Training in progress, step 5000, 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 +1530 -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 sha256:77bdd07402c0fe434c587ece44b2edeb5f86258e2a03ca9d156a6d48b5150f65
|
| 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 sha256:2cfa7170844cfd8c9e538ce6b712c134638973e02bcee820b7bd2a686a44027c
|
| 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 sha256:5504b8d722b425f58bab6aedf9a43fc8129b02036307d31c7a21e224d2412ace
|
| 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:e555860fd7a2cfb8945f188f7232baf938ce622886881cc422b3eb0e7444eda4
|
| 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:ea0c272cf77c9504efaa077bfa8f9229d461c16d6641be0e57a7f20f9b761399
|
| 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 sha256:921e0812b510be6ea788fc2c6aa7541f3ff4eb1bb3dd7c230340a35d8e1e764b
|
| 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:c3603a5403f0489f4be4d27720a0fa7e0fe0d08dbde5d58c1060cef37b9084d2
|
| 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 sha256:2b9f685b83b5545cc2db9c29e88184590e89acb7836b4bb92a6a1df01b4bf43f
|
| 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 sha256:5843202dd0ab5bc42fd0b6fa35e7cc2dca365d38fb379a2faf93bf274ef023e6
|
| 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 sha256:79f72ffc2afb7672fc32ddd050c69181c2c0c16f8eac79a352eecb064fb5a9c7
|
| 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 sha256:10f98a9bf3c827b7f13510b7bc00db936ca6b0dcd935745bc447f03aae03112f
|
| 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,
|
|
@@ -2299,6 +2299,1534 @@
|
|
| 2299 |
"eval_spearman_manhattan": 0.7629469399526556,
|
| 2300 |
"eval_steps_per_second": 7.41,
|
| 2301 |
"step": 3000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2302 |
}
|
| 2303 |
],
|
| 2304 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.686035613870665,
|
| 5 |
"eval_steps": 250,
|
| 6 |
+
"global_step": 5000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2299 |
"eval_spearman_manhattan": 0.7629469399526556,
|
| 2300 |
"eval_steps_per_second": 7.41,
|
| 2301 |
"step": 3000
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 2.820993439550141,
|
| 2305 |
+
"grad_norm": 1.4202260971069336,
|
| 2306 |
+
"learning_rate": 9.977961311485013e-06,
|
| 2307 |
+
"loss": 0.1456,
|
| 2308 |
+
"step": 3010
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 2.830365510777882,
|
| 2312 |
+
"grad_norm": 1.3678419589996338,
|
| 2313 |
+
"learning_rate": 9.977888093250743e-06,
|
| 2314 |
+
"loss": 0.1445,
|
| 2315 |
+
"step": 3020
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.839737582005623,
|
| 2319 |
+
"grad_norm": 1.168271541595459,
|
| 2320 |
+
"learning_rate": 9.977814875016474e-06,
|
| 2321 |
+
"loss": 0.1428,
|
| 2322 |
+
"step": 3030
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 2.8491096532333646,
|
| 2326 |
+
"grad_norm": 1.5929275751113892,
|
| 2327 |
+
"learning_rate": 9.977741656782207e-06,
|
| 2328 |
+
"loss": 0.1593,
|
| 2329 |
+
"step": 3040
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 2.858481724461106,
|
| 2333 |
+
"grad_norm": 1.265101432800293,
|
| 2334 |
+
"learning_rate": 9.977668438547937e-06,
|
| 2335 |
+
"loss": 0.1519,
|
| 2336 |
+
"step": 3050
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 2.8678537956888475,
|
| 2340 |
+
"grad_norm": 1.1187818050384521,
|
| 2341 |
+
"learning_rate": 9.977595220313666e-06,
|
| 2342 |
+
"loss": 0.1454,
|
| 2343 |
+
"step": 3060
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 2.8772258669165884,
|
| 2347 |
+
"grad_norm": 1.1976639032363892,
|
| 2348 |
+
"learning_rate": 9.977522002079399e-06,
|
| 2349 |
+
"loss": 0.1321,
|
| 2350 |
+
"step": 3070
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 2.88659793814433,
|
| 2354 |
+
"grad_norm": 1.7162209749221802,
|
| 2355 |
+
"learning_rate": 9.97744878384513e-06,
|
| 2356 |
+
"loss": 0.147,
|
| 2357 |
+
"step": 3080
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.8959700093720713,
|
| 2361 |
+
"grad_norm": 1.3301661014556885,
|
| 2362 |
+
"learning_rate": 9.97737556561086e-06,
|
| 2363 |
+
"loss": 0.1341,
|
| 2364 |
+
"step": 3090
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 2.9053420805998127,
|
| 2368 |
+
"grad_norm": 1.279984951019287,
|
| 2369 |
+
"learning_rate": 9.977302347376591e-06,
|
| 2370 |
+
"loss": 0.1342,
|
| 2371 |
+
"step": 3100
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 2.914714151827554,
|
| 2375 |
+
"grad_norm": 1.6548879146575928,
|
| 2376 |
+
"learning_rate": 9.977229129142324e-06,
|
| 2377 |
+
"loss": 0.1429,
|
| 2378 |
+
"step": 3110
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 2.924086223055295,
|
| 2382 |
+
"grad_norm": 0.9662721753120422,
|
| 2383 |
+
"learning_rate": 9.977155910908053e-06,
|
| 2384 |
+
"loss": 0.1524,
|
| 2385 |
+
"step": 3120
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 2.9334582942830365,
|
| 2389 |
+
"grad_norm": 1.5336380004882812,
|
| 2390 |
+
"learning_rate": 9.977082692673783e-06,
|
| 2391 |
+
"loss": 0.1445,
|
| 2392 |
+
"step": 3130
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 2.942830365510778,
|
| 2396 |
+
"grad_norm": 1.4380927085876465,
|
| 2397 |
+
"learning_rate": 9.977009474439516e-06,
|
| 2398 |
+
"loss": 0.1371,
|
| 2399 |
+
"step": 3140
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.9522024367385193,
|
| 2403 |
+
"grad_norm": 1.551700472831726,
|
| 2404 |
+
"learning_rate": 9.976936256205247e-06,
|
| 2405 |
+
"loss": 0.135,
|
| 2406 |
+
"step": 3150
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 2.9615745079662608,
|
| 2410 |
+
"grad_norm": 1.32683265209198,
|
| 2411 |
+
"learning_rate": 9.976863037970977e-06,
|
| 2412 |
+
"loss": 0.1444,
|
| 2413 |
+
"step": 3160
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 2.9709465791940017,
|
| 2417 |
+
"grad_norm": 1.3574503660202026,
|
| 2418 |
+
"learning_rate": 9.976789819736708e-06,
|
| 2419 |
+
"loss": 0.1391,
|
| 2420 |
+
"step": 3170
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 2.980318650421743,
|
| 2424 |
+
"grad_norm": 1.506625771522522,
|
| 2425 |
+
"learning_rate": 9.976716601502439e-06,
|
| 2426 |
+
"loss": 0.1552,
|
| 2427 |
+
"step": 3180
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 2.9896907216494846,
|
| 2431 |
+
"grad_norm": 1.3970105648040771,
|
| 2432 |
+
"learning_rate": 9.97664338326817e-06,
|
| 2433 |
+
"loss": 0.147,
|
| 2434 |
+
"step": 3190
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 2.999062792877226,
|
| 2438 |
+
"grad_norm": 1.4303011894226074,
|
| 2439 |
+
"learning_rate": 9.9765701650339e-06,
|
| 2440 |
+
"loss": 0.1559,
|
| 2441 |
+
"step": 3200
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 3.0084348641049674,
|
| 2445 |
+
"grad_norm": 1.377488613128662,
|
| 2446 |
+
"learning_rate": 9.976496946799633e-06,
|
| 2447 |
+
"loss": 0.1187,
|
| 2448 |
+
"step": 3210
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 3.0178069353327084,
|
| 2452 |
+
"grad_norm": 1.1664360761642456,
|
| 2453 |
+
"learning_rate": 9.976423728565364e-06,
|
| 2454 |
+
"loss": 0.1101,
|
| 2455 |
+
"step": 3220
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 3.02717900656045,
|
| 2459 |
+
"grad_norm": 0.9129014015197754,
|
| 2460 |
+
"learning_rate": 9.976350510331093e-06,
|
| 2461 |
+
"loss": 0.111,
|
| 2462 |
+
"step": 3230
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 3.036551077788191,
|
| 2466 |
+
"grad_norm": 1.2628843784332275,
|
| 2467 |
+
"learning_rate": 9.976277292096825e-06,
|
| 2468 |
+
"loss": 0.1141,
|
| 2469 |
+
"step": 3240
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 3.0459231490159326,
|
| 2473 |
+
"grad_norm": 1.1534360647201538,
|
| 2474 |
+
"learning_rate": 9.976204073862556e-06,
|
| 2475 |
+
"loss": 0.1227,
|
| 2476 |
+
"step": 3250
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 3.0459231490159326,
|
| 2480 |
+
"eval_loss": 0.051736850291490555,
|
| 2481 |
+
"eval_pearson_cosine": 0.763727605342865,
|
| 2482 |
+
"eval_pearson_dot": 0.673626720905304,
|
| 2483 |
+
"eval_pearson_euclidean": 0.756030797958374,
|
| 2484 |
+
"eval_pearson_manhattan": 0.7567305564880371,
|
| 2485 |
+
"eval_runtime": 21.997,
|
| 2486 |
+
"eval_samples_per_second": 68.191,
|
| 2487 |
+
"eval_spearman_cosine": 0.7669834916269708,
|
| 2488 |
+
"eval_spearman_dot": 0.6714383880600381,
|
| 2489 |
+
"eval_spearman_euclidean": 0.7611960037220876,
|
| 2490 |
+
"eval_spearman_manhattan": 0.7615680957541558,
|
| 2491 |
+
"eval_steps_per_second": 8.547,
|
| 2492 |
+
"step": 3250
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 3.055295220243674,
|
| 2496 |
+
"grad_norm": 1.4779927730560303,
|
| 2497 |
+
"learning_rate": 9.976130855628287e-06,
|
| 2498 |
+
"loss": 0.1186,
|
| 2499 |
+
"step": 3260
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 3.064667291471415,
|
| 2503 |
+
"grad_norm": 1.2425293922424316,
|
| 2504 |
+
"learning_rate": 9.976057637394017e-06,
|
| 2505 |
+
"loss": 0.1213,
|
| 2506 |
+
"step": 3270
|
| 2507 |
+
},
|
| 2508 |
+
{
|
| 2509 |
+
"epoch": 3.0740393626991565,
|
| 2510 |
+
"grad_norm": 1.6161679029464722,
|
| 2511 |
+
"learning_rate": 9.975984419159748e-06,
|
| 2512 |
+
"loss": 0.1127,
|
| 2513 |
+
"step": 3280
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 3.083411433926898,
|
| 2517 |
+
"grad_norm": 1.199263334274292,
|
| 2518 |
+
"learning_rate": 9.975911200925479e-06,
|
| 2519 |
+
"loss": 0.0971,
|
| 2520 |
+
"step": 3290
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 3.0927835051546393,
|
| 2524 |
+
"grad_norm": 1.5749520063400269,
|
| 2525 |
+
"learning_rate": 9.97583798269121e-06,
|
| 2526 |
+
"loss": 0.1162,
|
| 2527 |
+
"step": 3300
|
| 2528 |
+
},
|
| 2529 |
+
{
|
| 2530 |
+
"epoch": 3.1021555763823807,
|
| 2531 |
+
"grad_norm": 1.558112382888794,
|
| 2532 |
+
"learning_rate": 9.97576476445694e-06,
|
| 2533 |
+
"loss": 0.125,
|
| 2534 |
+
"step": 3310
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 3.1115276476101217,
|
| 2538 |
+
"grad_norm": 1.5197752714157104,
|
| 2539 |
+
"learning_rate": 9.975691546222673e-06,
|
| 2540 |
+
"loss": 0.1199,
|
| 2541 |
+
"step": 3320
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 3.120899718837863,
|
| 2545 |
+
"grad_norm": 1.1978933811187744,
|
| 2546 |
+
"learning_rate": 9.975618327988404e-06,
|
| 2547 |
+
"loss": 0.0975,
|
| 2548 |
+
"step": 3330
|
| 2549 |
+
},
|
| 2550 |
+
{
|
| 2551 |
+
"epoch": 3.1302717900656045,
|
| 2552 |
+
"grad_norm": 1.0790154933929443,
|
| 2553 |
+
"learning_rate": 9.975545109754134e-06,
|
| 2554 |
+
"loss": 0.1078,
|
| 2555 |
+
"step": 3340
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 3.139643861293346,
|
| 2559 |
+
"grad_norm": 1.7810611724853516,
|
| 2560 |
+
"learning_rate": 9.975471891519865e-06,
|
| 2561 |
+
"loss": 0.1065,
|
| 2562 |
+
"step": 3350
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 3.1490159325210874,
|
| 2566 |
+
"grad_norm": 1.2899665832519531,
|
| 2567 |
+
"learning_rate": 9.975398673285596e-06,
|
| 2568 |
+
"loss": 0.1104,
|
| 2569 |
+
"step": 3360
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 3.1583880037488283,
|
| 2573 |
+
"grad_norm": 1.1923859119415283,
|
| 2574 |
+
"learning_rate": 9.975325455051327e-06,
|
| 2575 |
+
"loss": 0.1143,
|
| 2576 |
+
"step": 3370
|
| 2577 |
+
},
|
| 2578 |
+
{
|
| 2579 |
+
"epoch": 3.1677600749765698,
|
| 2580 |
+
"grad_norm": 1.428306221961975,
|
| 2581 |
+
"learning_rate": 9.975252236817057e-06,
|
| 2582 |
+
"loss": 0.101,
|
| 2583 |
+
"step": 3380
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 3.177132146204311,
|
| 2587 |
+
"grad_norm": 1.323941946029663,
|
| 2588 |
+
"learning_rate": 9.97517901858279e-06,
|
| 2589 |
+
"loss": 0.1115,
|
| 2590 |
+
"step": 3390
|
| 2591 |
+
},
|
| 2592 |
+
{
|
| 2593 |
+
"epoch": 3.1865042174320526,
|
| 2594 |
+
"grad_norm": 1.4079722166061401,
|
| 2595 |
+
"learning_rate": 9.97510580034852e-06,
|
| 2596 |
+
"loss": 0.1032,
|
| 2597 |
+
"step": 3400
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 3.195876288659794,
|
| 2601 |
+
"grad_norm": 1.2919671535491943,
|
| 2602 |
+
"learning_rate": 9.97503258211425e-06,
|
| 2603 |
+
"loss": 0.1145,
|
| 2604 |
+
"step": 3410
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 3.205248359887535,
|
| 2608 |
+
"grad_norm": 1.1800559759140015,
|
| 2609 |
+
"learning_rate": 9.974959363879982e-06,
|
| 2610 |
+
"loss": 0.106,
|
| 2611 |
+
"step": 3420
|
| 2612 |
+
},
|
| 2613 |
+
{
|
| 2614 |
+
"epoch": 3.2146204311152764,
|
| 2615 |
+
"grad_norm": 1.5425052642822266,
|
| 2616 |
+
"learning_rate": 9.974886145645713e-06,
|
| 2617 |
+
"loss": 0.1156,
|
| 2618 |
+
"step": 3430
|
| 2619 |
+
},
|
| 2620 |
+
{
|
| 2621 |
+
"epoch": 3.223992502343018,
|
| 2622 |
+
"grad_norm": 1.7271355390548706,
|
| 2623 |
+
"learning_rate": 9.974812927411443e-06,
|
| 2624 |
+
"loss": 0.1121,
|
| 2625 |
+
"step": 3440
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 3.2333645735707592,
|
| 2629 |
+
"grad_norm": 1.3295711278915405,
|
| 2630 |
+
"learning_rate": 9.974739709177174e-06,
|
| 2631 |
+
"loss": 0.1072,
|
| 2632 |
+
"step": 3450
|
| 2633 |
+
},
|
| 2634 |
+
{
|
| 2635 |
+
"epoch": 3.2427366447985007,
|
| 2636 |
+
"grad_norm": 1.658498764038086,
|
| 2637 |
+
"learning_rate": 9.974666490942905e-06,
|
| 2638 |
+
"loss": 0.1131,
|
| 2639 |
+
"step": 3460
|
| 2640 |
+
},
|
| 2641 |
+
{
|
| 2642 |
+
"epoch": 3.2521087160262416,
|
| 2643 |
+
"grad_norm": 1.6077649593353271,
|
| 2644 |
+
"learning_rate": 9.974593272708636e-06,
|
| 2645 |
+
"loss": 0.1143,
|
| 2646 |
+
"step": 3470
|
| 2647 |
+
},
|
| 2648 |
+
{
|
| 2649 |
+
"epoch": 3.261480787253983,
|
| 2650 |
+
"grad_norm": 1.4552775621414185,
|
| 2651 |
+
"learning_rate": 9.974520054474366e-06,
|
| 2652 |
+
"loss": 0.1065,
|
| 2653 |
+
"step": 3480
|
| 2654 |
+
},
|
| 2655 |
+
{
|
| 2656 |
+
"epoch": 3.2708528584817245,
|
| 2657 |
+
"grad_norm": 1.586267113685608,
|
| 2658 |
+
"learning_rate": 9.974446836240099e-06,
|
| 2659 |
+
"loss": 0.1137,
|
| 2660 |
+
"step": 3490
|
| 2661 |
+
},
|
| 2662 |
+
{
|
| 2663 |
+
"epoch": 3.280224929709466,
|
| 2664 |
+
"grad_norm": 0.9890511631965637,
|
| 2665 |
+
"learning_rate": 9.97437361800583e-06,
|
| 2666 |
+
"loss": 0.103,
|
| 2667 |
+
"step": 3500
|
| 2668 |
+
},
|
| 2669 |
+
{
|
| 2670 |
+
"epoch": 3.280224929709466,
|
| 2671 |
+
"eval_loss": 0.04644956439733505,
|
| 2672 |
+
"eval_pearson_cosine": 0.760254442691803,
|
| 2673 |
+
"eval_pearson_dot": 0.6812557578086853,
|
| 2674 |
+
"eval_pearson_euclidean": 0.7475454807281494,
|
| 2675 |
+
"eval_pearson_manhattan": 0.7483712434768677,
|
| 2676 |
+
"eval_runtime": 22.2407,
|
| 2677 |
+
"eval_samples_per_second": 67.444,
|
| 2678 |
+
"eval_spearman_cosine": 0.7642516190492565,
|
| 2679 |
+
"eval_spearman_dot": 0.6795590047108491,
|
| 2680 |
+
"eval_spearman_euclidean": 0.7527436591109528,
|
| 2681 |
+
"eval_spearman_manhattan": 0.7534967017417152,
|
| 2682 |
+
"eval_steps_per_second": 8.453,
|
| 2683 |
+
"step": 3500
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 3.2895970009372073,
|
| 2687 |
+
"grad_norm": 1.4361557960510254,
|
| 2688 |
+
"learning_rate": 9.97430039977156e-06,
|
| 2689 |
+
"loss": 0.1078,
|
| 2690 |
+
"step": 3510
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 3.2989690721649483,
|
| 2694 |
+
"grad_norm": 1.307634949684143,
|
| 2695 |
+
"learning_rate": 9.974227181537291e-06,
|
| 2696 |
+
"loss": 0.105,
|
| 2697 |
+
"step": 3520
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 3.3083411433926897,
|
| 2701 |
+
"grad_norm": 1.103812336921692,
|
| 2702 |
+
"learning_rate": 9.974153963303022e-06,
|
| 2703 |
+
"loss": 0.1021,
|
| 2704 |
+
"step": 3530
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 3.317713214620431,
|
| 2708 |
+
"grad_norm": 1.485766887664795,
|
| 2709 |
+
"learning_rate": 9.974080745068753e-06,
|
| 2710 |
+
"loss": 0.1055,
|
| 2711 |
+
"step": 3540
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 3.3270852858481725,
|
| 2715 |
+
"grad_norm": 1.4017934799194336,
|
| 2716 |
+
"learning_rate": 9.974007526834483e-06,
|
| 2717 |
+
"loss": 0.0991,
|
| 2718 |
+
"step": 3550
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 3.336457357075914,
|
| 2722 |
+
"grad_norm": 1.1994048357009888,
|
| 2723 |
+
"learning_rate": 9.973934308600214e-06,
|
| 2724 |
+
"loss": 0.1176,
|
| 2725 |
+
"step": 3560
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 3.345829428303655,
|
| 2729 |
+
"grad_norm": 1.0661845207214355,
|
| 2730 |
+
"learning_rate": 9.973861090365947e-06,
|
| 2731 |
+
"loss": 0.1036,
|
| 2732 |
+
"step": 3570
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 3.3552014995313963,
|
| 2736 |
+
"grad_norm": 1.273992896080017,
|
| 2737 |
+
"learning_rate": 9.973787872131676e-06,
|
| 2738 |
+
"loss": 0.1069,
|
| 2739 |
+
"step": 3580
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 3.3645735707591378,
|
| 2743 |
+
"grad_norm": 1.157599687576294,
|
| 2744 |
+
"learning_rate": 9.973714653897406e-06,
|
| 2745 |
+
"loss": 0.1154,
|
| 2746 |
+
"step": 3590
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 3.373945641986879,
|
| 2750 |
+
"grad_norm": 1.567265272140503,
|
| 2751 |
+
"learning_rate": 9.973641435663139e-06,
|
| 2752 |
+
"loss": 0.1104,
|
| 2753 |
+
"step": 3600
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 3.3833177132146206,
|
| 2757 |
+
"grad_norm": 1.509450078010559,
|
| 2758 |
+
"learning_rate": 9.97356821742887e-06,
|
| 2759 |
+
"loss": 0.1123,
|
| 2760 |
+
"step": 3610
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 3.3926897844423616,
|
| 2764 |
+
"grad_norm": 1.6206624507904053,
|
| 2765 |
+
"learning_rate": 9.9734949991946e-06,
|
| 2766 |
+
"loss": 0.0915,
|
| 2767 |
+
"step": 3620
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 3.402061855670103,
|
| 2771 |
+
"grad_norm": 1.3384416103363037,
|
| 2772 |
+
"learning_rate": 9.973421780960331e-06,
|
| 2773 |
+
"loss": 0.1286,
|
| 2774 |
+
"step": 3630
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 3.4114339268978444,
|
| 2778 |
+
"grad_norm": 1.4834225177764893,
|
| 2779 |
+
"learning_rate": 9.973348562726062e-06,
|
| 2780 |
+
"loss": 0.1129,
|
| 2781 |
+
"step": 3640
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 3.420805998125586,
|
| 2785 |
+
"grad_norm": 1.486007809638977,
|
| 2786 |
+
"learning_rate": 9.973275344491793e-06,
|
| 2787 |
+
"loss": 0.1037,
|
| 2788 |
+
"step": 3650
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 3.4301780693533273,
|
| 2792 |
+
"grad_norm": 1.5038363933563232,
|
| 2793 |
+
"learning_rate": 9.973202126257523e-06,
|
| 2794 |
+
"loss": 0.104,
|
| 2795 |
+
"step": 3660
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 3.4395501405810682,
|
| 2799 |
+
"grad_norm": 1.3018808364868164,
|
| 2800 |
+
"learning_rate": 9.973128908023256e-06,
|
| 2801 |
+
"loss": 0.1068,
|
| 2802 |
+
"step": 3670
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 3.4489222118088096,
|
| 2806 |
+
"grad_norm": 1.733067512512207,
|
| 2807 |
+
"learning_rate": 9.973055689788987e-06,
|
| 2808 |
+
"loss": 0.1011,
|
| 2809 |
+
"step": 3680
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 3.458294283036551,
|
| 2813 |
+
"grad_norm": 1.3246439695358276,
|
| 2814 |
+
"learning_rate": 9.972982471554716e-06,
|
| 2815 |
+
"loss": 0.0989,
|
| 2816 |
+
"step": 3690
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 3.4676663542642925,
|
| 2820 |
+
"grad_norm": 1.7354522943496704,
|
| 2821 |
+
"learning_rate": 9.972909253320448e-06,
|
| 2822 |
+
"loss": 0.1174,
|
| 2823 |
+
"step": 3700
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 3.477038425492034,
|
| 2827 |
+
"grad_norm": 1.5907713174819946,
|
| 2828 |
+
"learning_rate": 9.972836035086179e-06,
|
| 2829 |
+
"loss": 0.1067,
|
| 2830 |
+
"step": 3710
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 3.486410496719775,
|
| 2834 |
+
"grad_norm": 1.4252599477767944,
|
| 2835 |
+
"learning_rate": 9.97276281685191e-06,
|
| 2836 |
+
"loss": 0.1064,
|
| 2837 |
+
"step": 3720
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 3.4957825679475163,
|
| 2841 |
+
"grad_norm": 1.3505686521530151,
|
| 2842 |
+
"learning_rate": 9.97268959861764e-06,
|
| 2843 |
+
"loss": 0.1168,
|
| 2844 |
+
"step": 3730
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 3.5051546391752577,
|
| 2848 |
+
"grad_norm": 1.3022727966308594,
|
| 2849 |
+
"learning_rate": 9.972616380383373e-06,
|
| 2850 |
+
"loss": 0.1111,
|
| 2851 |
+
"step": 3740
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 3.514526710402999,
|
| 2855 |
+
"grad_norm": 1.080246090888977,
|
| 2856 |
+
"learning_rate": 9.972543162149102e-06,
|
| 2857 |
+
"loss": 0.0982,
|
| 2858 |
+
"step": 3750
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 3.514526710402999,
|
| 2862 |
+
"eval_loss": 0.04514094442129135,
|
| 2863 |
+
"eval_pearson_cosine": 0.7656620144844055,
|
| 2864 |
+
"eval_pearson_dot": 0.6821019649505615,
|
| 2865 |
+
"eval_pearson_euclidean": 0.7441372871398926,
|
| 2866 |
+
"eval_pearson_manhattan": 0.7452259659767151,
|
| 2867 |
+
"eval_runtime": 22.4556,
|
| 2868 |
+
"eval_samples_per_second": 66.798,
|
| 2869 |
+
"eval_spearman_cosine": 0.7694518035767811,
|
| 2870 |
+
"eval_spearman_dot": 0.6821838150409313,
|
| 2871 |
+
"eval_spearman_euclidean": 0.7516165395512334,
|
| 2872 |
+
"eval_spearman_manhattan": 0.7527176854515762,
|
| 2873 |
+
"eval_steps_per_second": 8.372,
|
| 2874 |
+
"step": 3750
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 3.5238987816307406,
|
| 2878 |
+
"grad_norm": 1.3396129608154297,
|
| 2879 |
+
"learning_rate": 9.972469943914833e-06,
|
| 2880 |
+
"loss": 0.1145,
|
| 2881 |
+
"step": 3760
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 3.5332708528584815,
|
| 2885 |
+
"grad_norm": 1.5277647972106934,
|
| 2886 |
+
"learning_rate": 9.972396725680565e-06,
|
| 2887 |
+
"loss": 0.1101,
|
| 2888 |
+
"step": 3770
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 3.542642924086223,
|
| 2892 |
+
"grad_norm": 1.8469972610473633,
|
| 2893 |
+
"learning_rate": 9.972323507446296e-06,
|
| 2894 |
+
"loss": 0.1129,
|
| 2895 |
+
"step": 3780
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 3.5520149953139644,
|
| 2899 |
+
"grad_norm": 1.2464599609375,
|
| 2900 |
+
"learning_rate": 9.972250289212027e-06,
|
| 2901 |
+
"loss": 0.1103,
|
| 2902 |
+
"step": 3790
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 3.561387066541706,
|
| 2906 |
+
"grad_norm": 1.7863965034484863,
|
| 2907 |
+
"learning_rate": 9.972177070977757e-06,
|
| 2908 |
+
"loss": 0.1084,
|
| 2909 |
+
"step": 3800
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 3.570759137769447,
|
| 2913 |
+
"grad_norm": 1.3085591793060303,
|
| 2914 |
+
"learning_rate": 9.972103852743488e-06,
|
| 2915 |
+
"loss": 0.11,
|
| 2916 |
+
"step": 3810
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 3.580131208997188,
|
| 2920 |
+
"grad_norm": 1.5875599384307861,
|
| 2921 |
+
"learning_rate": 9.972030634509219e-06,
|
| 2922 |
+
"loss": 0.1213,
|
| 2923 |
+
"step": 3820
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 3.5895032802249296,
|
| 2927 |
+
"grad_norm": 1.2654856443405151,
|
| 2928 |
+
"learning_rate": 9.97195741627495e-06,
|
| 2929 |
+
"loss": 0.1045,
|
| 2930 |
+
"step": 3830
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 3.598875351452671,
|
| 2934 |
+
"grad_norm": 1.4713581800460815,
|
| 2935 |
+
"learning_rate": 9.97188419804068e-06,
|
| 2936 |
+
"loss": 0.1123,
|
| 2937 |
+
"step": 3840
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 3.6082474226804124,
|
| 2941 |
+
"grad_norm": 1.3559589385986328,
|
| 2942 |
+
"learning_rate": 9.971810979806413e-06,
|
| 2943 |
+
"loss": 0.1171,
|
| 2944 |
+
"step": 3850
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 3.617619493908154,
|
| 2948 |
+
"grad_norm": 1.7482990026474,
|
| 2949 |
+
"learning_rate": 9.971737761572142e-06,
|
| 2950 |
+
"loss": 0.1141,
|
| 2951 |
+
"step": 3860
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 3.626991565135895,
|
| 2955 |
+
"grad_norm": 1.7189960479736328,
|
| 2956 |
+
"learning_rate": 9.971664543337873e-06,
|
| 2957 |
+
"loss": 0.107,
|
| 2958 |
+
"step": 3870
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 3.6363636363636362,
|
| 2962 |
+
"grad_norm": 1.8246538639068604,
|
| 2963 |
+
"learning_rate": 9.971591325103605e-06,
|
| 2964 |
+
"loss": 0.1161,
|
| 2965 |
+
"step": 3880
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 3.6457357075913777,
|
| 2969 |
+
"grad_norm": 1.0778300762176514,
|
| 2970 |
+
"learning_rate": 9.971518106869336e-06,
|
| 2971 |
+
"loss": 0.1084,
|
| 2972 |
+
"step": 3890
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 3.655107778819119,
|
| 2976 |
+
"grad_norm": 1.5588942766189575,
|
| 2977 |
+
"learning_rate": 9.971444888635066e-06,
|
| 2978 |
+
"loss": 0.1038,
|
| 2979 |
+
"step": 3900
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 3.6644798500468605,
|
| 2983 |
+
"grad_norm": 1.3670451641082764,
|
| 2984 |
+
"learning_rate": 9.971371670400797e-06,
|
| 2985 |
+
"loss": 0.1069,
|
| 2986 |
+
"step": 3910
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 3.6738519212746015,
|
| 2990 |
+
"grad_norm": 1.437696099281311,
|
| 2991 |
+
"learning_rate": 9.971298452166528e-06,
|
| 2992 |
+
"loss": 0.1129,
|
| 2993 |
+
"step": 3920
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 3.683223992502343,
|
| 2997 |
+
"grad_norm": 1.39695143699646,
|
| 2998 |
+
"learning_rate": 9.971225233932259e-06,
|
| 2999 |
+
"loss": 0.1113,
|
| 3000 |
+
"step": 3930
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 3.6925960637300843,
|
| 3004 |
+
"grad_norm": 1.3372693061828613,
|
| 3005 |
+
"learning_rate": 9.97115201569799e-06,
|
| 3006 |
+
"loss": 0.1042,
|
| 3007 |
+
"step": 3940
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 3.7019681349578257,
|
| 3011 |
+
"grad_norm": 1.4336313009262085,
|
| 3012 |
+
"learning_rate": 9.971078797463722e-06,
|
| 3013 |
+
"loss": 0.1224,
|
| 3014 |
+
"step": 3950
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 3.711340206185567,
|
| 3018 |
+
"grad_norm": 1.3641144037246704,
|
| 3019 |
+
"learning_rate": 9.971005579229453e-06,
|
| 3020 |
+
"loss": 0.1082,
|
| 3021 |
+
"step": 3960
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 3.720712277413308,
|
| 3025 |
+
"grad_norm": 1.1231974363327026,
|
| 3026 |
+
"learning_rate": 9.970932360995183e-06,
|
| 3027 |
+
"loss": 0.1108,
|
| 3028 |
+
"step": 3970
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 3.7300843486410495,
|
| 3032 |
+
"grad_norm": 1.0743800401687622,
|
| 3033 |
+
"learning_rate": 9.970859142760914e-06,
|
| 3034 |
+
"loss": 0.1148,
|
| 3035 |
+
"step": 3980
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 3.739456419868791,
|
| 3039 |
+
"grad_norm": 1.5260711908340454,
|
| 3040 |
+
"learning_rate": 9.970785924526645e-06,
|
| 3041 |
+
"loss": 0.1248,
|
| 3042 |
+
"step": 3990
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 3.7488284910965324,
|
| 3046 |
+
"grad_norm": 1.1183910369873047,
|
| 3047 |
+
"learning_rate": 9.970712706292376e-06,
|
| 3048 |
+
"loss": 0.0987,
|
| 3049 |
+
"step": 4000
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 3.7488284910965324,
|
| 3053 |
+
"eval_loss": 0.046661876142024994,
|
| 3054 |
+
"eval_pearson_cosine": 0.7576525807380676,
|
| 3055 |
+
"eval_pearson_dot": 0.6644298434257507,
|
| 3056 |
+
"eval_pearson_euclidean": 0.7384845614433289,
|
| 3057 |
+
"eval_pearson_manhattan": 0.7396556735038757,
|
| 3058 |
+
"eval_runtime": 23.8808,
|
| 3059 |
+
"eval_samples_per_second": 62.812,
|
| 3060 |
+
"eval_spearman_cosine": 0.7607075839895016,
|
| 3061 |
+
"eval_spearman_dot": 0.6622737418861694,
|
| 3062 |
+
"eval_spearman_euclidean": 0.7433752629911805,
|
| 3063 |
+
"eval_spearman_manhattan": 0.7446298314535014,
|
| 3064 |
+
"eval_steps_per_second": 7.872,
|
| 3065 |
+
"step": 4000
|
| 3066 |
+
},
|
| 3067 |
+
{
|
| 3068 |
+
"epoch": 3.758200562324274,
|
| 3069 |
+
"grad_norm": 1.3608311414718628,
|
| 3070 |
+
"learning_rate": 9.970639488058106e-06,
|
| 3071 |
+
"loss": 0.1179,
|
| 3072 |
+
"step": 4010
|
| 3073 |
+
},
|
| 3074 |
+
{
|
| 3075 |
+
"epoch": 3.7675726335520148,
|
| 3076 |
+
"grad_norm": 1.6313430070877075,
|
| 3077 |
+
"learning_rate": 9.970566269823839e-06,
|
| 3078 |
+
"loss": 0.1186,
|
| 3079 |
+
"step": 4020
|
| 3080 |
+
},
|
| 3081 |
+
{
|
| 3082 |
+
"epoch": 3.776944704779756,
|
| 3083 |
+
"grad_norm": 1.4092051982879639,
|
| 3084 |
+
"learning_rate": 9.970493051589568e-06,
|
| 3085 |
+
"loss": 0.1048,
|
| 3086 |
+
"step": 4030
|
| 3087 |
+
},
|
| 3088 |
+
{
|
| 3089 |
+
"epoch": 3.7863167760074976,
|
| 3090 |
+
"grad_norm": 1.4106525182724,
|
| 3091 |
+
"learning_rate": 9.970419833355299e-06,
|
| 3092 |
+
"loss": 0.1233,
|
| 3093 |
+
"step": 4040
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 3.795688847235239,
|
| 3097 |
+
"grad_norm": 1.498146891593933,
|
| 3098 |
+
"learning_rate": 9.970346615121031e-06,
|
| 3099 |
+
"loss": 0.1164,
|
| 3100 |
+
"step": 4050
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 3.8050609184629804,
|
| 3104 |
+
"grad_norm": 1.68582284450531,
|
| 3105 |
+
"learning_rate": 9.970273396886762e-06,
|
| 3106 |
+
"loss": 0.1194,
|
| 3107 |
+
"step": 4060
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"epoch": 3.8144329896907214,
|
| 3111 |
+
"grad_norm": 1.329270362854004,
|
| 3112 |
+
"learning_rate": 9.970200178652493e-06,
|
| 3113 |
+
"loss": 0.1001,
|
| 3114 |
+
"step": 4070
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 3.823805060918463,
|
| 3118 |
+
"grad_norm": 1.6010513305664062,
|
| 3119 |
+
"learning_rate": 9.970126960418223e-06,
|
| 3120 |
+
"loss": 0.107,
|
| 3121 |
+
"step": 4080
|
| 3122 |
+
},
|
| 3123 |
+
{
|
| 3124 |
+
"epoch": 3.8331771321462043,
|
| 3125 |
+
"grad_norm": 1.213576078414917,
|
| 3126 |
+
"learning_rate": 9.970053742183954e-06,
|
| 3127 |
+
"loss": 0.1108,
|
| 3128 |
+
"step": 4090
|
| 3129 |
+
},
|
| 3130 |
+
{
|
| 3131 |
+
"epoch": 3.8425492033739457,
|
| 3132 |
+
"grad_norm": 1.585524320602417,
|
| 3133 |
+
"learning_rate": 9.969980523949685e-06,
|
| 3134 |
+
"loss": 0.1079,
|
| 3135 |
+
"step": 4100
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 3.851921274601687,
|
| 3139 |
+
"grad_norm": 1.6043713092803955,
|
| 3140 |
+
"learning_rate": 9.969907305715416e-06,
|
| 3141 |
+
"loss": 0.1141,
|
| 3142 |
+
"step": 4110
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 3.861293345829428,
|
| 3146 |
+
"grad_norm": 1.3566473722457886,
|
| 3147 |
+
"learning_rate": 9.969834087481146e-06,
|
| 3148 |
+
"loss": 0.1148,
|
| 3149 |
+
"step": 4120
|
| 3150 |
+
},
|
| 3151 |
+
{
|
| 3152 |
+
"epoch": 3.8706654170571695,
|
| 3153 |
+
"grad_norm": 1.390787124633789,
|
| 3154 |
+
"learning_rate": 9.969760869246879e-06,
|
| 3155 |
+
"loss": 0.1024,
|
| 3156 |
+
"step": 4130
|
| 3157 |
+
},
|
| 3158 |
+
{
|
| 3159 |
+
"epoch": 3.880037488284911,
|
| 3160 |
+
"grad_norm": 1.689005970954895,
|
| 3161 |
+
"learning_rate": 9.96968765101261e-06,
|
| 3162 |
+
"loss": 0.111,
|
| 3163 |
+
"step": 4140
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 3.8894095595126523,
|
| 3167 |
+
"grad_norm": 1.850071907043457,
|
| 3168 |
+
"learning_rate": 9.96961443277834e-06,
|
| 3169 |
+
"loss": 0.1097,
|
| 3170 |
+
"step": 4150
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 3.8987816307403937,
|
| 3174 |
+
"grad_norm": 1.4834603071212769,
|
| 3175 |
+
"learning_rate": 9.969541214544071e-06,
|
| 3176 |
+
"loss": 0.1084,
|
| 3177 |
+
"step": 4160
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 3.9081537019681347,
|
| 3181 |
+
"grad_norm": 1.3408997058868408,
|
| 3182 |
+
"learning_rate": 9.969467996309802e-06,
|
| 3183 |
+
"loss": 0.1194,
|
| 3184 |
+
"step": 4170
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 3.917525773195876,
|
| 3188 |
+
"grad_norm": 1.3920304775238037,
|
| 3189 |
+
"learning_rate": 9.969394778075533e-06,
|
| 3190 |
+
"loss": 0.1091,
|
| 3191 |
+
"step": 4180
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 3.9268978444236176,
|
| 3195 |
+
"grad_norm": 1.0026508569717407,
|
| 3196 |
+
"learning_rate": 9.969321559841263e-06,
|
| 3197 |
+
"loss": 0.119,
|
| 3198 |
+
"step": 4190
|
| 3199 |
+
},
|
| 3200 |
+
{
|
| 3201 |
+
"epoch": 3.936269915651359,
|
| 3202 |
+
"grad_norm": 1.7984665632247925,
|
| 3203 |
+
"learning_rate": 9.969248341606996e-06,
|
| 3204 |
+
"loss": 0.1065,
|
| 3205 |
+
"step": 4200
|
| 3206 |
+
},
|
| 3207 |
+
{
|
| 3208 |
+
"epoch": 3.9456419868791004,
|
| 3209 |
+
"grad_norm": 1.6500909328460693,
|
| 3210 |
+
"learning_rate": 9.969175123372725e-06,
|
| 3211 |
+
"loss": 0.1083,
|
| 3212 |
+
"step": 4210
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 3.9550140581068414,
|
| 3216 |
+
"grad_norm": 1.7580713033676147,
|
| 3217 |
+
"learning_rate": 9.969101905138456e-06,
|
| 3218 |
+
"loss": 0.1237,
|
| 3219 |
+
"step": 4220
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 3.964386129334583,
|
| 3223 |
+
"grad_norm": 1.8374171257019043,
|
| 3224 |
+
"learning_rate": 9.969028686904188e-06,
|
| 3225 |
+
"loss": 0.1003,
|
| 3226 |
+
"step": 4230
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 3.973758200562324,
|
| 3230 |
+
"grad_norm": 1.5857341289520264,
|
| 3231 |
+
"learning_rate": 9.968955468669919e-06,
|
| 3232 |
+
"loss": 0.1012,
|
| 3233 |
+
"step": 4240
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 3.9831302717900656,
|
| 3237 |
+
"grad_norm": 1.627947211265564,
|
| 3238 |
+
"learning_rate": 9.96888225043565e-06,
|
| 3239 |
+
"loss": 0.1111,
|
| 3240 |
+
"step": 4250
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 3.9831302717900656,
|
| 3244 |
+
"eval_loss": 0.04063473269343376,
|
| 3245 |
+
"eval_pearson_cosine": 0.7690664529800415,
|
| 3246 |
+
"eval_pearson_dot": 0.6998196840286255,
|
| 3247 |
+
"eval_pearson_euclidean": 0.7456687092781067,
|
| 3248 |
+
"eval_pearson_manhattan": 0.7471497058868408,
|
| 3249 |
+
"eval_runtime": 23.0817,
|
| 3250 |
+
"eval_samples_per_second": 64.986,
|
| 3251 |
+
"eval_spearman_cosine": 0.7702784084250337,
|
| 3252 |
+
"eval_spearman_dot": 0.7005907360024843,
|
| 3253 |
+
"eval_spearman_euclidean": 0.7509877657044322,
|
| 3254 |
+
"eval_spearman_manhattan": 0.7524785559548752,
|
| 3255 |
+
"eval_steps_per_second": 8.145,
|
| 3256 |
+
"step": 4250
|
| 3257 |
+
},
|
| 3258 |
+
{
|
| 3259 |
+
"epoch": 3.992502343017807,
|
| 3260 |
+
"grad_norm": 1.3161486387252808,
|
| 3261 |
+
"learning_rate": 9.96880903220138e-06,
|
| 3262 |
+
"loss": 0.1114,
|
| 3263 |
+
"step": 4260
|
| 3264 |
+
},
|
| 3265 |
+
{
|
| 3266 |
+
"epoch": 4.001874414245548,
|
| 3267 |
+
"grad_norm": 0.9556475281715393,
|
| 3268 |
+
"learning_rate": 9.968735813967111e-06,
|
| 3269 |
+
"loss": 0.1141,
|
| 3270 |
+
"step": 4270
|
| 3271 |
+
},
|
| 3272 |
+
{
|
| 3273 |
+
"epoch": 4.01124648547329,
|
| 3274 |
+
"grad_norm": 1.0041595697402954,
|
| 3275 |
+
"learning_rate": 9.968662595732842e-06,
|
| 3276 |
+
"loss": 0.0807,
|
| 3277 |
+
"step": 4280
|
| 3278 |
+
},
|
| 3279 |
+
{
|
| 3280 |
+
"epoch": 4.020618556701031,
|
| 3281 |
+
"grad_norm": 1.1500684022903442,
|
| 3282 |
+
"learning_rate": 9.968589377498573e-06,
|
| 3283 |
+
"loss": 0.0701,
|
| 3284 |
+
"step": 4290
|
| 3285 |
+
},
|
| 3286 |
+
{
|
| 3287 |
+
"epoch": 4.029990627928772,
|
| 3288 |
+
"grad_norm": 1.3963230848312378,
|
| 3289 |
+
"learning_rate": 9.968516159264305e-06,
|
| 3290 |
+
"loss": 0.0863,
|
| 3291 |
+
"step": 4300
|
| 3292 |
+
},
|
| 3293 |
+
{
|
| 3294 |
+
"epoch": 4.039362699156514,
|
| 3295 |
+
"grad_norm": 1.4251878261566162,
|
| 3296 |
+
"learning_rate": 9.968442941030036e-06,
|
| 3297 |
+
"loss": 0.0746,
|
| 3298 |
+
"step": 4310
|
| 3299 |
+
},
|
| 3300 |
+
{
|
| 3301 |
+
"epoch": 4.048734770384255,
|
| 3302 |
+
"grad_norm": 1.0674968957901,
|
| 3303 |
+
"learning_rate": 9.968369722795765e-06,
|
| 3304 |
+
"loss": 0.0667,
|
| 3305 |
+
"step": 4320
|
| 3306 |
+
},
|
| 3307 |
+
{
|
| 3308 |
+
"epoch": 4.0581068416119965,
|
| 3309 |
+
"grad_norm": 1.2465558052062988,
|
| 3310 |
+
"learning_rate": 9.968296504561497e-06,
|
| 3311 |
+
"loss": 0.0773,
|
| 3312 |
+
"step": 4330
|
| 3313 |
+
},
|
| 3314 |
+
{
|
| 3315 |
+
"epoch": 4.0674789128397375,
|
| 3316 |
+
"grad_norm": 1.409511923789978,
|
| 3317 |
+
"learning_rate": 9.968223286327228e-06,
|
| 3318 |
+
"loss": 0.0775,
|
| 3319 |
+
"step": 4340
|
| 3320 |
+
},
|
| 3321 |
+
{
|
| 3322 |
+
"epoch": 4.0768509840674785,
|
| 3323 |
+
"grad_norm": 1.2048633098602295,
|
| 3324 |
+
"learning_rate": 9.968150068092959e-06,
|
| 3325 |
+
"loss": 0.0885,
|
| 3326 |
+
"step": 4350
|
| 3327 |
+
},
|
| 3328 |
+
{
|
| 3329 |
+
"epoch": 4.08622305529522,
|
| 3330 |
+
"grad_norm": 1.3504215478897095,
|
| 3331 |
+
"learning_rate": 9.96807684985869e-06,
|
| 3332 |
+
"loss": 0.0802,
|
| 3333 |
+
"step": 4360
|
| 3334 |
+
},
|
| 3335 |
+
{
|
| 3336 |
+
"epoch": 4.095595126522961,
|
| 3337 |
+
"grad_norm": 1.5094915628433228,
|
| 3338 |
+
"learning_rate": 9.96800363162442e-06,
|
| 3339 |
+
"loss": 0.0889,
|
| 3340 |
+
"step": 4370
|
| 3341 |
+
},
|
| 3342 |
+
{
|
| 3343 |
+
"epoch": 4.104967197750703,
|
| 3344 |
+
"grad_norm": 1.2075692415237427,
|
| 3345 |
+
"learning_rate": 9.967930413390151e-06,
|
| 3346 |
+
"loss": 0.0718,
|
| 3347 |
+
"step": 4380
|
| 3348 |
+
},
|
| 3349 |
+
{
|
| 3350 |
+
"epoch": 4.114339268978444,
|
| 3351 |
+
"grad_norm": 1.476462960243225,
|
| 3352 |
+
"learning_rate": 9.967857195155882e-06,
|
| 3353 |
+
"loss": 0.0809,
|
| 3354 |
+
"step": 4390
|
| 3355 |
+
},
|
| 3356 |
+
{
|
| 3357 |
+
"epoch": 4.123711340206185,
|
| 3358 |
+
"grad_norm": 1.4811893701553345,
|
| 3359 |
+
"learning_rate": 9.967783976921614e-06,
|
| 3360 |
+
"loss": 0.082,
|
| 3361 |
+
"step": 4400
|
| 3362 |
+
},
|
| 3363 |
+
{
|
| 3364 |
+
"epoch": 4.133083411433927,
|
| 3365 |
+
"grad_norm": 1.3016406297683716,
|
| 3366 |
+
"learning_rate": 9.967710758687345e-06,
|
| 3367 |
+
"loss": 0.0867,
|
| 3368 |
+
"step": 4410
|
| 3369 |
+
},
|
| 3370 |
+
{
|
| 3371 |
+
"epoch": 4.142455482661668,
|
| 3372 |
+
"grad_norm": 1.3254297971725464,
|
| 3373 |
+
"learning_rate": 9.967637540453076e-06,
|
| 3374 |
+
"loss": 0.0783,
|
| 3375 |
+
"step": 4420
|
| 3376 |
+
},
|
| 3377 |
+
{
|
| 3378 |
+
"epoch": 4.15182755388941,
|
| 3379 |
+
"grad_norm": 1.7814503908157349,
|
| 3380 |
+
"learning_rate": 9.967564322218806e-06,
|
| 3381 |
+
"loss": 0.0812,
|
| 3382 |
+
"step": 4430
|
| 3383 |
+
},
|
| 3384 |
+
{
|
| 3385 |
+
"epoch": 4.161199625117151,
|
| 3386 |
+
"grad_norm": 1.3375070095062256,
|
| 3387 |
+
"learning_rate": 9.967491103984537e-06,
|
| 3388 |
+
"loss": 0.0835,
|
| 3389 |
+
"step": 4440
|
| 3390 |
+
},
|
| 3391 |
+
{
|
| 3392 |
+
"epoch": 4.170571696344892,
|
| 3393 |
+
"grad_norm": 1.3573247194290161,
|
| 3394 |
+
"learning_rate": 9.967417885750268e-06,
|
| 3395 |
+
"loss": 0.0772,
|
| 3396 |
+
"step": 4450
|
| 3397 |
+
},
|
| 3398 |
+
{
|
| 3399 |
+
"epoch": 4.179943767572634,
|
| 3400 |
+
"grad_norm": 1.601321816444397,
|
| 3401 |
+
"learning_rate": 9.967344667515999e-06,
|
| 3402 |
+
"loss": 0.0785,
|
| 3403 |
+
"step": 4460
|
| 3404 |
+
},
|
| 3405 |
+
{
|
| 3406 |
+
"epoch": 4.189315838800375,
|
| 3407 |
+
"grad_norm": 1.0777158737182617,
|
| 3408 |
+
"learning_rate": 9.96727144928173e-06,
|
| 3409 |
+
"loss": 0.0789,
|
| 3410 |
+
"step": 4470
|
| 3411 |
+
},
|
| 3412 |
+
{
|
| 3413 |
+
"epoch": 4.1986879100281165,
|
| 3414 |
+
"grad_norm": 1.717281699180603,
|
| 3415 |
+
"learning_rate": 9.967198231047462e-06,
|
| 3416 |
+
"loss": 0.0876,
|
| 3417 |
+
"step": 4480
|
| 3418 |
+
},
|
| 3419 |
+
{
|
| 3420 |
+
"epoch": 4.2080599812558575,
|
| 3421 |
+
"grad_norm": 1.6537655591964722,
|
| 3422 |
+
"learning_rate": 9.967125012813191e-06,
|
| 3423 |
+
"loss": 0.0859,
|
| 3424 |
+
"step": 4490
|
| 3425 |
+
},
|
| 3426 |
+
{
|
| 3427 |
+
"epoch": 4.217432052483598,
|
| 3428 |
+
"grad_norm": 1.3347113132476807,
|
| 3429 |
+
"learning_rate": 9.967051794578922e-06,
|
| 3430 |
+
"loss": 0.0888,
|
| 3431 |
+
"step": 4500
|
| 3432 |
+
},
|
| 3433 |
+
{
|
| 3434 |
+
"epoch": 4.217432052483598,
|
| 3435 |
+
"eval_loss": 0.042121224105358124,
|
| 3436 |
+
"eval_pearson_cosine": 0.7580196857452393,
|
| 3437 |
+
"eval_pearson_dot": 0.6874213814735413,
|
| 3438 |
+
"eval_pearson_euclidean": 0.740117073059082,
|
| 3439 |
+
"eval_pearson_manhattan": 0.7411655187606812,
|
| 3440 |
+
"eval_runtime": 22.046,
|
| 3441 |
+
"eval_samples_per_second": 68.04,
|
| 3442 |
+
"eval_spearman_cosine": 0.7598083870591178,
|
| 3443 |
+
"eval_spearman_dot": 0.6866180590359211,
|
| 3444 |
+
"eval_spearman_euclidean": 0.7457408658977246,
|
| 3445 |
+
"eval_spearman_manhattan": 0.7467901472090236,
|
| 3446 |
+
"eval_steps_per_second": 8.528,
|
| 3447 |
+
"step": 4500
|
| 3448 |
+
},
|
| 3449 |
+
{
|
| 3450 |
+
"epoch": 4.22680412371134,
|
| 3451 |
+
"grad_norm": 1.283334732055664,
|
| 3452 |
+
"learning_rate": 9.966978576344654e-06,
|
| 3453 |
+
"loss": 0.0824,
|
| 3454 |
+
"step": 4510
|
| 3455 |
+
},
|
| 3456 |
+
{
|
| 3457 |
+
"epoch": 4.236176194939081,
|
| 3458 |
+
"grad_norm": 1.4807559251785278,
|
| 3459 |
+
"learning_rate": 9.966905358110385e-06,
|
| 3460 |
+
"loss": 0.0812,
|
| 3461 |
+
"step": 4520
|
| 3462 |
+
},
|
| 3463 |
+
{
|
| 3464 |
+
"epoch": 4.245548266166823,
|
| 3465 |
+
"grad_norm": 1.1873483657836914,
|
| 3466 |
+
"learning_rate": 9.966832139876116e-06,
|
| 3467 |
+
"loss": 0.0788,
|
| 3468 |
+
"step": 4530
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"epoch": 4.254920337394564,
|
| 3472 |
+
"grad_norm": 1.27379310131073,
|
| 3473 |
+
"learning_rate": 9.966758921641846e-06,
|
| 3474 |
+
"loss": 0.0802,
|
| 3475 |
+
"step": 4540
|
| 3476 |
+
},
|
| 3477 |
+
{
|
| 3478 |
+
"epoch": 4.264292408622305,
|
| 3479 |
+
"grad_norm": 1.3721706867218018,
|
| 3480 |
+
"learning_rate": 9.966685703407577e-06,
|
| 3481 |
+
"loss": 0.0776,
|
| 3482 |
+
"step": 4550
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 4.273664479850047,
|
| 3486 |
+
"grad_norm": 1.4129197597503662,
|
| 3487 |
+
"learning_rate": 9.966612485173308e-06,
|
| 3488 |
+
"loss": 0.0924,
|
| 3489 |
+
"step": 4560
|
| 3490 |
+
},
|
| 3491 |
+
{
|
| 3492 |
+
"epoch": 4.283036551077788,
|
| 3493 |
+
"grad_norm": 1.453730821609497,
|
| 3494 |
+
"learning_rate": 9.966539266939039e-06,
|
| 3495 |
+
"loss": 0.0823,
|
| 3496 |
+
"step": 4570
|
| 3497 |
+
},
|
| 3498 |
+
{
|
| 3499 |
+
"epoch": 4.29240862230553,
|
| 3500 |
+
"grad_norm": 1.4608802795410156,
|
| 3501 |
+
"learning_rate": 9.966466048704771e-06,
|
| 3502 |
+
"loss": 0.0806,
|
| 3503 |
+
"step": 4580
|
| 3504 |
+
},
|
| 3505 |
+
{
|
| 3506 |
+
"epoch": 4.301780693533271,
|
| 3507 |
+
"grad_norm": 1.0814175605773926,
|
| 3508 |
+
"learning_rate": 9.966392830470502e-06,
|
| 3509 |
+
"loss": 0.0781,
|
| 3510 |
+
"step": 4590
|
| 3511 |
+
},
|
| 3512 |
+
{
|
| 3513 |
+
"epoch": 4.311152764761012,
|
| 3514 |
+
"grad_norm": 1.9891834259033203,
|
| 3515 |
+
"learning_rate": 9.966319612236233e-06,
|
| 3516 |
+
"loss": 0.0792,
|
| 3517 |
+
"step": 4600
|
| 3518 |
+
},
|
| 3519 |
+
{
|
| 3520 |
+
"epoch": 4.320524835988754,
|
| 3521 |
+
"grad_norm": 0.7774847745895386,
|
| 3522 |
+
"learning_rate": 9.966246394001963e-06,
|
| 3523 |
+
"loss": 0.0734,
|
| 3524 |
+
"step": 4610
|
| 3525 |
+
},
|
| 3526 |
+
{
|
| 3527 |
+
"epoch": 4.329896907216495,
|
| 3528 |
+
"grad_norm": 2.0921082496643066,
|
| 3529 |
+
"learning_rate": 9.966173175767694e-06,
|
| 3530 |
+
"loss": 0.0789,
|
| 3531 |
+
"step": 4620
|
| 3532 |
+
},
|
| 3533 |
+
{
|
| 3534 |
+
"epoch": 4.339268978444236,
|
| 3535 |
+
"grad_norm": 1.4378306865692139,
|
| 3536 |
+
"learning_rate": 9.966099957533425e-06,
|
| 3537 |
+
"loss": 0.0829,
|
| 3538 |
+
"step": 4630
|
| 3539 |
+
},
|
| 3540 |
+
{
|
| 3541 |
+
"epoch": 4.348641049671977,
|
| 3542 |
+
"grad_norm": 1.5577812194824219,
|
| 3543 |
+
"learning_rate": 9.966026739299156e-06,
|
| 3544 |
+
"loss": 0.0782,
|
| 3545 |
+
"step": 4640
|
| 3546 |
+
},
|
| 3547 |
+
{
|
| 3548 |
+
"epoch": 4.358013120899718,
|
| 3549 |
+
"grad_norm": 1.8791301250457764,
|
| 3550 |
+
"learning_rate": 9.965953521064888e-06,
|
| 3551 |
+
"loss": 0.088,
|
| 3552 |
+
"step": 4650
|
| 3553 |
+
},
|
| 3554 |
+
{
|
| 3555 |
+
"epoch": 4.36738519212746,
|
| 3556 |
+
"grad_norm": 0.8537359833717346,
|
| 3557 |
+
"learning_rate": 9.965880302830617e-06,
|
| 3558 |
+
"loss": 0.0766,
|
| 3559 |
+
"step": 4660
|
| 3560 |
+
},
|
| 3561 |
+
{
|
| 3562 |
+
"epoch": 4.376757263355201,
|
| 3563 |
+
"grad_norm": 1.258042573928833,
|
| 3564 |
+
"learning_rate": 9.965807084596348e-06,
|
| 3565 |
+
"loss": 0.0877,
|
| 3566 |
+
"step": 4670
|
| 3567 |
+
},
|
| 3568 |
+
{
|
| 3569 |
+
"epoch": 4.386129334582943,
|
| 3570 |
+
"grad_norm": 1.5519142150878906,
|
| 3571 |
+
"learning_rate": 9.96573386636208e-06,
|
| 3572 |
+
"loss": 0.0881,
|
| 3573 |
+
"step": 4680
|
| 3574 |
+
},
|
| 3575 |
+
{
|
| 3576 |
+
"epoch": 4.395501405810684,
|
| 3577 |
+
"grad_norm": 1.1437076330184937,
|
| 3578 |
+
"learning_rate": 9.965660648127811e-06,
|
| 3579 |
+
"loss": 0.0816,
|
| 3580 |
+
"step": 4690
|
| 3581 |
+
},
|
| 3582 |
+
{
|
| 3583 |
+
"epoch": 4.404873477038425,
|
| 3584 |
+
"grad_norm": 1.3333864212036133,
|
| 3585 |
+
"learning_rate": 9.965587429893542e-06,
|
| 3586 |
+
"loss": 0.0818,
|
| 3587 |
+
"step": 4700
|
| 3588 |
+
},
|
| 3589 |
+
{
|
| 3590 |
+
"epoch": 4.414245548266167,
|
| 3591 |
+
"grad_norm": 1.403075098991394,
|
| 3592 |
+
"learning_rate": 9.965514211659273e-06,
|
| 3593 |
+
"loss": 0.0771,
|
| 3594 |
+
"step": 4710
|
| 3595 |
+
},
|
| 3596 |
+
{
|
| 3597 |
+
"epoch": 4.423617619493908,
|
| 3598 |
+
"grad_norm": 1.3652963638305664,
|
| 3599 |
+
"learning_rate": 9.965440993425003e-06,
|
| 3600 |
+
"loss": 0.0692,
|
| 3601 |
+
"step": 4720
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"epoch": 4.43298969072165,
|
| 3605 |
+
"grad_norm": 1.4429869651794434,
|
| 3606 |
+
"learning_rate": 9.965367775190734e-06,
|
| 3607 |
+
"loss": 0.0846,
|
| 3608 |
+
"step": 4730
|
| 3609 |
+
},
|
| 3610 |
+
{
|
| 3611 |
+
"epoch": 4.442361761949391,
|
| 3612 |
+
"grad_norm": 1.291710376739502,
|
| 3613 |
+
"learning_rate": 9.965294556956465e-06,
|
| 3614 |
+
"loss": 0.0796,
|
| 3615 |
+
"step": 4740
|
| 3616 |
+
},
|
| 3617 |
+
{
|
| 3618 |
+
"epoch": 4.451733833177133,
|
| 3619 |
+
"grad_norm": 1.4110385179519653,
|
| 3620 |
+
"learning_rate": 9.965221338722196e-06,
|
| 3621 |
+
"loss": 0.0756,
|
| 3622 |
+
"step": 4750
|
| 3623 |
+
},
|
| 3624 |
+
{
|
| 3625 |
+
"epoch": 4.451733833177133,
|
| 3626 |
+
"eval_loss": 0.039456192404031754,
|
| 3627 |
+
"eval_pearson_cosine": 0.7664028406143188,
|
| 3628 |
+
"eval_pearson_dot": 0.7008457779884338,
|
| 3629 |
+
"eval_pearson_euclidean": 0.7418538928031921,
|
| 3630 |
+
"eval_pearson_manhattan": 0.7431594133377075,
|
| 3631 |
+
"eval_runtime": 23.3602,
|
| 3632 |
+
"eval_samples_per_second": 64.212,
|
| 3633 |
+
"eval_spearman_cosine": 0.7673929323503452,
|
| 3634 |
+
"eval_spearman_dot": 0.7011750025269451,
|
| 3635 |
+
"eval_spearman_euclidean": 0.7464768579915497,
|
| 3636 |
+
"eval_spearman_manhattan": 0.7479944496608657,
|
| 3637 |
+
"eval_steps_per_second": 8.048,
|
| 3638 |
+
"step": 4750
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 4.4611059044048735,
|
| 3642 |
+
"grad_norm": 1.1584782600402832,
|
| 3643 |
+
"learning_rate": 9.965148120487928e-06,
|
| 3644 |
+
"loss": 0.0834,
|
| 3645 |
+
"step": 4760
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 4.4704779756326145,
|
| 3649 |
+
"grad_norm": 1.2065712213516235,
|
| 3650 |
+
"learning_rate": 9.965074902253659e-06,
|
| 3651 |
+
"loss": 0.0865,
|
| 3652 |
+
"step": 4770
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 4.479850046860356,
|
| 3656 |
+
"grad_norm": 1.3458271026611328,
|
| 3657 |
+
"learning_rate": 9.965001684019388e-06,
|
| 3658 |
+
"loss": 0.0764,
|
| 3659 |
+
"step": 4780
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 4.489222118088097,
|
| 3663 |
+
"grad_norm": 2.0091888904571533,
|
| 3664 |
+
"learning_rate": 9.96492846578512e-06,
|
| 3665 |
+
"loss": 0.0773,
|
| 3666 |
+
"step": 4790
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 4.498594189315839,
|
| 3670 |
+
"grad_norm": 1.3832370042800903,
|
| 3671 |
+
"learning_rate": 9.964855247550851e-06,
|
| 3672 |
+
"loss": 0.0806,
|
| 3673 |
+
"step": 4800
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 4.50796626054358,
|
| 3677 |
+
"grad_norm": 1.4656741619110107,
|
| 3678 |
+
"learning_rate": 9.964782029316582e-06,
|
| 3679 |
+
"loss": 0.0852,
|
| 3680 |
+
"step": 4810
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 4.517338331771321,
|
| 3684 |
+
"grad_norm": 1.3915668725967407,
|
| 3685 |
+
"learning_rate": 9.964708811082312e-06,
|
| 3686 |
+
"loss": 0.086,
|
| 3687 |
+
"step": 4820
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 4.526710402999063,
|
| 3691 |
+
"grad_norm": 1.2182085514068604,
|
| 3692 |
+
"learning_rate": 9.964635592848043e-06,
|
| 3693 |
+
"loss": 0.0777,
|
| 3694 |
+
"step": 4830
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 4.536082474226804,
|
| 3698 |
+
"grad_norm": 1.2041029930114746,
|
| 3699 |
+
"learning_rate": 9.964562374613774e-06,
|
| 3700 |
+
"loss": 0.0738,
|
| 3701 |
+
"step": 4840
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 4.545454545454545,
|
| 3705 |
+
"grad_norm": 1.289475917816162,
|
| 3706 |
+
"learning_rate": 9.964489156379505e-06,
|
| 3707 |
+
"loss": 0.0723,
|
| 3708 |
+
"step": 4850
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 4.554826616682287,
|
| 3712 |
+
"grad_norm": 1.8206441402435303,
|
| 3713 |
+
"learning_rate": 9.964415938145237e-06,
|
| 3714 |
+
"loss": 0.0823,
|
| 3715 |
+
"step": 4860
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 4.564198687910028,
|
| 3719 |
+
"grad_norm": 1.393254280090332,
|
| 3720 |
+
"learning_rate": 9.964342719910968e-06,
|
| 3721 |
+
"loss": 0.0869,
|
| 3722 |
+
"step": 4870
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 4.57357075913777,
|
| 3726 |
+
"grad_norm": 1.6424909830093384,
|
| 3727 |
+
"learning_rate": 9.964269501676699e-06,
|
| 3728 |
+
"loss": 0.0721,
|
| 3729 |
+
"step": 4880
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 4.582942830365511,
|
| 3733 |
+
"grad_norm": 1.6760517358779907,
|
| 3734 |
+
"learning_rate": 9.96419628344243e-06,
|
| 3735 |
+
"loss": 0.0849,
|
| 3736 |
+
"step": 4890
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 4.592314901593252,
|
| 3740 |
+
"grad_norm": 1.4797537326812744,
|
| 3741 |
+
"learning_rate": 9.96412306520816e-06,
|
| 3742 |
+
"loss": 0.0815,
|
| 3743 |
+
"step": 4900
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 4.6016869728209935,
|
| 3747 |
+
"grad_norm": 1.3184549808502197,
|
| 3748 |
+
"learning_rate": 9.964049846973891e-06,
|
| 3749 |
+
"loss": 0.0875,
|
| 3750 |
+
"step": 4910
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 4.6110590440487345,
|
| 3754 |
+
"grad_norm": 1.0524438619613647,
|
| 3755 |
+
"learning_rate": 9.963976628739622e-06,
|
| 3756 |
+
"loss": 0.0821,
|
| 3757 |
+
"step": 4920
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 4.620431115276476,
|
| 3761 |
+
"grad_norm": 0.8284000158309937,
|
| 3762 |
+
"learning_rate": 9.963903410505354e-06,
|
| 3763 |
+
"loss": 0.0737,
|
| 3764 |
+
"step": 4930
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 4.629803186504217,
|
| 3768 |
+
"grad_norm": 1.2979810237884521,
|
| 3769 |
+
"learning_rate": 9.963830192271085e-06,
|
| 3770 |
+
"loss": 0.1031,
|
| 3771 |
+
"step": 4940
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 4.639175257731958,
|
| 3775 |
+
"grad_norm": 1.2484486103057861,
|
| 3776 |
+
"learning_rate": 9.963756974036814e-06,
|
| 3777 |
+
"loss": 0.0853,
|
| 3778 |
+
"step": 4950
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 4.6485473289597,
|
| 3782 |
+
"grad_norm": 1.4267854690551758,
|
| 3783 |
+
"learning_rate": 9.963683755802546e-06,
|
| 3784 |
+
"loss": 0.0784,
|
| 3785 |
+
"step": 4960
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 4.657919400187441,
|
| 3789 |
+
"grad_norm": 1.2631357908248901,
|
| 3790 |
+
"learning_rate": 9.963610537568277e-06,
|
| 3791 |
+
"loss": 0.0814,
|
| 3792 |
+
"step": 4970
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 4.667291471415183,
|
| 3796 |
+
"grad_norm": 1.5679900646209717,
|
| 3797 |
+
"learning_rate": 9.963537319334008e-06,
|
| 3798 |
+
"loss": 0.0851,
|
| 3799 |
+
"step": 4980
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 4.676663542642924,
|
| 3803 |
+
"grad_norm": 1.216604471206665,
|
| 3804 |
+
"learning_rate": 9.963464101099739e-06,
|
| 3805 |
+
"loss": 0.0747,
|
| 3806 |
+
"step": 4990
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 4.686035613870665,
|
| 3810 |
+
"grad_norm": 1.3772624731063843,
|
| 3811 |
+
"learning_rate": 9.96339088286547e-06,
|
| 3812 |
+
"loss": 0.0871,
|
| 3813 |
+
"step": 5000
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 4.686035613870665,
|
| 3817 |
+
"eval_loss": 0.041086822748184204,
|
| 3818 |
+
"eval_pearson_cosine": 0.7587878704071045,
|
| 3819 |
+
"eval_pearson_dot": 0.6872098445892334,
|
| 3820 |
+
"eval_pearson_euclidean": 0.7388917207717896,
|
| 3821 |
+
"eval_pearson_manhattan": 0.7404583692550659,
|
| 3822 |
+
"eval_runtime": 22.5042,
|
| 3823 |
+
"eval_samples_per_second": 66.654,
|
| 3824 |
+
"eval_spearman_cosine": 0.7603871650644157,
|
| 3825 |
+
"eval_spearman_dot": 0.6866960900397536,
|
| 3826 |
+
"eval_spearman_euclidean": 0.7440960862957542,
|
| 3827 |
+
"eval_spearman_manhattan": 0.745568766414613,
|
| 3828 |
+
"eval_steps_per_second": 8.354,
|
| 3829 |
+
"step": 5000
|
| 3830 |
}
|
| 3831 |
],
|
| 3832 |
"logging_steps": 10,
|