Training in progress, step 3000, checkpoint
Browse files- last-checkpoint/model.safetensors +1 -1
- last-checkpoint/optimizer.pt +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/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +766 -2
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 738367848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:719ef6a0a30163105c21960f0dd8b4a42fea79f576ca2432e13b51c25b82ea80
|
| 3 |
size 738367848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1476823354
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6e48bb60ca108c3441e9bb2125a1aa87430a3765f91e3afbcc5c959898ed887
|
| 3 |
size 1476823354
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57d1be83d8248a4d086961979df8c8adf273c0891e791d7b637d9e752cbaf971
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:df16bc8587b83b59d73ffcb4774bab640ed2bbf6249aba7b7112751df7280b58
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b8ebd66766c47747d9d34f4ee4e6f1e09fb1843f9769ec17242277c256d80133
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e44591b56351d86ebac6b6310a6b9a58bf9ebd5af691efd9614e457180a22080
|
| 3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2488c2baf1f7983e7e82c869c2ff023bdc7796ba97390c46686a4df8544a046
|
| 3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a864d68e543f00211ae2c48a5b9f47a92cf862dc03f0cda64f0647177108efe6
|
| 3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4ab980d3568f3d6a91c3cc4b09b1c84c8bbbd77347d21d918824619ddb9bc7f
|
| 3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a7008a4087300200a04419d46f39b98daf870297f179e965bf970ef908f90f3
|
| 3 |
size 15984
|
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:d6c54bb82c2097d343d35db0c4f1a616af19cd7b594e6327d1786008f1daedc5
|
| 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": 0.7747616492851076,
|
| 1536 |
"eval_steps_per_second": 8.47,
|
| 1537 |
"step": 2000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1538 |
}
|
| 1539 |
],
|
| 1540 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.8116213683223994,
|
| 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": 0.7747616492851076,
|
| 1536 |
"eval_steps_per_second": 8.47,
|
| 1537 |
"step": 2000
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.8837863167760074,
|
| 1541 |
+
"grad_norm": 1.116107702255249,
|
| 1542 |
+
"learning_rate": 9.985283134911919e-06,
|
| 1543 |
+
"loss": 0.1775,
|
| 1544 |
+
"step": 2010
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 1.8931583880037488,
|
| 1548 |
+
"grad_norm": 1.280927300453186,
|
| 1549 |
+
"learning_rate": 9.985209916677651e-06,
|
| 1550 |
+
"loss": 0.1853,
|
| 1551 |
+
"step": 2020
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 1.9025304592314902,
|
| 1555 |
+
"grad_norm": 1.419044852256775,
|
| 1556 |
+
"learning_rate": 9.98513669844338e-06,
|
| 1557 |
+
"loss": 0.1767,
|
| 1558 |
+
"step": 2030
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 1.9119025304592316,
|
| 1562 |
+
"grad_norm": 1.4140015840530396,
|
| 1563 |
+
"learning_rate": 9.985063480209111e-06,
|
| 1564 |
+
"loss": 0.1968,
|
| 1565 |
+
"step": 2040
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 1.9212746016869728,
|
| 1569 |
+
"grad_norm": 1.23015296459198,
|
| 1570 |
+
"learning_rate": 9.984990261974844e-06,
|
| 1571 |
+
"loss": 0.1559,
|
| 1572 |
+
"step": 2050
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 1.930646672914714,
|
| 1576 |
+
"grad_norm": 1.4209731817245483,
|
| 1577 |
+
"learning_rate": 9.984917043740574e-06,
|
| 1578 |
+
"loss": 0.18,
|
| 1579 |
+
"step": 2060
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.9400187441424555,
|
| 1583 |
+
"grad_norm": 1.5270899534225464,
|
| 1584 |
+
"learning_rate": 9.984843825506305e-06,
|
| 1585 |
+
"loss": 0.1858,
|
| 1586 |
+
"step": 2070
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 1.9493908153701969,
|
| 1590 |
+
"grad_norm": 2.0037920475006104,
|
| 1591 |
+
"learning_rate": 9.984770607272036e-06,
|
| 1592 |
+
"loss": 0.1812,
|
| 1593 |
+
"step": 2080
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 1.9587628865979383,
|
| 1597 |
+
"grad_norm": 1.4397103786468506,
|
| 1598 |
+
"learning_rate": 9.984697389037767e-06,
|
| 1599 |
+
"loss": 0.1853,
|
| 1600 |
+
"step": 2090
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 1.9681349578256795,
|
| 1604 |
+
"grad_norm": 1.555161476135254,
|
| 1605 |
+
"learning_rate": 9.984624170803497e-06,
|
| 1606 |
+
"loss": 0.1758,
|
| 1607 |
+
"step": 2100
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 1.9775070290534207,
|
| 1611 |
+
"grad_norm": 1.1453354358673096,
|
| 1612 |
+
"learning_rate": 9.984550952569228e-06,
|
| 1613 |
+
"loss": 0.1821,
|
| 1614 |
+
"step": 2110
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 1.986879100281162,
|
| 1618 |
+
"grad_norm": 1.3050484657287598,
|
| 1619 |
+
"learning_rate": 9.984477734334959e-06,
|
| 1620 |
+
"loss": 0.1828,
|
| 1621 |
+
"step": 2120
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 1.9962511715089035,
|
| 1625 |
+
"grad_norm": 1.1858463287353516,
|
| 1626 |
+
"learning_rate": 9.984404516100691e-06,
|
| 1627 |
+
"loss": 0.1801,
|
| 1628 |
+
"step": 2130
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 2.005623242736645,
|
| 1632 |
+
"grad_norm": 1.2467753887176514,
|
| 1633 |
+
"learning_rate": 9.984331297866422e-06,
|
| 1634 |
+
"loss": 0.1651,
|
| 1635 |
+
"step": 2140
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 2.014995313964386,
|
| 1639 |
+
"grad_norm": 1.9730074405670166,
|
| 1640 |
+
"learning_rate": 9.984258079632151e-06,
|
| 1641 |
+
"loss": 0.1654,
|
| 1642 |
+
"step": 2150
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 2.0243673851921273,
|
| 1646 |
+
"grad_norm": 1.384181261062622,
|
| 1647 |
+
"learning_rate": 9.984184861397884e-06,
|
| 1648 |
+
"loss": 0.151,
|
| 1649 |
+
"step": 2160
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 2.0337394564198688,
|
| 1653 |
+
"grad_norm": 1.2262136936187744,
|
| 1654 |
+
"learning_rate": 9.984111643163614e-06,
|
| 1655 |
+
"loss": 0.1338,
|
| 1656 |
+
"step": 2170
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 2.04311152764761,
|
| 1660 |
+
"grad_norm": 1.3417856693267822,
|
| 1661 |
+
"learning_rate": 9.984038424929345e-06,
|
| 1662 |
+
"loss": 0.1445,
|
| 1663 |
+
"step": 2180
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 2.0524835988753516,
|
| 1667 |
+
"grad_norm": 1.3032526969909668,
|
| 1668 |
+
"learning_rate": 9.983965206695076e-06,
|
| 1669 |
+
"loss": 0.1675,
|
| 1670 |
+
"step": 2190
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 2.0618556701030926,
|
| 1674 |
+
"grad_norm": 1.4586397409439087,
|
| 1675 |
+
"learning_rate": 9.983891988460808e-06,
|
| 1676 |
+
"loss": 0.1503,
|
| 1677 |
+
"step": 2200
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 2.071227741330834,
|
| 1681 |
+
"grad_norm": 1.8017582893371582,
|
| 1682 |
+
"learning_rate": 9.983818770226537e-06,
|
| 1683 |
+
"loss": 0.1614,
|
| 1684 |
+
"step": 2210
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 2.0805998125585754,
|
| 1688 |
+
"grad_norm": 1.1136542558670044,
|
| 1689 |
+
"learning_rate": 9.983745551992268e-06,
|
| 1690 |
+
"loss": 0.1385,
|
| 1691 |
+
"step": 2220
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 2.089971883786317,
|
| 1695 |
+
"grad_norm": 1.48130202293396,
|
| 1696 |
+
"learning_rate": 9.983672333758e-06,
|
| 1697 |
+
"loss": 0.1448,
|
| 1698 |
+
"step": 2230
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 2.0993439550140582,
|
| 1702 |
+
"grad_norm": 1.1847114562988281,
|
| 1703 |
+
"learning_rate": 9.983599115523731e-06,
|
| 1704 |
+
"loss": 0.1263,
|
| 1705 |
+
"step": 2240
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 2.108716026241799,
|
| 1709 |
+
"grad_norm": 1.068515419960022,
|
| 1710 |
+
"learning_rate": 9.983525897289462e-06,
|
| 1711 |
+
"loss": 0.1448,
|
| 1712 |
+
"step": 2250
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 2.108716026241799,
|
| 1716 |
+
"eval_loss": 0.0637284442782402,
|
| 1717 |
+
"eval_pearson_cosine": 0.766581654548645,
|
| 1718 |
+
"eval_pearson_dot": 0.652958333492279,
|
| 1719 |
+
"eval_pearson_euclidean": 0.76385897397995,
|
| 1720 |
+
"eval_pearson_manhattan": 0.7643536329269409,
|
| 1721 |
+
"eval_runtime": 24.9836,
|
| 1722 |
+
"eval_samples_per_second": 60.039,
|
| 1723 |
+
"eval_spearman_cosine": 0.7736502023043434,
|
| 1724 |
+
"eval_spearman_dot": 0.6506365364740643,
|
| 1725 |
+
"eval_spearman_euclidean": 0.7701725336122238,
|
| 1726 |
+
"eval_spearman_manhattan": 0.7705851416924343,
|
| 1727 |
+
"eval_steps_per_second": 7.525,
|
| 1728 |
+
"step": 2250
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 2.1180880974695406,
|
| 1732 |
+
"grad_norm": 1.2607600688934326,
|
| 1733 |
+
"learning_rate": 9.983452679055193e-06,
|
| 1734 |
+
"loss": 0.1405,
|
| 1735 |
+
"step": 2260
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"epoch": 2.127460168697282,
|
| 1739 |
+
"grad_norm": 1.3096617460250854,
|
| 1740 |
+
"learning_rate": 9.983379460820924e-06,
|
| 1741 |
+
"loss": 0.159,
|
| 1742 |
+
"step": 2270
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 2.1368322399250235,
|
| 1746 |
+
"grad_norm": 1.4220956563949585,
|
| 1747 |
+
"learning_rate": 9.983306242586654e-06,
|
| 1748 |
+
"loss": 0.1634,
|
| 1749 |
+
"step": 2280
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 2.146204311152765,
|
| 1753 |
+
"grad_norm": 1.5565595626831055,
|
| 1754 |
+
"learning_rate": 9.983233024352385e-06,
|
| 1755 |
+
"loss": 0.1549,
|
| 1756 |
+
"step": 2290
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 2.155576382380506,
|
| 1760 |
+
"grad_norm": 1.357906460762024,
|
| 1761 |
+
"learning_rate": 9.983159806118118e-06,
|
| 1762 |
+
"loss": 0.1503,
|
| 1763 |
+
"step": 2300
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 2.1649484536082473,
|
| 1767 |
+
"grad_norm": 1.0181514024734497,
|
| 1768 |
+
"learning_rate": 9.983086587883848e-06,
|
| 1769 |
+
"loss": 0.1242,
|
| 1770 |
+
"step": 2310
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 2.1743205248359887,
|
| 1774 |
+
"grad_norm": 1.2936785221099854,
|
| 1775 |
+
"learning_rate": 9.983013369649577e-06,
|
| 1776 |
+
"loss": 0.1516,
|
| 1777 |
+
"step": 2320
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 2.18369259606373,
|
| 1781 |
+
"grad_norm": 1.353125810623169,
|
| 1782 |
+
"learning_rate": 9.98294015141531e-06,
|
| 1783 |
+
"loss": 0.1576,
|
| 1784 |
+
"step": 2330
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 2.1930646672914715,
|
| 1788 |
+
"grad_norm": 1.5978926420211792,
|
| 1789 |
+
"learning_rate": 9.98286693318104e-06,
|
| 1790 |
+
"loss": 0.143,
|
| 1791 |
+
"step": 2340
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 2.2024367385192125,
|
| 1795 |
+
"grad_norm": 1.643609642982483,
|
| 1796 |
+
"learning_rate": 9.982793714946771e-06,
|
| 1797 |
+
"loss": 0.1509,
|
| 1798 |
+
"step": 2350
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 2.211808809746954,
|
| 1802 |
+
"grad_norm": 1.2868740558624268,
|
| 1803 |
+
"learning_rate": 9.982720496712502e-06,
|
| 1804 |
+
"loss": 0.1407,
|
| 1805 |
+
"step": 2360
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.2211808809746953,
|
| 1809 |
+
"grad_norm": 1.662234902381897,
|
| 1810 |
+
"learning_rate": 9.982647278478233e-06,
|
| 1811 |
+
"loss": 0.1499,
|
| 1812 |
+
"step": 2370
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 2.2305529522024368,
|
| 1816 |
+
"grad_norm": 1.7390748262405396,
|
| 1817 |
+
"learning_rate": 9.982574060243964e-06,
|
| 1818 |
+
"loss": 0.139,
|
| 1819 |
+
"step": 2380
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 2.239925023430178,
|
| 1823 |
+
"grad_norm": 1.2645044326782227,
|
| 1824 |
+
"learning_rate": 9.982500842009694e-06,
|
| 1825 |
+
"loss": 0.1541,
|
| 1826 |
+
"step": 2390
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 2.2492970946579196,
|
| 1830 |
+
"grad_norm": 1.5143808126449585,
|
| 1831 |
+
"learning_rate": 9.982427623775425e-06,
|
| 1832 |
+
"loss": 0.15,
|
| 1833 |
+
"step": 2400
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 2.2586691658856606,
|
| 1837 |
+
"grad_norm": 1.516233205795288,
|
| 1838 |
+
"learning_rate": 9.982354405541158e-06,
|
| 1839 |
+
"loss": 0.1387,
|
| 1840 |
+
"step": 2410
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 2.268041237113402,
|
| 1844 |
+
"grad_norm": 1.607926368713379,
|
| 1845 |
+
"learning_rate": 9.982281187306888e-06,
|
| 1846 |
+
"loss": 0.1459,
|
| 1847 |
+
"step": 2420
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.2774133083411434,
|
| 1851 |
+
"grad_norm": 1.433325171470642,
|
| 1852 |
+
"learning_rate": 9.982207969072617e-06,
|
| 1853 |
+
"loss": 0.145,
|
| 1854 |
+
"step": 2430
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 2.286785379568885,
|
| 1858 |
+
"grad_norm": 1.4051145315170288,
|
| 1859 |
+
"learning_rate": 9.98213475083835e-06,
|
| 1860 |
+
"loss": 0.1433,
|
| 1861 |
+
"step": 2440
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 2.296157450796626,
|
| 1865 |
+
"grad_norm": 1.5076231956481934,
|
| 1866 |
+
"learning_rate": 9.98206153260408e-06,
|
| 1867 |
+
"loss": 0.1514,
|
| 1868 |
+
"step": 2450
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 2.3055295220243672,
|
| 1872 |
+
"grad_norm": 1.185927152633667,
|
| 1873 |
+
"learning_rate": 9.981988314369811e-06,
|
| 1874 |
+
"loss": 0.1315,
|
| 1875 |
+
"step": 2460
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 2.3149015932521086,
|
| 1879 |
+
"grad_norm": 1.1687299013137817,
|
| 1880 |
+
"learning_rate": 9.981915096135542e-06,
|
| 1881 |
+
"loss": 0.1611,
|
| 1882 |
+
"step": 2470
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 2.32427366447985,
|
| 1886 |
+
"grad_norm": 1.205338716506958,
|
| 1887 |
+
"learning_rate": 9.981841877901274e-06,
|
| 1888 |
+
"loss": 0.1587,
|
| 1889 |
+
"step": 2480
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 2.3336457357075915,
|
| 1893 |
+
"grad_norm": 1.1079684495925903,
|
| 1894 |
+
"learning_rate": 9.981768659667004e-06,
|
| 1895 |
+
"loss": 0.142,
|
| 1896 |
+
"step": 2490
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 2.3430178069353325,
|
| 1900 |
+
"grad_norm": 1.1689645051956177,
|
| 1901 |
+
"learning_rate": 9.981695441432734e-06,
|
| 1902 |
+
"loss": 0.1449,
|
| 1903 |
+
"step": 2500
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 2.3430178069353325,
|
| 1907 |
+
"eval_loss": 0.05785529315471649,
|
| 1908 |
+
"eval_pearson_cosine": 0.7640599012374878,
|
| 1909 |
+
"eval_pearson_dot": 0.6659318208694458,
|
| 1910 |
+
"eval_pearson_euclidean": 0.7584241628646851,
|
| 1911 |
+
"eval_pearson_manhattan": 0.7589800357818604,
|
| 1912 |
+
"eval_runtime": 27.3942,
|
| 1913 |
+
"eval_samples_per_second": 54.756,
|
| 1914 |
+
"eval_spearman_cosine": 0.7698402659202235,
|
| 1915 |
+
"eval_spearman_dot": 0.6637382071207051,
|
| 1916 |
+
"eval_spearman_euclidean": 0.765183939076614,
|
| 1917 |
+
"eval_spearman_manhattan": 0.7654494135153407,
|
| 1918 |
+
"eval_steps_per_second": 6.863,
|
| 1919 |
+
"step": 2500
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 2.352389878163074,
|
| 1923 |
+
"grad_norm": 1.1410503387451172,
|
| 1924 |
+
"learning_rate": 9.981622223198467e-06,
|
| 1925 |
+
"loss": 0.1253,
|
| 1926 |
+
"step": 2510
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 2.3617619493908153,
|
| 1930 |
+
"grad_norm": 1.6562408208847046,
|
| 1931 |
+
"learning_rate": 9.981549004964197e-06,
|
| 1932 |
+
"loss": 0.1363,
|
| 1933 |
+
"step": 2520
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 2.3711340206185567,
|
| 1937 |
+
"grad_norm": 1.3503327369689941,
|
| 1938 |
+
"learning_rate": 9.981475786729928e-06,
|
| 1939 |
+
"loss": 0.141,
|
| 1940 |
+
"step": 2530
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 2.380506091846298,
|
| 1944 |
+
"grad_norm": 1.4653688669204712,
|
| 1945 |
+
"learning_rate": 9.981402568495659e-06,
|
| 1946 |
+
"loss": 0.1452,
|
| 1947 |
+
"step": 2540
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 2.3898781630740396,
|
| 1951 |
+
"grad_norm": 1.4135221242904663,
|
| 1952 |
+
"learning_rate": 9.98132935026139e-06,
|
| 1953 |
+
"loss": 0.1387,
|
| 1954 |
+
"step": 2550
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 2.3992502343017805,
|
| 1958 |
+
"grad_norm": 1.1758474111557007,
|
| 1959 |
+
"learning_rate": 9.98125613202712e-06,
|
| 1960 |
+
"loss": 0.1402,
|
| 1961 |
+
"step": 2560
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 2.408622305529522,
|
| 1965 |
+
"grad_norm": 1.6394227743148804,
|
| 1966 |
+
"learning_rate": 9.981182913792851e-06,
|
| 1967 |
+
"loss": 0.1434,
|
| 1968 |
+
"step": 2570
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 2.4179943767572634,
|
| 1972 |
+
"grad_norm": 1.5223402976989746,
|
| 1973 |
+
"learning_rate": 9.981109695558584e-06,
|
| 1974 |
+
"loss": 0.1433,
|
| 1975 |
+
"step": 2580
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 2.427366447985005,
|
| 1979 |
+
"grad_norm": 1.3722361326217651,
|
| 1980 |
+
"learning_rate": 9.981036477324314e-06,
|
| 1981 |
+
"loss": 0.145,
|
| 1982 |
+
"step": 2590
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 2.436738519212746,
|
| 1986 |
+
"grad_norm": 1.4288251399993896,
|
| 1987 |
+
"learning_rate": 9.980963259090045e-06,
|
| 1988 |
+
"loss": 0.1419,
|
| 1989 |
+
"step": 2600
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 2.446110590440487,
|
| 1993 |
+
"grad_norm": 1.3789891004562378,
|
| 1994 |
+
"learning_rate": 9.980890040855776e-06,
|
| 1995 |
+
"loss": 0.1428,
|
| 1996 |
+
"step": 2610
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 2.4554826616682286,
|
| 2000 |
+
"grad_norm": 1.3833218812942505,
|
| 2001 |
+
"learning_rate": 9.980816822621507e-06,
|
| 2002 |
+
"loss": 0.163,
|
| 2003 |
+
"step": 2620
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 2.46485473289597,
|
| 2007 |
+
"grad_norm": 1.2749391794204712,
|
| 2008 |
+
"learning_rate": 9.980743604387237e-06,
|
| 2009 |
+
"loss": 0.1457,
|
| 2010 |
+
"step": 2630
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 2.4742268041237114,
|
| 2014 |
+
"grad_norm": 1.3677037954330444,
|
| 2015 |
+
"learning_rate": 9.980670386152968e-06,
|
| 2016 |
+
"loss": 0.1393,
|
| 2017 |
+
"step": 2640
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 2.483598875351453,
|
| 2021 |
+
"grad_norm": 1.2386823892593384,
|
| 2022 |
+
"learning_rate": 9.980597167918699e-06,
|
| 2023 |
+
"loss": 0.1446,
|
| 2024 |
+
"step": 2650
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 2.492970946579194,
|
| 2028 |
+
"grad_norm": 1.6553146839141846,
|
| 2029 |
+
"learning_rate": 9.98052394968443e-06,
|
| 2030 |
+
"loss": 0.1399,
|
| 2031 |
+
"step": 2660
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 2.5023430178069352,
|
| 2035 |
+
"grad_norm": 1.2258574962615967,
|
| 2036 |
+
"learning_rate": 9.98045073145016e-06,
|
| 2037 |
+
"loss": 0.1557,
|
| 2038 |
+
"step": 2670
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 2.5117150890346767,
|
| 2042 |
+
"grad_norm": 1.1680238246917725,
|
| 2043 |
+
"learning_rate": 9.980377513215891e-06,
|
| 2044 |
+
"loss": 0.14,
|
| 2045 |
+
"step": 2680
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 2.521087160262418,
|
| 2049 |
+
"grad_norm": 1.3764533996582031,
|
| 2050 |
+
"learning_rate": 9.980304294981624e-06,
|
| 2051 |
+
"loss": 0.1429,
|
| 2052 |
+
"step": 2690
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 2.530459231490159,
|
| 2056 |
+
"grad_norm": 1.1607757806777954,
|
| 2057 |
+
"learning_rate": 9.980231076747354e-06,
|
| 2058 |
+
"loss": 0.156,
|
| 2059 |
+
"step": 2700
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 2.539831302717901,
|
| 2063 |
+
"grad_norm": 1.30258309841156,
|
| 2064 |
+
"learning_rate": 9.980157858513085e-06,
|
| 2065 |
+
"loss": 0.1334,
|
| 2066 |
+
"step": 2710
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 2.549203373945642,
|
| 2070 |
+
"grad_norm": 1.3965803384780884,
|
| 2071 |
+
"learning_rate": 9.980084640278816e-06,
|
| 2072 |
+
"loss": 0.1532,
|
| 2073 |
+
"step": 2720
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 2.5585754451733833,
|
| 2077 |
+
"grad_norm": 1.2492479085922241,
|
| 2078 |
+
"learning_rate": 9.980011422044547e-06,
|
| 2079 |
+
"loss": 0.1538,
|
| 2080 |
+
"step": 2730
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 2.5679475164011247,
|
| 2084 |
+
"grad_norm": 1.5879229307174683,
|
| 2085 |
+
"learning_rate": 9.979938203810277e-06,
|
| 2086 |
+
"loss": 0.1393,
|
| 2087 |
+
"step": 2740
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 2.5773195876288657,
|
| 2091 |
+
"grad_norm": 1.5499955415725708,
|
| 2092 |
+
"learning_rate": 9.979864985576008e-06,
|
| 2093 |
+
"loss": 0.1443,
|
| 2094 |
+
"step": 2750
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 2.5773195876288657,
|
| 2098 |
+
"eval_loss": 0.059572458267211914,
|
| 2099 |
+
"eval_pearson_cosine": 0.7583234310150146,
|
| 2100 |
+
"eval_pearson_dot": 0.6585268378257751,
|
| 2101 |
+
"eval_pearson_euclidean": 0.7594324946403503,
|
| 2102 |
+
"eval_pearson_manhattan": 0.7599164843559265,
|
| 2103 |
+
"eval_runtime": 25.1198,
|
| 2104 |
+
"eval_samples_per_second": 59.714,
|
| 2105 |
+
"eval_spearman_cosine": 0.7658877891929784,
|
| 2106 |
+
"eval_spearman_dot": 0.6550703356470525,
|
| 2107 |
+
"eval_spearman_euclidean": 0.7651954936870381,
|
| 2108 |
+
"eval_spearman_manhattan": 0.7656066832066194,
|
| 2109 |
+
"eval_steps_per_second": 7.484,
|
| 2110 |
+
"step": 2750
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 2.5866916588566076,
|
| 2114 |
+
"grad_norm": 1.1182575225830078,
|
| 2115 |
+
"learning_rate": 9.97979176734174e-06,
|
| 2116 |
+
"loss": 0.1449,
|
| 2117 |
+
"step": 2760
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.5960637300843485,
|
| 2121 |
+
"grad_norm": 1.3228731155395508,
|
| 2122 |
+
"learning_rate": 9.979718549107471e-06,
|
| 2123 |
+
"loss": 0.1339,
|
| 2124 |
+
"step": 2770
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 2.60543580131209,
|
| 2128 |
+
"grad_norm": 1.3763021230697632,
|
| 2129 |
+
"learning_rate": 9.9796453308732e-06,
|
| 2130 |
+
"loss": 0.1379,
|
| 2131 |
+
"step": 2780
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 2.6148078725398314,
|
| 2135 |
+
"grad_norm": 1.6708637475967407,
|
| 2136 |
+
"learning_rate": 9.979572112638933e-06,
|
| 2137 |
+
"loss": 0.1491,
|
| 2138 |
+
"step": 2790
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.624179943767573,
|
| 2142 |
+
"grad_norm": 1.0826717615127563,
|
| 2143 |
+
"learning_rate": 9.979498894404664e-06,
|
| 2144 |
+
"loss": 0.1447,
|
| 2145 |
+
"step": 2800
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 2.633552014995314,
|
| 2149 |
+
"grad_norm": 1.4416155815124512,
|
| 2150 |
+
"learning_rate": 9.979425676170394e-06,
|
| 2151 |
+
"loss": 0.1398,
|
| 2152 |
+
"step": 2810
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 2.642924086223055,
|
| 2156 |
+
"grad_norm": 1.3966304063796997,
|
| 2157 |
+
"learning_rate": 9.979352457936125e-06,
|
| 2158 |
+
"loss": 0.1332,
|
| 2159 |
+
"step": 2820
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.6522961574507966,
|
| 2163 |
+
"grad_norm": 1.5255811214447021,
|
| 2164 |
+
"learning_rate": 9.979279239701856e-06,
|
| 2165 |
+
"loss": 0.1423,
|
| 2166 |
+
"step": 2830
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 2.661668228678538,
|
| 2170 |
+
"grad_norm": 1.3866652250289917,
|
| 2171 |
+
"learning_rate": 9.979206021467587e-06,
|
| 2172 |
+
"loss": 0.1554,
|
| 2173 |
+
"step": 2840
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 2.6710402999062794,
|
| 2177 |
+
"grad_norm": 1.3477802276611328,
|
| 2178 |
+
"learning_rate": 9.979132803233317e-06,
|
| 2179 |
+
"loss": 0.1547,
|
| 2180 |
+
"step": 2850
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.680412371134021,
|
| 2184 |
+
"grad_norm": 1.540963053703308,
|
| 2185 |
+
"learning_rate": 9.97905958499905e-06,
|
| 2186 |
+
"loss": 0.1229,
|
| 2187 |
+
"step": 2860
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.689784442361762,
|
| 2191 |
+
"grad_norm": 1.697350025177002,
|
| 2192 |
+
"learning_rate": 9.97898636676478e-06,
|
| 2193 |
+
"loss": 0.153,
|
| 2194 |
+
"step": 2870
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 2.6991565135895033,
|
| 2198 |
+
"grad_norm": 1.6020257472991943,
|
| 2199 |
+
"learning_rate": 9.978913148530511e-06,
|
| 2200 |
+
"loss": 0.1334,
|
| 2201 |
+
"step": 2880
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.7085285848172447,
|
| 2205 |
+
"grad_norm": 1.7637958526611328,
|
| 2206 |
+
"learning_rate": 9.978839930296242e-06,
|
| 2207 |
+
"loss": 0.1513,
|
| 2208 |
+
"step": 2890
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 2.717900656044986,
|
| 2212 |
+
"grad_norm": 1.2917182445526123,
|
| 2213 |
+
"learning_rate": 9.978766712061973e-06,
|
| 2214 |
+
"loss": 0.1296,
|
| 2215 |
+
"step": 2900
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 2.7272727272727275,
|
| 2219 |
+
"grad_norm": 1.42876136302948,
|
| 2220 |
+
"learning_rate": 9.978693493827704e-06,
|
| 2221 |
+
"loss": 0.1276,
|
| 2222 |
+
"step": 2910
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 2.7366447985004685,
|
| 2226 |
+
"grad_norm": 1.340184211730957,
|
| 2227 |
+
"learning_rate": 9.978620275593434e-06,
|
| 2228 |
+
"loss": 0.164,
|
| 2229 |
+
"step": 2920
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 2.74601686972821,
|
| 2233 |
+
"grad_norm": 1.1638396978378296,
|
| 2234 |
+
"learning_rate": 9.978547057359165e-06,
|
| 2235 |
+
"loss": 0.1372,
|
| 2236 |
+
"step": 2930
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 2.7553889409559513,
|
| 2240 |
+
"grad_norm": 1.5060447454452515,
|
| 2241 |
+
"learning_rate": 9.978473839124897e-06,
|
| 2242 |
+
"loss": 0.1489,
|
| 2243 |
+
"step": 2940
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.7647610121836927,
|
| 2247 |
+
"grad_norm": 1.3632638454437256,
|
| 2248 |
+
"learning_rate": 9.978400620890627e-06,
|
| 2249 |
+
"loss": 0.1242,
|
| 2250 |
+
"step": 2950
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 2.774133083411434,
|
| 2254 |
+
"grad_norm": 1.6402980089187622,
|
| 2255 |
+
"learning_rate": 9.978327402656359e-06,
|
| 2256 |
+
"loss": 0.1395,
|
| 2257 |
+
"step": 2960
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 2.783505154639175,
|
| 2261 |
+
"grad_norm": 1.8350452184677124,
|
| 2262 |
+
"learning_rate": 9.97825418442209e-06,
|
| 2263 |
+
"loss": 0.1501,
|
| 2264 |
+
"step": 2970
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 2.7928772258669166,
|
| 2268 |
+
"grad_norm": 1.6517874002456665,
|
| 2269 |
+
"learning_rate": 9.97818096618782e-06,
|
| 2270 |
+
"loss": 0.1596,
|
| 2271 |
+
"step": 2980
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 2.802249297094658,
|
| 2275 |
+
"grad_norm": 1.7441259622573853,
|
| 2276 |
+
"learning_rate": 9.978107747953551e-06,
|
| 2277 |
+
"loss": 0.1344,
|
| 2278 |
+
"step": 2990
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 2.8116213683223994,
|
| 2282 |
+
"grad_norm": 1.4474517107009888,
|
| 2283 |
+
"learning_rate": 9.978034529719282e-06,
|
| 2284 |
+
"loss": 0.1363,
|
| 2285 |
+
"step": 3000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.8116213683223994,
|
| 2289 |
+
"eval_loss": 0.05750729516148567,
|
| 2290 |
+
"eval_pearson_cosine": 0.767126202583313,
|
| 2291 |
+
"eval_pearson_dot": 0.676889181137085,
|
| 2292 |
+
"eval_pearson_euclidean": 0.756407618522644,
|
| 2293 |
+
"eval_pearson_manhattan": 0.7570176124572754,
|
| 2294 |
+
"eval_runtime": 25.3699,
|
| 2295 |
+
"eval_samples_per_second": 59.125,
|
| 2296 |
+
"eval_spearman_cosine": 0.7727339030438767,
|
| 2297 |
+
"eval_spearman_dot": 0.6755843192398268,
|
| 2298 |
+
"eval_spearman_euclidean": 0.7624238185076594,
|
| 2299 |
+
"eval_spearman_manhattan": 0.7629469399526556,
|
| 2300 |
+
"eval_steps_per_second": 7.41,
|
| 2301 |
+
"step": 3000
|
| 2302 |
}
|
| 2303 |
],
|
| 2304 |
"logging_steps": 10,
|