 |
+
"learning_rate": 9.26780693533271e-06,
|
| 1902 |
+
"loss": 0.4828,
|
| 1903 |
+
"step": 2500
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 1.1715089034676662,
|
| 1907 |
+
"eval_loss": 0.05214480683207512,
|
| 1908 |
+
"eval_pearson_cosine": 0.7792755247272061,
|
| 1909 |
+
"eval_pearson_dot": 0.6569300577465214,
|
| 1910 |
+
"eval_pearson_euclidean": 0.7718322585231894,
|
| 1911 |
+
"eval_pearson_manhattan": 0.7703922250718165,
|
| 1912 |
+
"eval_runtime": 47.8089,
|
| 1913 |
+
"eval_samples_per_second": 31.375,
|
| 1914 |
+
"eval_spearman_cosine": 0.7799819701975583,
|
| 1915 |
+
"eval_spearman_dot": 0.662507389274304,
|
| 1916 |
+
"eval_spearman_euclidean": 0.7818437831063969,
|
| 1917 |
+
"eval_spearman_manhattan": 0.7805341558401507,
|
| 1918 |
+
"eval_steps_per_second": 31.375,
|
| 1919 |
+
"step": 2500
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 1.176194939081537,
|
| 1923 |
+
"grad_norm": 3.5287399291992188,
|
| 1924 |
+
"learning_rate": 9.26487816307404e-06,
|
| 1925 |
+
"loss": 0.4591,
|
| 1926 |
+
"step": 2510
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 1.1808809746954076,
|
| 1930 |
+
"grad_norm": 3.277655601501465,
|
| 1931 |
+
"learning_rate": 9.261949390815372e-06,
|
| 1932 |
+
"loss": 0.4479,
|
| 1933 |
+
"step": 2520
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 1.1855670103092784,
|
| 1937 |
+
"grad_norm": 4.732039451599121,
|
| 1938 |
+
"learning_rate": 9.259020618556703e-06,
|
| 1939 |
+
"loss": 0.461,
|
| 1940 |
+
"step": 2530
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 1.190253045923149,
|
| 1944 |
+
"grad_norm": 4.4760966300964355,
|
| 1945 |
+
"learning_rate": 9.256091846298032e-06,
|
| 1946 |
+
"loss": 0.4652,
|
| 1947 |
+
"step": 2540
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 1.1949390815370198,
|
| 1951 |
+
"grad_norm": 7.485498428344727,
|
| 1952 |
+
"learning_rate": 9.253163074039364e-06,
|
| 1953 |
+
"loss": 0.4779,
|
| 1954 |
+
"step": 2550
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 1.1996251171508903,
|
| 1958 |
+
"grad_norm": 3.9956140518188477,
|
| 1959 |
+
"learning_rate": 9.250234301780693e-06,
|
| 1960 |
+
"loss": 0.4567,
|
| 1961 |
+
"step": 2560
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 1.204311152764761,
|
| 1965 |
+
"grad_norm": 3.547563314437866,
|
| 1966 |
+
"learning_rate": 9.247305529522024e-06,
|
| 1967 |
+
"loss": 0.4988,
|
| 1968 |
+
"step": 2570
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 1.2089971883786317,
|
| 1972 |
+
"grad_norm": 5.354389667510986,
|
| 1973 |
+
"learning_rate": 9.244376757263355e-06,
|
| 1974 |
+
"loss": 0.464,
|
| 1975 |
+
"step": 2580
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 1.2136832239925024,
|
| 1979 |
+
"grad_norm": 3.791760206222534,
|
| 1980 |
+
"learning_rate": 9.241447985004687e-06,
|
| 1981 |
+
"loss": 0.4441,
|
| 1982 |
+
"step": 2590
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 1.218369259606373,
|
| 1986 |
+
"grad_norm": 4.77889347076416,
|
| 1987 |
+
"learning_rate": 9.238519212746018e-06,
|
| 1988 |
+
"loss": 0.4655,
|
| 1989 |
+
"step": 2600
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 1.2230552952202436,
|
| 1993 |
+
"grad_norm": 5.804917335510254,
|
| 1994 |
+
"learning_rate": 9.235590440487349e-06,
|
| 1995 |
+
"loss": 0.4912,
|
| 1996 |
+
"step": 2610
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 1.2277413308341143,
|
| 2000 |
+
"grad_norm": 3.841860771179199,
|
| 2001 |
+
"learning_rate": 9.23266166822868e-06,
|
| 2002 |
+
"loss": 0.472,
|
| 2003 |
+
"step": 2620
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 1.232427366447985,
|
| 2007 |
+
"grad_norm": 4.4197540283203125,
|
| 2008 |
+
"learning_rate": 9.22973289597001e-06,
|
| 2009 |
+
"loss": 0.4821,
|
| 2010 |
+
"step": 2630
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 1.2371134020618557,
|
| 2014 |
+
"grad_norm": 5.844490051269531,
|
| 2015 |
+
"learning_rate": 9.226804123711341e-06,
|
| 2016 |
+
"loss": 0.5655,
|
| 2017 |
+
"step": 2640
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 1.2417994376757264,
|
| 2021 |
+
"grad_norm": 3.5442116260528564,
|
| 2022 |
+
"learning_rate": 9.223875351452672e-06,
|
| 2023 |
+
"loss": 0.4532,
|
| 2024 |
+
"step": 2650
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 1.246485473289597,
|
| 2028 |
+
"grad_norm": 5.259571075439453,
|
| 2029 |
+
"learning_rate": 9.220946579194002e-06,
|
| 2030 |
+
"loss": 0.4856,
|
| 2031 |
+
"step": 2660
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 1.2511715089034676,
|
| 2035 |
+
"grad_norm": 4.675846576690674,
|
| 2036 |
+
"learning_rate": 9.218017806935333e-06,
|
| 2037 |
+
"loss": 0.4576,
|
| 2038 |
+
"step": 2670
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 1.2558575445173383,
|
| 2042 |
+
"grad_norm": 5.236482620239258,
|
| 2043 |
+
"learning_rate": 9.215089034676664e-06,
|
| 2044 |
+
"loss": 0.513,
|
| 2045 |
+
"step": 2680
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 1.260543580131209,
|
| 2049 |
+
"grad_norm": 4.658278465270996,
|
| 2050 |
+
"learning_rate": 9.212160262417995e-06,
|
| 2051 |
+
"loss": 0.4734,
|
| 2052 |
+
"step": 2690
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 1.2652296157450795,
|
| 2056 |
+
"grad_norm": 3.7085494995117188,
|
| 2057 |
+
"learning_rate": 9.209231490159326e-06,
|
| 2058 |
+
"loss": 0.5279,
|
| 2059 |
+
"step": 2700
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 1.2699156513589505,
|
| 2063 |
+
"grad_norm": 3.4627673625946045,
|
| 2064 |
+
"learning_rate": 9.206302717900658e-06,
|
| 2065 |
+
"loss": 0.4773,
|
| 2066 |
+
"step": 2710
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 1.274601686972821,
|
| 2070 |
+
"grad_norm": 4.618409633636475,
|
| 2071 |
+
"learning_rate": 9.203373945641987e-06,
|
| 2072 |
+
"loss": 0.4354,
|
| 2073 |
+
"step": 2720
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 1.2792877225866917,
|
| 2077 |
+
"grad_norm": 3.1090590953826904,
|
| 2078 |
+
"learning_rate": 9.200445173383318e-06,
|
| 2079 |
+
"loss": 0.4409,
|
| 2080 |
+
"step": 2730
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 1.2839737582005624,
|
| 2084 |
+
"grad_norm": 4.328725337982178,
|
| 2085 |
+
"learning_rate": 9.19751640112465e-06,
|
| 2086 |
+
"loss": 0.4799,
|
| 2087 |
+
"step": 2740
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 1.2886597938144329,
|
| 2091 |
+
"grad_norm": 3.8362419605255127,
|
| 2092 |
+
"learning_rate": 9.194587628865979e-06,
|
| 2093 |
+
"loss": 0.5062,
|
| 2094 |
+
"step": 2750
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 1.2886597938144329,
|
| 2098 |
+
"eval_loss": 0.05263364687561989,
|
| 2099 |
+
"eval_pearson_cosine": 0.7755555336434341,
|
| 2100 |
+
"eval_pearson_dot": 0.6502184577290961,
|
| 2101 |
+
"eval_pearson_euclidean": 0.7709853609297426,
|
| 2102 |
+
"eval_pearson_manhattan": 0.769572635033791,
|
| 2103 |
+
"eval_runtime": 44.8508,
|
| 2104 |
+
"eval_samples_per_second": 33.444,
|
| 2105 |
+
"eval_spearman_cosine": 0.7765036654281985,
|
| 2106 |
+
"eval_spearman_dot": 0.6558936409143281,
|
| 2107 |
+
"eval_spearman_euclidean": 0.7808945633743188,
|
| 2108 |
+
"eval_spearman_manhattan": 0.7795729380744477,
|
| 2109 |
+
"eval_steps_per_second": 33.444,
|
| 2110 |
+
"step": 2750
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.2933458294283038,
|
| 2114 |
+
"grad_norm": 3.6972432136535645,
|
| 2115 |
+
"learning_rate": 9.19165885660731e-06,
|
| 2116 |
+
"loss": 0.488,
|
| 2117 |
+
"step": 2760
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.2980318650421743,
|
| 2121 |
+
"grad_norm": 6.73103141784668,
|
| 2122 |
+
"learning_rate": 9.188730084348641e-06,
|
| 2123 |
+
"loss": 0.4553,
|
| 2124 |
+
"step": 2770
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.302717900656045,
|
| 2128 |
+
"grad_norm": 4.371028423309326,
|
| 2129 |
+
"learning_rate": 9.185801312089973e-06,
|
| 2130 |
+
"loss": 0.4555,
|
| 2131 |
+
"step": 2780
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.3074039362699157,
|
| 2135 |
+
"grad_norm": 3.4788401126861572,
|
| 2136 |
+
"learning_rate": 9.182872539831304e-06,
|
| 2137 |
+
"loss": 0.4561,
|
| 2138 |
+
"step": 2790
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.3120899718837864,
|
| 2142 |
+
"grad_norm": 3.832277774810791,
|
| 2143 |
+
"learning_rate": 9.179943767572635e-06,
|
| 2144 |
+
"loss": 0.4838,
|
| 2145 |
+
"step": 2800
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 1.316776007497657,
|
| 2149 |
+
"grad_norm": 3.5579423904418945,
|
| 2150 |
+
"learning_rate": 9.177014995313966e-06,
|
| 2151 |
+
"loss": 0.4404,
|
| 2152 |
+
"step": 2810
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.3214620431115276,
|
| 2156 |
+
"grad_norm": 3.7768073081970215,
|
| 2157 |
+
"learning_rate": 9.174086223055296e-06,
|
| 2158 |
+
"loss": 0.4724,
|
| 2159 |
+
"step": 2820
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.3261480787253983,
|
| 2163 |
+
"grad_norm": 3.957035779953003,
|
| 2164 |
+
"learning_rate": 9.171157450796627e-06,
|
| 2165 |
+
"loss": 0.471,
|
| 2166 |
+
"step": 2830
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.330834114339269,
|
| 2170 |
+
"grad_norm": 3.6035895347595215,
|
| 2171 |
+
"learning_rate": 9.168228678537958e-06,
|
| 2172 |
+
"loss": 0.4645,
|
| 2173 |
+
"step": 2840
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 1.3355201499531397,
|
| 2177 |
+
"grad_norm": 4.358327388763428,
|
| 2178 |
+
"learning_rate": 9.165299906279288e-06,
|
| 2179 |
+
"loss": 0.4301,
|
| 2180 |
+
"step": 2850
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 1.3402061855670104,
|
| 2184 |
+
"grad_norm": 3.4666709899902344,
|
| 2185 |
+
"learning_rate": 9.162371134020619e-06,
|
| 2186 |
+
"loss": 0.4508,
|
| 2187 |
+
"step": 2860
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 1.344892221180881,
|
| 2191 |
+
"grad_norm": 3.912290096282959,
|
| 2192 |
+
"learning_rate": 9.15944236176195e-06,
|
| 2193 |
+
"loss": 0.4379,
|
| 2194 |
+
"step": 2870
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 1.3495782567947516,
|
| 2198 |
+
"grad_norm": 4.305796146392822,
|
| 2199 |
+
"learning_rate": 9.156513589503281e-06,
|
| 2200 |
+
"loss": 0.4194,
|
| 2201 |
+
"step": 2880
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 1.3542642924086223,
|
| 2205 |
+
"grad_norm": 4.231681823730469,
|
| 2206 |
+
"learning_rate": 9.153584817244612e-06,
|
| 2207 |
+
"loss": 0.4017,
|
| 2208 |
+
"step": 2890
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 1.358950328022493,
|
| 2212 |
+
"grad_norm": 4.43821382522583,
|
| 2213 |
+
"learning_rate": 9.150656044985944e-06,
|
| 2214 |
+
"loss": 0.4185,
|
| 2215 |
+
"step": 2900
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 1.3636363636363638,
|
| 2219 |
+
"grad_norm": 4.922164440155029,
|
| 2220 |
+
"learning_rate": 9.147727272727273e-06,
|
| 2221 |
+
"loss": 0.5199,
|
| 2222 |
+
"step": 2910
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 1.3683223992502342,
|
| 2226 |
+
"grad_norm": 4.577489852905273,
|
| 2227 |
+
"learning_rate": 9.144798500468604e-06,
|
| 2228 |
+
"loss": 0.4237,
|
| 2229 |
+
"step": 2920
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 1.373008434864105,
|
| 2233 |
+
"grad_norm": 3.9537651538848877,
|
| 2234 |
+
"learning_rate": 9.141869728209935e-06,
|
| 2235 |
+
"loss": 0.4888,
|
| 2236 |
+
"step": 2930
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 1.3776944704779757,
|
| 2240 |
+
"grad_norm": 4.165870189666748,
|
| 2241 |
+
"learning_rate": 9.138940955951267e-06,
|
| 2242 |
+
"loss": 0.4476,
|
| 2243 |
+
"step": 2940
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 1.3823805060918464,
|
| 2247 |
+
"grad_norm": 4.492893218994141,
|
| 2248 |
+
"learning_rate": 9.136012183692596e-06,
|
| 2249 |
+
"loss": 0.5159,
|
| 2250 |
+
"step": 2950
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 1.387066541705717,
|
| 2254 |
+
"grad_norm": 3.847490072250366,
|
| 2255 |
+
"learning_rate": 9.133083411433927e-06,
|
| 2256 |
+
"loss": 0.4497,
|
| 2257 |
+
"step": 2960
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 1.3917525773195876,
|
| 2261 |
+
"grad_norm": 6.766137599945068,
|
| 2262 |
+
"learning_rate": 9.130154639175258e-06,
|
| 2263 |
+
"loss": 0.4379,
|
| 2264 |
+
"step": 2970
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 1.3964386129334583,
|
| 2268 |
+
"grad_norm": 3.9198007583618164,
|
| 2269 |
+
"learning_rate": 9.12722586691659e-06,
|
| 2270 |
+
"loss": 0.4519,
|
| 2271 |
+
"step": 2980
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 1.401124648547329,
|
| 2275 |
+
"grad_norm": 3.67480731010437,
|
| 2276 |
+
"learning_rate": 9.124297094657921e-06,
|
| 2277 |
+
"loss": 0.4108,
|
| 2278 |
+
"step": 2990
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 1.4058106841611997,
|
| 2282 |
+
"grad_norm": 3.3013832569122314,
|
| 2283 |
+
"learning_rate": 9.12136832239925e-06,
|
| 2284 |
+
"loss": 0.433,
|
| 2285 |
+
"step": 3000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 1.4058106841611997,
|
| 2289 |
+
"eval_loss": 0.0497601218521595,
|
| 2290 |
+
"eval_pearson_cosine": 0.7834985989633054,
|
| 2291 |
+
"eval_pearson_dot": 0.6669524421664974,
|
| 2292 |
+
"eval_pearson_euclidean": 0.7743874834934843,
|
| 2293 |
+
"eval_pearson_manhattan": 0.7730376146204847,
|
| 2294 |
+
"eval_runtime": 47.8141,
|
| 2295 |
+
"eval_samples_per_second": 31.371,
|
| 2296 |
+
"eval_spearman_cosine": 0.7845889452017747,
|
| 2297 |
+
"eval_spearman_dot": 0.6729435548765089,
|
| 2298 |
+
"eval_spearman_euclidean": 0.784591658726837,
|
| 2299 |
+
"eval_spearman_manhattan": 0.7832975474858643,
|
| 2300 |
+
"eval_steps_per_second": 31.371,
|
| 2301 |
+
"step": 3000
|
| 2302 |
}
|
| 2303 |
],
|
| 2304 |
"logging_steps": 10,
|