Training in progress, step 300, checkpoint
Browse files- last-checkpoint/optimizer_0/.metadata +0 -0
- last-checkpoint/optimizer_0/__0_0.distcp +1 -1
- last-checkpoint/optimizer_0/__1_0.distcp +1 -1
- last-checkpoint/optimizer_0/__2_0.distcp +1 -1
- last-checkpoint/optimizer_0/__3_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/.metadata +0 -0
- last-checkpoint/pytorch_model_fsdp_0/__0_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__1_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__2_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__3_0.distcp +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +763 -3
last-checkpoint/optimizer_0/.metadata
CHANGED
Binary files a/last-checkpoint/optimizer_0/.metadata and b/last-checkpoint/optimizer_0/.metadata differ
|
|
last-checkpoint/optimizer_0/__0_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 13934748
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3720c9d7c3d147e3c6a4c31b748a401804a80a6f28683e7c9983f2e8c8a0f20a
|
3 |
size 13934748
|
last-checkpoint/optimizer_0/__1_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 13999412
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3638a66f29f136c13174bee0dd43e693f5fc102e10bee4ca9b5d7060756ced7c
|
3 |
size 13999412
|
last-checkpoint/optimizer_0/__2_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 13990904
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:59454dac3aeba9c46e0b8ed50eb871c1d98271ecbabd9754c8618cdc65584ad9
|
3 |
size 13990904
|
last-checkpoint/optimizer_0/__3_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 13990904
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:93ca40f1a301b546ff3c8e51d511eb49571367df816e5ac6c43d69b073ba1e35
|
3 |
size 13990904
|
last-checkpoint/pytorch_model_fsdp_0/.metadata
CHANGED
Binary files a/last-checkpoint/pytorch_model_fsdp_0/.metadata and b/last-checkpoint/pytorch_model_fsdp_0/.metadata differ
|
|
last-checkpoint/pytorch_model_fsdp_0/__0_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6966784
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7e305eb64bbd004a3712d4d0f3b65560d9c0d8b3920c2789be35be33fef333cd
|
3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__1_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6966784
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:16b6721393030c99c98218e1bcc44fa93cc347e7c920295cebba18bf69ebf311
|
3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__2_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6966784
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f9c2af98820448e537c2aa09618f8c2299b2ed8c9040abdad7cc23d455398738
|
3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__3_0.distcp
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6966784
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fa8dfa614952af057b305d24646b204cdfa9b2fbf5610aa112de72f4d1903dd4
|
3 |
size 6966784
|
last-checkpoint/rng_state_0.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14960
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6185843c50764de20922699c89193c33e1e13037719a5d55479aa190e715e4fc
|
3 |
size 14960
|
last-checkpoint/rng_state_1.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14960
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:35e51ecf57078c2d652964726d8abc8157e10e9fdddf8cacb5700305b465147a
|
3 |
size 14960
|
last-checkpoint/rng_state_2.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14960
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ebe10fe55b3a58ae13fa7a58fca8f2486fa82c4aa360522ee9cde43cc43ba473
|
3 |
size 14960
|
last-checkpoint/rng_state_3.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14960
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fcda73faaa8d5a9ab0a72d2fef1c1af0341c8e7f8ec0eede744acae39dd22f43
|
3 |
size 14960
|
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:a3e47edb1b664bc04c493b0996774157c1ffdb9f0b12df515a0b32829d748704
|
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": 20,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -1539,6 +1539,766 @@
|
|
1539 |
"eval_samples_per_second": 6.49,
|
1540 |
"eval_steps_per_second": 0.245,
|
1541 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1542 |
}
|
1543 |
],
|
1544 |
"logging_steps": 1,
|
@@ -1558,7 +2318,7 @@
|
|
1558 |
"attributes": {}
|
1559 |
}
|
1560 |
},
|
1561 |
-
"total_flos":
|
1562 |
"train_batch_size": 8,
|
1563 |
"trial_name": null,
|
1564 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.8620689655172413,
|
5 |
"eval_steps": 20,
|
6 |
+
"global_step": 300,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
1539 |
"eval_samples_per_second": 6.49,
|
1540 |
"eval_steps_per_second": 0.245,
|
1541 |
"step": 200
|
1542 |
+
},
|
1543 |
+
{
|
1544 |
+
"epoch": 0.5775862068965517,
|
1545 |
+
"grad_norm": 3.761861801147461,
|
1546 |
+
"learning_rate": 9.047925700872552e-06,
|
1547 |
+
"loss": 0.3089,
|
1548 |
+
"step": 201
|
1549 |
+
},
|
1550 |
+
{
|
1551 |
+
"epoch": 0.5804597701149425,
|
1552 |
+
"grad_norm": 2.631603240966797,
|
1553 |
+
"learning_rate": 8.948060899634846e-06,
|
1554 |
+
"loss": 0.2943,
|
1555 |
+
"step": 202
|
1556 |
+
},
|
1557 |
+
{
|
1558 |
+
"epoch": 0.5833333333333334,
|
1559 |
+
"grad_norm": 2.613485336303711,
|
1560 |
+
"learning_rate": 8.848302072078762e-06,
|
1561 |
+
"loss": 0.2866,
|
1562 |
+
"step": 203
|
1563 |
+
},
|
1564 |
+
{
|
1565 |
+
"epoch": 0.5862068965517241,
|
1566 |
+
"grad_norm": 2.1991889476776123,
|
1567 |
+
"learning_rate": 8.748659268035339e-06,
|
1568 |
+
"loss": 0.2749,
|
1569 |
+
"step": 204
|
1570 |
+
},
|
1571 |
+
{
|
1572 |
+
"epoch": 0.5890804597701149,
|
1573 |
+
"grad_norm": 2.7884528636932373,
|
1574 |
+
"learning_rate": 8.649142525647271e-06,
|
1575 |
+
"loss": 0.3378,
|
1576 |
+
"step": 205
|
1577 |
+
},
|
1578 |
+
{
|
1579 |
+
"epoch": 0.5919540229885057,
|
1580 |
+
"grad_norm": 4.045213222503662,
|
1581 |
+
"learning_rate": 8.549761870357633e-06,
|
1582 |
+
"loss": 0.3398,
|
1583 |
+
"step": 206
|
1584 |
+
},
|
1585 |
+
{
|
1586 |
+
"epoch": 0.5948275862068966,
|
1587 |
+
"grad_norm": 2.83146595954895,
|
1588 |
+
"learning_rate": 8.450527313899923e-06,
|
1589 |
+
"loss": 0.2951,
|
1590 |
+
"step": 207
|
1591 |
+
},
|
1592 |
+
{
|
1593 |
+
"epoch": 0.5977011494252874,
|
1594 |
+
"grad_norm": 2.664379358291626,
|
1595 |
+
"learning_rate": 8.351448853289448e-06,
|
1596 |
+
"loss": 0.3452,
|
1597 |
+
"step": 208
|
1598 |
+
},
|
1599 |
+
{
|
1600 |
+
"epoch": 0.6005747126436781,
|
1601 |
+
"grad_norm": 2.2802395820617676,
|
1602 |
+
"learning_rate": 8.25253646981622e-06,
|
1603 |
+
"loss": 0.2993,
|
1604 |
+
"step": 209
|
1605 |
+
},
|
1606 |
+
{
|
1607 |
+
"epoch": 0.603448275862069,
|
1608 |
+
"grad_norm": 2.1689095497131348,
|
1609 |
+
"learning_rate": 8.153800128039441e-06,
|
1610 |
+
"loss": 0.2847,
|
1611 |
+
"step": 210
|
1612 |
+
},
|
1613 |
+
{
|
1614 |
+
"epoch": 0.6063218390804598,
|
1615 |
+
"grad_norm": 3.4026296138763428,
|
1616 |
+
"learning_rate": 8.05524977478364e-06,
|
1617 |
+
"loss": 0.3192,
|
1618 |
+
"step": 211
|
1619 |
+
},
|
1620 |
+
{
|
1621 |
+
"epoch": 0.6091954022988506,
|
1622 |
+
"grad_norm": 3.272836923599243,
|
1623 |
+
"learning_rate": 7.956895338136618e-06,
|
1624 |
+
"loss": 0.3129,
|
1625 |
+
"step": 212
|
1626 |
+
},
|
1627 |
+
{
|
1628 |
+
"epoch": 0.6120689655172413,
|
1629 |
+
"grad_norm": 4.036294460296631,
|
1630 |
+
"learning_rate": 7.858746726449309e-06,
|
1631 |
+
"loss": 0.3199,
|
1632 |
+
"step": 213
|
1633 |
+
},
|
1634 |
+
{
|
1635 |
+
"epoch": 0.6149425287356322,
|
1636 |
+
"grad_norm": 2.3944919109344482,
|
1637 |
+
"learning_rate": 7.760813827337555e-06,
|
1638 |
+
"loss": 0.2513,
|
1639 |
+
"step": 214
|
1640 |
+
},
|
1641 |
+
{
|
1642 |
+
"epoch": 0.617816091954023,
|
1643 |
+
"grad_norm": 4.011943340301514,
|
1644 |
+
"learning_rate": 7.663106506686057e-06,
|
1645 |
+
"loss": 0.3026,
|
1646 |
+
"step": 215
|
1647 |
+
},
|
1648 |
+
{
|
1649 |
+
"epoch": 0.6206896551724138,
|
1650 |
+
"grad_norm": 2.424299716949463,
|
1651 |
+
"learning_rate": 7.565634607654453e-06,
|
1652 |
+
"loss": 0.3029,
|
1653 |
+
"step": 216
|
1654 |
+
},
|
1655 |
+
{
|
1656 |
+
"epoch": 0.6235632183908046,
|
1657 |
+
"grad_norm": 5.497297286987305,
|
1658 |
+
"learning_rate": 7.468407949685695e-06,
|
1659 |
+
"loss": 0.3601,
|
1660 |
+
"step": 217
|
1661 |
+
},
|
1662 |
+
{
|
1663 |
+
"epoch": 0.6264367816091954,
|
1664 |
+
"grad_norm": 2.9515810012817383,
|
1665 |
+
"learning_rate": 7.371436327516854e-06,
|
1666 |
+
"loss": 0.328,
|
1667 |
+
"step": 218
|
1668 |
+
},
|
1669 |
+
{
|
1670 |
+
"epoch": 0.6293103448275862,
|
1671 |
+
"grad_norm": 3.359109878540039,
|
1672 |
+
"learning_rate": 7.274729510192367e-06,
|
1673 |
+
"loss": 0.3104,
|
1674 |
+
"step": 219
|
1675 |
+
},
|
1676 |
+
{
|
1677 |
+
"epoch": 0.632183908045977,
|
1678 |
+
"grad_norm": 3.1870110034942627,
|
1679 |
+
"learning_rate": 7.1782972400798825e-06,
|
1680 |
+
"loss": 0.34,
|
1681 |
+
"step": 220
|
1682 |
+
},
|
1683 |
+
{
|
1684 |
+
"epoch": 0.632183908045977,
|
1685 |
+
"eval_accuracy": 0.8688118811881188,
|
1686 |
+
"eval_f1": 0.7464114832535885,
|
1687 |
+
"eval_loss": 0.31775230169296265,
|
1688 |
+
"eval_precision": 0.7572815533980582,
|
1689 |
+
"eval_recall": 0.7358490566037735,
|
1690 |
+
"eval_runtime": 17.2507,
|
1691 |
+
"eval_samples_per_second": 6.145,
|
1692 |
+
"eval_steps_per_second": 0.232,
|
1693 |
+
"step": 220
|
1694 |
+
},
|
1695 |
+
{
|
1696 |
+
"epoch": 0.6350574712643678,
|
1697 |
+
"grad_norm": 2.9183595180511475,
|
1698 |
+
"learning_rate": 7.082149231888833e-06,
|
1699 |
+
"loss": 0.2827,
|
1700 |
+
"step": 221
|
1701 |
+
},
|
1702 |
+
{
|
1703 |
+
"epoch": 0.6379310344827587,
|
1704 |
+
"grad_norm": 4.59254789352417,
|
1705 |
+
"learning_rate": 6.986295171691727e-06,
|
1706 |
+
"loss": 0.3284,
|
1707 |
+
"step": 222
|
1708 |
+
},
|
1709 |
+
{
|
1710 |
+
"epoch": 0.6408045977011494,
|
1711 |
+
"grad_norm": 2.9099929332733154,
|
1712 |
+
"learning_rate": 6.890744715948388e-06,
|
1713 |
+
"loss": 0.2972,
|
1714 |
+
"step": 223
|
1715 |
+
},
|
1716 |
+
{
|
1717 |
+
"epoch": 0.6436781609195402,
|
1718 |
+
"grad_norm": 2.407136917114258,
|
1719 |
+
"learning_rate": 6.795507490533142e-06,
|
1720 |
+
"loss": 0.2973,
|
1721 |
+
"step": 224
|
1722 |
+
},
|
1723 |
+
{
|
1724 |
+
"epoch": 0.646551724137931,
|
1725 |
+
"grad_norm": 2.786597728729248,
|
1726 |
+
"learning_rate": 6.700593089765086e-06,
|
1727 |
+
"loss": 0.3426,
|
1728 |
+
"step": 225
|
1729 |
+
},
|
1730 |
+
{
|
1731 |
+
"epoch": 0.6494252873563219,
|
1732 |
+
"grad_norm": 2.642282485961914,
|
1733 |
+
"learning_rate": 6.606011075441556e-06,
|
1734 |
+
"loss": 0.3201,
|
1735 |
+
"step": 226
|
1736 |
+
},
|
1737 |
+
{
|
1738 |
+
"epoch": 0.6522988505747126,
|
1739 |
+
"grad_norm": 2.1382575035095215,
|
1740 |
+
"learning_rate": 6.511770975874862e-06,
|
1741 |
+
"loss": 0.2767,
|
1742 |
+
"step": 227
|
1743 |
+
},
|
1744 |
+
{
|
1745 |
+
"epoch": 0.6551724137931034,
|
1746 |
+
"grad_norm": 4.03010892868042,
|
1747 |
+
"learning_rate": 6.417882284932373e-06,
|
1748 |
+
"loss": 0.2742,
|
1749 |
+
"step": 228
|
1750 |
+
},
|
1751 |
+
{
|
1752 |
+
"epoch": 0.6580459770114943,
|
1753 |
+
"grad_norm": 2.754526138305664,
|
1754 |
+
"learning_rate": 6.324354461080121e-06,
|
1755 |
+
"loss": 0.2672,
|
1756 |
+
"step": 229
|
1757 |
+
},
|
1758 |
+
{
|
1759 |
+
"epoch": 0.6609195402298851,
|
1760 |
+
"grad_norm": 2.4932737350463867,
|
1761 |
+
"learning_rate": 6.231196926429913e-06,
|
1762 |
+
"loss": 0.2835,
|
1763 |
+
"step": 230
|
1764 |
+
},
|
1765 |
+
{
|
1766 |
+
"epoch": 0.6637931034482759,
|
1767 |
+
"grad_norm": 4.364743232727051,
|
1768 |
+
"learning_rate": 6.138419065790169e-06,
|
1769 |
+
"loss": 0.3079,
|
1770 |
+
"step": 231
|
1771 |
+
},
|
1772 |
+
{
|
1773 |
+
"epoch": 0.6666666666666666,
|
1774 |
+
"grad_norm": 2.9517085552215576,
|
1775 |
+
"learning_rate": 6.046030225720456e-06,
|
1776 |
+
"loss": 0.2422,
|
1777 |
+
"step": 232
|
1778 |
+
},
|
1779 |
+
{
|
1780 |
+
"epoch": 0.6695402298850575,
|
1781 |
+
"grad_norm": 2.8469698429107666,
|
1782 |
+
"learning_rate": 5.95403971358991e-06,
|
1783 |
+
"loss": 0.2641,
|
1784 |
+
"step": 233
|
1785 |
+
},
|
1786 |
+
{
|
1787 |
+
"epoch": 0.6724137931034483,
|
1788 |
+
"grad_norm": 3.698885440826416,
|
1789 |
+
"learning_rate": 5.86245679663962e-06,
|
1790 |
+
"loss": 0.3572,
|
1791 |
+
"step": 234
|
1792 |
+
},
|
1793 |
+
{
|
1794 |
+
"epoch": 0.6752873563218391,
|
1795 |
+
"grad_norm": 2.205653667449951,
|
1796 |
+
"learning_rate": 5.7712907010490036e-06,
|
1797 |
+
"loss": 0.252,
|
1798 |
+
"step": 235
|
1799 |
+
},
|
1800 |
+
{
|
1801 |
+
"epoch": 0.6781609195402298,
|
1802 |
+
"grad_norm": 3.1082050800323486,
|
1803 |
+
"learning_rate": 5.680550611006372e-06,
|
1804 |
+
"loss": 0.3062,
|
1805 |
+
"step": 236
|
1806 |
+
},
|
1807 |
+
{
|
1808 |
+
"epoch": 0.6810344827586207,
|
1809 |
+
"grad_norm": 4.51684045791626,
|
1810 |
+
"learning_rate": 5.590245667783701e-06,
|
1811 |
+
"loss": 0.281,
|
1812 |
+
"step": 237
|
1813 |
+
},
|
1814 |
+
{
|
1815 |
+
"epoch": 0.6839080459770115,
|
1816 |
+
"grad_norm": 2.53916335105896,
|
1817 |
+
"learning_rate": 5.5003849688157075e-06,
|
1818 |
+
"loss": 0.3273,
|
1819 |
+
"step": 238
|
1820 |
+
},
|
1821 |
+
{
|
1822 |
+
"epoch": 0.6867816091954023,
|
1823 |
+
"grad_norm": 3.1908535957336426,
|
1824 |
+
"learning_rate": 5.4109775667833866e-06,
|
1825 |
+
"loss": 0.3091,
|
1826 |
+
"step": 239
|
1827 |
+
},
|
1828 |
+
{
|
1829 |
+
"epoch": 0.6896551724137931,
|
1830 |
+
"grad_norm": 2.92702579498291,
|
1831 |
+
"learning_rate": 5.322032468702037e-06,
|
1832 |
+
"loss": 0.2969,
|
1833 |
+
"step": 240
|
1834 |
+
},
|
1835 |
+
{
|
1836 |
+
"epoch": 0.6896551724137931,
|
1837 |
+
"eval_accuracy": 0.8564356435643564,
|
1838 |
+
"eval_f1": 0.7289719626168224,
|
1839 |
+
"eval_loss": 0.3177642822265625,
|
1840 |
+
"eval_precision": 0.7222222222222222,
|
1841 |
+
"eval_recall": 0.7358490566037735,
|
1842 |
+
"eval_runtime": 16.3833,
|
1843 |
+
"eval_samples_per_second": 6.47,
|
1844 |
+
"eval_steps_per_second": 0.244,
|
1845 |
+
"step": 240
|
1846 |
+
},
|
1847 |
+
{
|
1848 |
+
"epoch": 0.6925287356321839,
|
1849 |
+
"grad_norm": 2.631377696990967,
|
1850 |
+
"learning_rate": 5.233558635013842e-06,
|
1851 |
+
"loss": 0.3108,
|
1852 |
+
"step": 241
|
1853 |
+
},
|
1854 |
+
{
|
1855 |
+
"epoch": 0.6954022988505747,
|
1856 |
+
"grad_norm": 2.5703439712524414,
|
1857 |
+
"learning_rate": 5.145564978685234e-06,
|
1858 |
+
"loss": 0.2965,
|
1859 |
+
"step": 242
|
1860 |
+
},
|
1861 |
+
{
|
1862 |
+
"epoch": 0.6982758620689655,
|
1863 |
+
"grad_norm": 5.4829254150390625,
|
1864 |
+
"learning_rate": 5.058060364308965e-06,
|
1865 |
+
"loss": 0.3302,
|
1866 |
+
"step": 243
|
1867 |
+
},
|
1868 |
+
{
|
1869 |
+
"epoch": 0.7011494252873564,
|
1870 |
+
"grad_norm": 4.077176570892334,
|
1871 |
+
"learning_rate": 4.971053607211069e-06,
|
1872 |
+
"loss": 0.3833,
|
1873 |
+
"step": 244
|
1874 |
+
},
|
1875 |
+
{
|
1876 |
+
"epoch": 0.7040229885057471,
|
1877 |
+
"grad_norm": 3.5669100284576416,
|
1878 |
+
"learning_rate": 4.884553472562809e-06,
|
1879 |
+
"loss": 0.3203,
|
1880 |
+
"step": 245
|
1881 |
+
},
|
1882 |
+
{
|
1883 |
+
"epoch": 0.7068965517241379,
|
1884 |
+
"grad_norm": 2.850348472595215,
|
1885 |
+
"learning_rate": 4.7985686744976714e-06,
|
1886 |
+
"loss": 0.2846,
|
1887 |
+
"step": 246
|
1888 |
+
},
|
1889 |
+
{
|
1890 |
+
"epoch": 0.7097701149425287,
|
1891 |
+
"grad_norm": 3.9147164821624756,
|
1892 |
+
"learning_rate": 4.713107875233459e-06,
|
1893 |
+
"loss": 0.3315,
|
1894 |
+
"step": 247
|
1895 |
+
},
|
1896 |
+
{
|
1897 |
+
"epoch": 0.7126436781609196,
|
1898 |
+
"grad_norm": 3.5606236457824707,
|
1899 |
+
"learning_rate": 4.628179684199685e-06,
|
1900 |
+
"loss": 0.3101,
|
1901 |
+
"step": 248
|
1902 |
+
},
|
1903 |
+
{
|
1904 |
+
"epoch": 0.7155172413793104,
|
1905 |
+
"grad_norm": 2.9054181575775146,
|
1906 |
+
"learning_rate": 4.543792657170228e-06,
|
1907 |
+
"loss": 0.3086,
|
1908 |
+
"step": 249
|
1909 |
+
},
|
1910 |
+
{
|
1911 |
+
"epoch": 0.7183908045977011,
|
1912 |
+
"grad_norm": 2.9038889408111572,
|
1913 |
+
"learning_rate": 4.459955295401415e-06,
|
1914 |
+
"loss": 0.2979,
|
1915 |
+
"step": 250
|
1916 |
+
},
|
1917 |
+
{
|
1918 |
+
"epoch": 0.7212643678160919,
|
1919 |
+
"grad_norm": 2.952456474304199,
|
1920 |
+
"learning_rate": 4.376676044775601e-06,
|
1921 |
+
"loss": 0.3221,
|
1922 |
+
"step": 251
|
1923 |
+
},
|
1924 |
+
{
|
1925 |
+
"epoch": 0.7241379310344828,
|
1926 |
+
"grad_norm": 2.3187882900238037,
|
1927 |
+
"learning_rate": 4.293963294950313e-06,
|
1928 |
+
"loss": 0.296,
|
1929 |
+
"step": 252
|
1930 |
+
},
|
1931 |
+
{
|
1932 |
+
"epoch": 0.7270114942528736,
|
1933 |
+
"grad_norm": 4.975540637969971,
|
1934 |
+
"learning_rate": 4.211825378513066e-06,
|
1935 |
+
"loss": 0.3873,
|
1936 |
+
"step": 253
|
1937 |
+
},
|
1938 |
+
{
|
1939 |
+
"epoch": 0.7298850574712644,
|
1940 |
+
"grad_norm": 2.431337833404541,
|
1941 |
+
"learning_rate": 4.130270570141931e-06,
|
1942 |
+
"loss": 0.3386,
|
1943 |
+
"step": 254
|
1944 |
+
},
|
1945 |
+
{
|
1946 |
+
"epoch": 0.7327586206896551,
|
1947 |
+
"grad_norm": 3.258333921432495,
|
1948 |
+
"learning_rate": 4.0493070857719305e-06,
|
1949 |
+
"loss": 0.3402,
|
1950 |
+
"step": 255
|
1951 |
+
},
|
1952 |
+
{
|
1953 |
+
"epoch": 0.735632183908046,
|
1954 |
+
"grad_norm": 2.442279577255249,
|
1955 |
+
"learning_rate": 3.968943081767358e-06,
|
1956 |
+
"loss": 0.2268,
|
1957 |
+
"step": 256
|
1958 |
+
},
|
1959 |
+
{
|
1960 |
+
"epoch": 0.7385057471264368,
|
1961 |
+
"grad_norm": 3.3889667987823486,
|
1962 |
+
"learning_rate": 3.889186654100089e-06,
|
1963 |
+
"loss": 0.3008,
|
1964 |
+
"step": 257
|
1965 |
+
},
|
1966 |
+
{
|
1967 |
+
"epoch": 0.7413793103448276,
|
1968 |
+
"grad_norm": 2.3388214111328125,
|
1969 |
+
"learning_rate": 3.81004583753399e-06,
|
1970 |
+
"loss": 0.3399,
|
1971 |
+
"step": 258
|
1972 |
+
},
|
1973 |
+
{
|
1974 |
+
"epoch": 0.7442528735632183,
|
1975 |
+
"grad_norm": 2.363194704055786,
|
1976 |
+
"learning_rate": 3.7315286048154862e-06,
|
1977 |
+
"loss": 0.2797,
|
1978 |
+
"step": 259
|
1979 |
+
},
|
1980 |
+
{
|
1981 |
+
"epoch": 0.7471264367816092,
|
1982 |
+
"grad_norm": 3.6801648139953613,
|
1983 |
+
"learning_rate": 3.6536428658703594e-06,
|
1984 |
+
"loss": 0.3179,
|
1985 |
+
"step": 260
|
1986 |
+
},
|
1987 |
+
{
|
1988 |
+
"epoch": 0.7471264367816092,
|
1989 |
+
"eval_accuracy": 0.8663366336633663,
|
1990 |
+
"eval_f1": 0.7452830188679245,
|
1991 |
+
"eval_loss": 0.3128357529640198,
|
1992 |
+
"eval_precision": 0.7452830188679245,
|
1993 |
+
"eval_recall": 0.7452830188679245,
|
1994 |
+
"eval_runtime": 16.9937,
|
1995 |
+
"eval_samples_per_second": 6.238,
|
1996 |
+
"eval_steps_per_second": 0.235,
|
1997 |
+
"step": 260
|
1998 |
+
},
|
1999 |
+
{
|
2000 |
+
"epoch": 0.75,
|
2001 |
+
"grad_norm": 3.5845208168029785,
|
2002 |
+
"learning_rate": 3.576396467006925e-06,
|
2003 |
+
"loss": 0.3121,
|
2004 |
+
"step": 261
|
2005 |
+
},
|
2006 |
+
{
|
2007 |
+
"epoch": 0.7528735632183908,
|
2008 |
+
"grad_norm": 2.5115549564361572,
|
2009 |
+
"learning_rate": 3.4997971901255588e-06,
|
2010 |
+
"loss": 0.2695,
|
2011 |
+
"step": 262
|
2012 |
+
},
|
2013 |
+
{
|
2014 |
+
"epoch": 0.7557471264367817,
|
2015 |
+
"grad_norm": 3.1949312686920166,
|
2016 |
+
"learning_rate": 3.4238527519347353e-06,
|
2017 |
+
"loss": 0.29,
|
2018 |
+
"step": 263
|
2019 |
+
},
|
2020 |
+
{
|
2021 |
+
"epoch": 0.7586206896551724,
|
2022 |
+
"grad_norm": 3.134657859802246,
|
2023 |
+
"learning_rate": 3.3485708031736698e-06,
|
2024 |
+
"loss": 0.2959,
|
2025 |
+
"step": 264
|
2026 |
+
},
|
2027 |
+
{
|
2028 |
+
"epoch": 0.7614942528735632,
|
2029 |
+
"grad_norm": 2.359828472137451,
|
2030 |
+
"learning_rate": 3.2739589278415252e-06,
|
2031 |
+
"loss": 0.299,
|
2032 |
+
"step": 265
|
2033 |
+
},
|
2034 |
+
{
|
2035 |
+
"epoch": 0.764367816091954,
|
2036 |
+
"grad_norm": 2.662598133087158,
|
2037 |
+
"learning_rate": 3.2000246424334315e-06,
|
2038 |
+
"loss": 0.2887,
|
2039 |
+
"step": 266
|
2040 |
+
},
|
2041 |
+
{
|
2042 |
+
"epoch": 0.7672413793103449,
|
2043 |
+
"grad_norm": 2.822681188583374,
|
2044 |
+
"learning_rate": 3.1267753951832523e-06,
|
2045 |
+
"loss": 0.3337,
|
2046 |
+
"step": 267
|
2047 |
+
},
|
2048 |
+
{
|
2049 |
+
"epoch": 0.7701149425287356,
|
2050 |
+
"grad_norm": 3.435675859451294,
|
2051 |
+
"learning_rate": 3.0542185653132216e-06,
|
2052 |
+
"loss": 0.2431,
|
2053 |
+
"step": 268
|
2054 |
+
},
|
2055 |
+
{
|
2056 |
+
"epoch": 0.7729885057471264,
|
2057 |
+
"grad_norm": 3.8508052825927734,
|
2058 |
+
"learning_rate": 2.982361462290575e-06,
|
2059 |
+
"loss": 0.3625,
|
2060 |
+
"step": 269
|
2061 |
+
},
|
2062 |
+
{
|
2063 |
+
"epoch": 0.7758620689655172,
|
2064 |
+
"grad_norm": 3.24882435798645,
|
2065 |
+
"learning_rate": 2.9112113250911844e-06,
|
2066 |
+
"loss": 0.3255,
|
2067 |
+
"step": 270
|
2068 |
+
},
|
2069 |
+
{
|
2070 |
+
"epoch": 0.7787356321839081,
|
2071 |
+
"grad_norm": 3.215721368789673,
|
2072 |
+
"learning_rate": 2.8407753214702694e-06,
|
2073 |
+
"loss": 0.3055,
|
2074 |
+
"step": 271
|
2075 |
+
},
|
2076 |
+
{
|
2077 |
+
"epoch": 0.7816091954022989,
|
2078 |
+
"grad_norm": 3.5768065452575684,
|
2079 |
+
"learning_rate": 2.7710605472403373e-06,
|
2080 |
+
"loss": 0.2593,
|
2081 |
+
"step": 272
|
2082 |
+
},
|
2083 |
+
{
|
2084 |
+
"epoch": 0.7844827586206896,
|
2085 |
+
"grad_norm": 3.4842770099639893,
|
2086 |
+
"learning_rate": 2.702074025556327e-06,
|
2087 |
+
"loss": 0.3183,
|
2088 |
+
"step": 273
|
2089 |
+
},
|
2090 |
+
{
|
2091 |
+
"epoch": 0.7873563218390804,
|
2092 |
+
"grad_norm": 2.8685038089752197,
|
2093 |
+
"learning_rate": 2.6338227062080924e-06,
|
2094 |
+
"loss": 0.2674,
|
2095 |
+
"step": 274
|
2096 |
+
},
|
2097 |
+
{
|
2098 |
+
"epoch": 0.7902298850574713,
|
2099 |
+
"grad_norm": 3.008521318435669,
|
2100 |
+
"learning_rate": 2.566313464920265e-06,
|
2101 |
+
"loss": 0.2944,
|
2102 |
+
"step": 275
|
2103 |
+
},
|
2104 |
+
{
|
2105 |
+
"epoch": 0.7931034482758621,
|
2106 |
+
"grad_norm": 2.9339377880096436,
|
2107 |
+
"learning_rate": 2.4995531026595952e-06,
|
2108 |
+
"loss": 0.295,
|
2109 |
+
"step": 276
|
2110 |
+
},
|
2111 |
+
{
|
2112 |
+
"epoch": 0.7959770114942529,
|
2113 |
+
"grad_norm": 4.123067378997803,
|
2114 |
+
"learning_rate": 2.4335483449498053e-06,
|
2115 |
+
"loss": 0.2295,
|
2116 |
+
"step": 277
|
2117 |
+
},
|
2118 |
+
{
|
2119 |
+
"epoch": 0.7988505747126436,
|
2120 |
+
"grad_norm": 2.862365245819092,
|
2121 |
+
"learning_rate": 2.3683058411940563e-06,
|
2122 |
+
"loss": 0.2967,
|
2123 |
+
"step": 278
|
2124 |
+
},
|
2125 |
+
{
|
2126 |
+
"epoch": 0.8017241379310345,
|
2127 |
+
"grad_norm": 4.078983783721924,
|
2128 |
+
"learning_rate": 2.3038321640050763e-06,
|
2129 |
+
"loss": 0.3005,
|
2130 |
+
"step": 279
|
2131 |
+
},
|
2132 |
+
{
|
2133 |
+
"epoch": 0.8045977011494253,
|
2134 |
+
"grad_norm": 4.147453308105469,
|
2135 |
+
"learning_rate": 2.2401338085430326e-06,
|
2136 |
+
"loss": 0.2901,
|
2137 |
+
"step": 280
|
2138 |
+
},
|
2139 |
+
{
|
2140 |
+
"epoch": 0.8045977011494253,
|
2141 |
+
"eval_accuracy": 0.8638613861386139,
|
2142 |
+
"eval_f1": 0.7417840375586855,
|
2143 |
+
"eval_loss": 0.3146108090877533,
|
2144 |
+
"eval_precision": 0.7383177570093458,
|
2145 |
+
"eval_recall": 0.7452830188679245,
|
2146 |
+
"eval_runtime": 16.7331,
|
2147 |
+
"eval_samples_per_second": 6.335,
|
2148 |
+
"eval_steps_per_second": 0.239,
|
2149 |
+
"step": 280
|
2150 |
+
},
|
2151 |
+
{
|
2152 |
+
"epoch": 0.8074712643678161,
|
2153 |
+
"grad_norm": 2.4641005992889404,
|
2154 |
+
"learning_rate": 2.177217191861183e-06,
|
2155 |
+
"loss": 0.2452,
|
2156 |
+
"step": 281
|
2157 |
+
},
|
2158 |
+
{
|
2159 |
+
"epoch": 0.8103448275862069,
|
2160 |
+
"grad_norm": 3.1481075286865234,
|
2161 |
+
"learning_rate": 2.115088652259446e-06,
|
2162 |
+
"loss": 0.3355,
|
2163 |
+
"step": 282
|
2164 |
+
},
|
2165 |
+
{
|
2166 |
+
"epoch": 0.8132183908045977,
|
2167 |
+
"grad_norm": 2.2011497020721436,
|
2168 |
+
"learning_rate": 2.053754448645846e-06,
|
2169 |
+
"loss": 0.2256,
|
2170 |
+
"step": 283
|
2171 |
+
},
|
2172 |
+
{
|
2173 |
+
"epoch": 0.8160919540229885,
|
2174 |
+
"grad_norm": 3.1297502517700195,
|
2175 |
+
"learning_rate": 1.9932207599059782e-06,
|
2176 |
+
"loss": 0.2899,
|
2177 |
+
"step": 284
|
2178 |
+
},
|
2179 |
+
{
|
2180 |
+
"epoch": 0.8189655172413793,
|
2181 |
+
"grad_norm": 2.566171646118164,
|
2182 |
+
"learning_rate": 1.933493684280574e-06,
|
2183 |
+
"loss": 0.2527,
|
2184 |
+
"step": 285
|
2185 |
+
},
|
2186 |
+
{
|
2187 |
+
"epoch": 0.8218390804597702,
|
2188 |
+
"grad_norm": 3.0499560832977295,
|
2189 |
+
"learning_rate": 1.8745792387511241e-06,
|
2190 |
+
"loss": 0.2979,
|
2191 |
+
"step": 286
|
2192 |
+
},
|
2193 |
+
{
|
2194 |
+
"epoch": 0.8247126436781609,
|
2195 |
+
"grad_norm": 3.5081562995910645,
|
2196 |
+
"learning_rate": 1.8164833584337216e-06,
|
2197 |
+
"loss": 0.2766,
|
2198 |
+
"step": 287
|
2199 |
+
},
|
2200 |
+
{
|
2201 |
+
"epoch": 0.8275862068965517,
|
2202 |
+
"grad_norm": 3.2664620876312256,
|
2203 |
+
"learning_rate": 1.75921189598118e-06,
|
2204 |
+
"loss": 0.3008,
|
2205 |
+
"step": 288
|
2206 |
+
},
|
2207 |
+
{
|
2208 |
+
"epoch": 0.8304597701149425,
|
2209 |
+
"grad_norm": 3.314521551132202,
|
2210 |
+
"learning_rate": 1.7027706209933903e-06,
|
2211 |
+
"loss": 0.3387,
|
2212 |
+
"step": 289
|
2213 |
+
},
|
2214 |
+
{
|
2215 |
+
"epoch": 0.8333333333333334,
|
2216 |
+
"grad_norm": 2.5149619579315186,
|
2217 |
+
"learning_rate": 1.6471652194361131e-06,
|
2218 |
+
"loss": 0.3032,
|
2219 |
+
"step": 290
|
2220 |
+
},
|
2221 |
+
{
|
2222 |
+
"epoch": 0.8362068965517241,
|
2223 |
+
"grad_norm": 3.847849130630493,
|
2224 |
+
"learning_rate": 1.5924012930681643e-06,
|
2225 |
+
"loss": 0.3218,
|
2226 |
+
"step": 291
|
2227 |
+
},
|
2228 |
+
{
|
2229 |
+
"epoch": 0.8390804597701149,
|
2230 |
+
"grad_norm": 6.620360374450684,
|
2231 |
+
"learning_rate": 1.5384843588770626e-06,
|
2232 |
+
"loss": 0.3464,
|
2233 |
+
"step": 292
|
2234 |
+
},
|
2235 |
+
{
|
2236 |
+
"epoch": 0.8419540229885057,
|
2237 |
+
"grad_norm": 3.233356475830078,
|
2238 |
+
"learning_rate": 1.4854198485232696e-06,
|
2239 |
+
"loss": 0.2475,
|
2240 |
+
"step": 293
|
2241 |
+
},
|
2242 |
+
{
|
2243 |
+
"epoch": 0.8448275862068966,
|
2244 |
+
"grad_norm": 4.268490314483643,
|
2245 |
+
"learning_rate": 1.433213107792991e-06,
|
2246 |
+
"loss": 0.3646,
|
2247 |
+
"step": 294
|
2248 |
+
},
|
2249 |
+
{
|
2250 |
+
"epoch": 0.8477011494252874,
|
2251 |
+
"grad_norm": 3.641005754470825,
|
2252 |
+
"learning_rate": 1.3818693960596186e-06,
|
2253 |
+
"loss": 0.3347,
|
2254 |
+
"step": 295
|
2255 |
+
},
|
2256 |
+
{
|
2257 |
+
"epoch": 0.8505747126436781,
|
2258 |
+
"grad_norm": 2.945902109146118,
|
2259 |
+
"learning_rate": 1.3313938857539133e-06,
|
2260 |
+
"loss": 0.2806,
|
2261 |
+
"step": 296
|
2262 |
+
},
|
2263 |
+
{
|
2264 |
+
"epoch": 0.853448275862069,
|
2265 |
+
"grad_norm": 3.552212715148926,
|
2266 |
+
"learning_rate": 1.2817916618429194e-06,
|
2267 |
+
"loss": 0.2993,
|
2268 |
+
"step": 297
|
2269 |
+
},
|
2270 |
+
{
|
2271 |
+
"epoch": 0.8563218390804598,
|
2272 |
+
"grad_norm": 3.9987523555755615,
|
2273 |
+
"learning_rate": 1.2330677213177034e-06,
|
2274 |
+
"loss": 0.2611,
|
2275 |
+
"step": 298
|
2276 |
+
},
|
2277 |
+
{
|
2278 |
+
"epoch": 0.8591954022988506,
|
2279 |
+
"grad_norm": 4.93873405456543,
|
2280 |
+
"learning_rate": 1.1852269726899423e-06,
|
2281 |
+
"loss": 0.3131,
|
2282 |
+
"step": 299
|
2283 |
+
},
|
2284 |
+
{
|
2285 |
+
"epoch": 0.8620689655172413,
|
2286 |
+
"grad_norm": 2.6833460330963135,
|
2287 |
+
"learning_rate": 1.138274235497443e-06,
|
2288 |
+
"loss": 0.2587,
|
2289 |
+
"step": 300
|
2290 |
+
},
|
2291 |
+
{
|
2292 |
+
"epoch": 0.8620689655172413,
|
2293 |
+
"eval_accuracy": 0.8638613861386139,
|
2294 |
+
"eval_f1": 0.7417840375586855,
|
2295 |
+
"eval_loss": 0.3137795925140381,
|
2296 |
+
"eval_precision": 0.7383177570093458,
|
2297 |
+
"eval_recall": 0.7452830188679245,
|
2298 |
+
"eval_runtime": 16.8912,
|
2299 |
+
"eval_samples_per_second": 6.275,
|
2300 |
+
"eval_steps_per_second": 0.237,
|
2301 |
+
"step": 300
|
2302 |
}
|
2303 |
],
|
2304 |
"logging_steps": 1,
|
|
|
2318 |
"attributes": {}
|
2319 |
}
|
2320 |
},
|
2321 |
+
"total_flos": 9.46781682579538e+16,
|
2322 |
"train_batch_size": 8,
|
2323 |
"trial_name": null,
|
2324 |
"trial_params": null
|