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 +1761 -3
last-checkpoint/model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 368988278
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f673b823be36f3eb97a0d1d83dac231599658a1bf8bcffea23f67cafcc109b48
|
3 |
size 368988278
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1107079290
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fe2f2bef345acda275dcd6003cb182f7f1e032680f3a4edeefc1b29e48691847
|
3 |
size 1107079290
|
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 sha256:74386f26f36ed67f56395205881e5db2d0c28ffcbeed50dd95b28771d2dac588
|
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 sha256:41c88f9de084200454883a13c3717941ea3fd433e2f8735507fc30611f9c5501
|
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 sha256:965b00d4cb4710ebab57c8787b9925bb3f77b8eeba94a186ec4bc1c2f326ef3f
|
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:d5dc374b8b9a4c45c950f9d136feab85a767081fa59f0c7d68ed3a62060c4949
|
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 sha256:5c7c212fb779217f1edac0baf44f67b608eefc1e0e4e3f5a9dd7eb557032c1bc
|
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 sha256:86e1effd626ce1e95dd68a0c8089fe19218f2b24dfe9e45ed2cab1c0ebc10ba1
|
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 sha256:799cc83f60dfc1c4243cfd6403592112414a2eec494e6832f10221c96ff62c20
|
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 sha256:586777c398770c3255d3a1f48c7fef44ea9d89117c627c9ea490e16bfd9a49ba
|
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 sha256:34fccfe9285534cc1461fe7c44d5b307057be9ab740b9c40800411e6c74ade04
|
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": 0.
|
5 |
"eval_steps": 2500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -1765,6 +1765,1764 @@
|
|
1765 |
"eval_samples_per_second": 2008.956,
|
1766 |
"eval_steps_per_second": 31.391,
|
1767 |
"step": 2500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1768 |
}
|
1769 |
],
|
1770 |
"logging_steps": 10,
|
@@ -1784,7 +3542,7 @@
|
|
1784 |
"attributes": {}
|
1785 |
}
|
1786 |
},
|
1787 |
-
"total_flos":
|
1788 |
"train_batch_size": 16,
|
1789 |
"trial_name": null,
|
1790 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.3616309556098002,
|
5 |
"eval_steps": 2500,
|
6 |
+
"global_step": 5000,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
1765 |
"eval_samples_per_second": 2008.956,
|
1766 |
"eval_steps_per_second": 31.391,
|
1767 |
"step": 2500
|
1768 |
+
},
|
1769 |
+
{
|
1770 |
+
"epoch": 0.1815387397161197,
|
1771 |
+
"grad_norm": 15.7265625,
|
1772 |
+
"learning_rate": 9.985817342056891e-07,
|
1773 |
+
"loss": 24.7492,
|
1774 |
+
"step": 2510
|
1775 |
+
},
|
1776 |
+
{
|
1777 |
+
"epoch": 0.1822620016273393,
|
1778 |
+
"grad_norm": 16.046875,
|
1779 |
+
"learning_rate": 9.985760837443573e-07,
|
1780 |
+
"loss": 24.7608,
|
1781 |
+
"step": 2520
|
1782 |
+
},
|
1783 |
+
{
|
1784 |
+
"epoch": 0.1829852635385589,
|
1785 |
+
"grad_norm": 16.90625,
|
1786 |
+
"learning_rate": 9.985704332830253e-07,
|
1787 |
+
"loss": 24.9355,
|
1788 |
+
"step": 2530
|
1789 |
+
},
|
1790 |
+
{
|
1791 |
+
"epoch": 0.1837085254497785,
|
1792 |
+
"grad_norm": 15.5625,
|
1793 |
+
"learning_rate": 9.985647828216935e-07,
|
1794 |
+
"loss": 25.3485,
|
1795 |
+
"step": 2540
|
1796 |
+
},
|
1797 |
+
{
|
1798 |
+
"epoch": 0.1844317873609981,
|
1799 |
+
"grad_norm": 15.8671875,
|
1800 |
+
"learning_rate": 9.985591323603615e-07,
|
1801 |
+
"loss": 25.1266,
|
1802 |
+
"step": 2550
|
1803 |
+
},
|
1804 |
+
{
|
1805 |
+
"epoch": 0.1851550492722177,
|
1806 |
+
"grad_norm": 17.3125,
|
1807 |
+
"learning_rate": 9.985534818990295e-07,
|
1808 |
+
"loss": 25.1881,
|
1809 |
+
"step": 2560
|
1810 |
+
},
|
1811 |
+
{
|
1812 |
+
"epoch": 0.1858783111834373,
|
1813 |
+
"grad_norm": 16.53125,
|
1814 |
+
"learning_rate": 9.985478314376977e-07,
|
1815 |
+
"loss": 24.9183,
|
1816 |
+
"step": 2570
|
1817 |
+
},
|
1818 |
+
{
|
1819 |
+
"epoch": 0.18660157309465691,
|
1820 |
+
"grad_norm": 15.8359375,
|
1821 |
+
"learning_rate": 9.985421809763657e-07,
|
1822 |
+
"loss": 25.1384,
|
1823 |
+
"step": 2580
|
1824 |
+
},
|
1825 |
+
{
|
1826 |
+
"epoch": 0.18732483500587652,
|
1827 |
+
"grad_norm": 17.84375,
|
1828 |
+
"learning_rate": 9.98536530515034e-07,
|
1829 |
+
"loss": 24.6692,
|
1830 |
+
"step": 2590
|
1831 |
+
},
|
1832 |
+
{
|
1833 |
+
"epoch": 0.1880480969170961,
|
1834 |
+
"grad_norm": 16.8125,
|
1835 |
+
"learning_rate": 9.98530880053702e-07,
|
1836 |
+
"loss": 24.7989,
|
1837 |
+
"step": 2600
|
1838 |
+
},
|
1839 |
+
{
|
1840 |
+
"epoch": 0.1887713588283157,
|
1841 |
+
"grad_norm": 16.078125,
|
1842 |
+
"learning_rate": 9.985252295923701e-07,
|
1843 |
+
"loss": 24.9476,
|
1844 |
+
"step": 2610
|
1845 |
+
},
|
1846 |
+
{
|
1847 |
+
"epoch": 0.1894946207395353,
|
1848 |
+
"grad_norm": 15.9765625,
|
1849 |
+
"learning_rate": 9.985195791310381e-07,
|
1850 |
+
"loss": 25.065,
|
1851 |
+
"step": 2620
|
1852 |
+
},
|
1853 |
+
{
|
1854 |
+
"epoch": 0.1902178826507549,
|
1855 |
+
"grad_norm": 16.171875,
|
1856 |
+
"learning_rate": 9.985139286697061e-07,
|
1857 |
+
"loss": 24.9492,
|
1858 |
+
"step": 2630
|
1859 |
+
},
|
1860 |
+
{
|
1861 |
+
"epoch": 0.1909411445619745,
|
1862 |
+
"grad_norm": 16.5,
|
1863 |
+
"learning_rate": 9.985082782083741e-07,
|
1864 |
+
"loss": 25.2619,
|
1865 |
+
"step": 2640
|
1866 |
+
},
|
1867 |
+
{
|
1868 |
+
"epoch": 0.1916644064731941,
|
1869 |
+
"grad_norm": 16.03125,
|
1870 |
+
"learning_rate": 9.985026277470424e-07,
|
1871 |
+
"loss": 25.2593,
|
1872 |
+
"step": 2650
|
1873 |
+
},
|
1874 |
+
{
|
1875 |
+
"epoch": 0.1923876683844137,
|
1876 |
+
"grad_norm": 15.625,
|
1877 |
+
"learning_rate": 9.984969772857106e-07,
|
1878 |
+
"loss": 24.9816,
|
1879 |
+
"step": 2660
|
1880 |
+
},
|
1881 |
+
{
|
1882 |
+
"epoch": 0.1931109302956333,
|
1883 |
+
"grad_norm": 15.890625,
|
1884 |
+
"learning_rate": 9.984913268243786e-07,
|
1885 |
+
"loss": 25.0609,
|
1886 |
+
"step": 2670
|
1887 |
+
},
|
1888 |
+
{
|
1889 |
+
"epoch": 0.1938341922068529,
|
1890 |
+
"grad_norm": 16.015625,
|
1891 |
+
"learning_rate": 9.984856763630466e-07,
|
1892 |
+
"loss": 25.2305,
|
1893 |
+
"step": 2680
|
1894 |
+
},
|
1895 |
+
{
|
1896 |
+
"epoch": 0.19455745411807251,
|
1897 |
+
"grad_norm": 19.59375,
|
1898 |
+
"learning_rate": 9.984800259017146e-07,
|
1899 |
+
"loss": 24.9294,
|
1900 |
+
"step": 2690
|
1901 |
+
},
|
1902 |
+
{
|
1903 |
+
"epoch": 0.19528071602929212,
|
1904 |
+
"grad_norm": 17.140625,
|
1905 |
+
"learning_rate": 9.984743754403828e-07,
|
1906 |
+
"loss": 25.3927,
|
1907 |
+
"step": 2700
|
1908 |
+
},
|
1909 |
+
{
|
1910 |
+
"epoch": 0.19600397794051172,
|
1911 |
+
"grad_norm": 17.21875,
|
1912 |
+
"learning_rate": 9.98468724979051e-07,
|
1913 |
+
"loss": 25.1831,
|
1914 |
+
"step": 2710
|
1915 |
+
},
|
1916 |
+
{
|
1917 |
+
"epoch": 0.1967272398517313,
|
1918 |
+
"grad_norm": 16.4375,
|
1919 |
+
"learning_rate": 9.98463074517719e-07,
|
1920 |
+
"loss": 25.128,
|
1921 |
+
"step": 2720
|
1922 |
+
},
|
1923 |
+
{
|
1924 |
+
"epoch": 0.1974505017629509,
|
1925 |
+
"grad_norm": 16.953125,
|
1926 |
+
"learning_rate": 9.98457424056387e-07,
|
1927 |
+
"loss": 25.0455,
|
1928 |
+
"step": 2730
|
1929 |
+
},
|
1930 |
+
{
|
1931 |
+
"epoch": 0.1981737636741705,
|
1932 |
+
"grad_norm": 16.5,
|
1933 |
+
"learning_rate": 9.984517735950552e-07,
|
1934 |
+
"loss": 24.7651,
|
1935 |
+
"step": 2740
|
1936 |
+
},
|
1937 |
+
{
|
1938 |
+
"epoch": 0.1988970255853901,
|
1939 |
+
"grad_norm": 16.40625,
|
1940 |
+
"learning_rate": 9.984461231337232e-07,
|
1941 |
+
"loss": 25.3457,
|
1942 |
+
"step": 2750
|
1943 |
+
},
|
1944 |
+
{
|
1945 |
+
"epoch": 0.1996202874966097,
|
1946 |
+
"grad_norm": 16.171875,
|
1947 |
+
"learning_rate": 9.984404726723914e-07,
|
1948 |
+
"loss": 25.3497,
|
1949 |
+
"step": 2760
|
1950 |
+
},
|
1951 |
+
{
|
1952 |
+
"epoch": 0.2003435494078293,
|
1953 |
+
"grad_norm": 16.125,
|
1954 |
+
"learning_rate": 9.984348222110594e-07,
|
1955 |
+
"loss": 25.1117,
|
1956 |
+
"step": 2770
|
1957 |
+
},
|
1958 |
+
{
|
1959 |
+
"epoch": 0.2010668113190489,
|
1960 |
+
"grad_norm": 16.09375,
|
1961 |
+
"learning_rate": 9.984291717497276e-07,
|
1962 |
+
"loss": 25.2813,
|
1963 |
+
"step": 2780
|
1964 |
+
},
|
1965 |
+
{
|
1966 |
+
"epoch": 0.2017900732302685,
|
1967 |
+
"grad_norm": 15.5078125,
|
1968 |
+
"learning_rate": 9.984235212883956e-07,
|
1969 |
+
"loss": 25.4148,
|
1970 |
+
"step": 2790
|
1971 |
+
},
|
1972 |
+
{
|
1973 |
+
"epoch": 0.20251333514148812,
|
1974 |
+
"grad_norm": 16.125,
|
1975 |
+
"learning_rate": 9.984178708270636e-07,
|
1976 |
+
"loss": 25.2476,
|
1977 |
+
"step": 2800
|
1978 |
+
},
|
1979 |
+
{
|
1980 |
+
"epoch": 0.20323659705270772,
|
1981 |
+
"grad_norm": 16.375,
|
1982 |
+
"learning_rate": 9.984122203657318e-07,
|
1983 |
+
"loss": 25.1569,
|
1984 |
+
"step": 2810
|
1985 |
+
},
|
1986 |
+
{
|
1987 |
+
"epoch": 0.20395985896392732,
|
1988 |
+
"grad_norm": 15.3984375,
|
1989 |
+
"learning_rate": 9.984065699043998e-07,
|
1990 |
+
"loss": 24.9938,
|
1991 |
+
"step": 2820
|
1992 |
+
},
|
1993 |
+
{
|
1994 |
+
"epoch": 0.20468312087514692,
|
1995 |
+
"grad_norm": 15.8984375,
|
1996 |
+
"learning_rate": 9.98400919443068e-07,
|
1997 |
+
"loss": 25.0097,
|
1998 |
+
"step": 2830
|
1999 |
+
},
|
2000 |
+
{
|
2001 |
+
"epoch": 0.20540638278636653,
|
2002 |
+
"grad_norm": 16.640625,
|
2003 |
+
"learning_rate": 9.98395268981736e-07,
|
2004 |
+
"loss": 25.2562,
|
2005 |
+
"step": 2840
|
2006 |
+
},
|
2007 |
+
{
|
2008 |
+
"epoch": 0.2061296446975861,
|
2009 |
+
"grad_norm": 15.3828125,
|
2010 |
+
"learning_rate": 9.98389618520404e-07,
|
2011 |
+
"loss": 24.8865,
|
2012 |
+
"step": 2850
|
2013 |
+
},
|
2014 |
+
{
|
2015 |
+
"epoch": 0.2068529066088057,
|
2016 |
+
"grad_norm": 15.5703125,
|
2017 |
+
"learning_rate": 9.983839680590722e-07,
|
2018 |
+
"loss": 25.2523,
|
2019 |
+
"step": 2860
|
2020 |
+
},
|
2021 |
+
{
|
2022 |
+
"epoch": 0.2075761685200253,
|
2023 |
+
"grad_norm": 15.90625,
|
2024 |
+
"learning_rate": 9.983783175977402e-07,
|
2025 |
+
"loss": 25.2208,
|
2026 |
+
"step": 2870
|
2027 |
+
},
|
2028 |
+
{
|
2029 |
+
"epoch": 0.2082994304312449,
|
2030 |
+
"grad_norm": 15.53125,
|
2031 |
+
"learning_rate": 9.983726671364084e-07,
|
2032 |
+
"loss": 25.2277,
|
2033 |
+
"step": 2880
|
2034 |
+
},
|
2035 |
+
{
|
2036 |
+
"epoch": 0.2090226923424645,
|
2037 |
+
"grad_norm": 15.375,
|
2038 |
+
"learning_rate": 9.983670166750764e-07,
|
2039 |
+
"loss": 24.9516,
|
2040 |
+
"step": 2890
|
2041 |
+
},
|
2042 |
+
{
|
2043 |
+
"epoch": 0.2097459542536841,
|
2044 |
+
"grad_norm": 17.4375,
|
2045 |
+
"learning_rate": 9.983613662137444e-07,
|
2046 |
+
"loss": 25.5087,
|
2047 |
+
"step": 2900
|
2048 |
+
},
|
2049 |
+
{
|
2050 |
+
"epoch": 0.21046921616490372,
|
2051 |
+
"grad_norm": 16.28125,
|
2052 |
+
"learning_rate": 9.983557157524126e-07,
|
2053 |
+
"loss": 25.1188,
|
2054 |
+
"step": 2910
|
2055 |
+
},
|
2056 |
+
{
|
2057 |
+
"epoch": 0.21119247807612332,
|
2058 |
+
"grad_norm": 16.390625,
|
2059 |
+
"learning_rate": 9.983500652910806e-07,
|
2060 |
+
"loss": 25.2662,
|
2061 |
+
"step": 2920
|
2062 |
+
},
|
2063 |
+
{
|
2064 |
+
"epoch": 0.21191573998734292,
|
2065 |
+
"grad_norm": 16.03125,
|
2066 |
+
"learning_rate": 9.983444148297488e-07,
|
2067 |
+
"loss": 24.9432,
|
2068 |
+
"step": 2930
|
2069 |
+
},
|
2070 |
+
{
|
2071 |
+
"epoch": 0.21263900189856252,
|
2072 |
+
"grad_norm": 17.234375,
|
2073 |
+
"learning_rate": 9.983387643684168e-07,
|
2074 |
+
"loss": 25.2459,
|
2075 |
+
"step": 2940
|
2076 |
+
},
|
2077 |
+
{
|
2078 |
+
"epoch": 0.21336226380978213,
|
2079 |
+
"grad_norm": 15.28125,
|
2080 |
+
"learning_rate": 9.983331139070848e-07,
|
2081 |
+
"loss": 24.9326,
|
2082 |
+
"step": 2950
|
2083 |
+
},
|
2084 |
+
{
|
2085 |
+
"epoch": 0.21408552572100173,
|
2086 |
+
"grad_norm": 18.0,
|
2087 |
+
"learning_rate": 9.98327463445753e-07,
|
2088 |
+
"loss": 25.2388,
|
2089 |
+
"step": 2960
|
2090 |
+
},
|
2091 |
+
{
|
2092 |
+
"epoch": 0.21480878763222133,
|
2093 |
+
"grad_norm": 17.1875,
|
2094 |
+
"learning_rate": 9.98321812984421e-07,
|
2095 |
+
"loss": 25.6366,
|
2096 |
+
"step": 2970
|
2097 |
+
},
|
2098 |
+
{
|
2099 |
+
"epoch": 0.2155320495434409,
|
2100 |
+
"grad_norm": 16.203125,
|
2101 |
+
"learning_rate": 9.983161625230892e-07,
|
2102 |
+
"loss": 25.1225,
|
2103 |
+
"step": 2980
|
2104 |
+
},
|
2105 |
+
{
|
2106 |
+
"epoch": 0.2162553114546605,
|
2107 |
+
"grad_norm": 16.828125,
|
2108 |
+
"learning_rate": 9.983105120617572e-07,
|
2109 |
+
"loss": 25.1165,
|
2110 |
+
"step": 2990
|
2111 |
+
},
|
2112 |
+
{
|
2113 |
+
"epoch": 0.2169785733658801,
|
2114 |
+
"grad_norm": 16.625,
|
2115 |
+
"learning_rate": 9.983048616004254e-07,
|
2116 |
+
"loss": 25.2762,
|
2117 |
+
"step": 3000
|
2118 |
+
},
|
2119 |
+
{
|
2120 |
+
"epoch": 0.21770183527709971,
|
2121 |
+
"grad_norm": 16.1875,
|
2122 |
+
"learning_rate": 9.982992111390934e-07,
|
2123 |
+
"loss": 24.9826,
|
2124 |
+
"step": 3010
|
2125 |
+
},
|
2126 |
+
{
|
2127 |
+
"epoch": 0.21842509718831932,
|
2128 |
+
"grad_norm": 16.890625,
|
2129 |
+
"learning_rate": 9.982935606777614e-07,
|
2130 |
+
"loss": 24.7336,
|
2131 |
+
"step": 3020
|
2132 |
+
},
|
2133 |
+
{
|
2134 |
+
"epoch": 0.21914835909953892,
|
2135 |
+
"grad_norm": 15.765625,
|
2136 |
+
"learning_rate": 9.982879102164296e-07,
|
2137 |
+
"loss": 25.1037,
|
2138 |
+
"step": 3030
|
2139 |
+
},
|
2140 |
+
{
|
2141 |
+
"epoch": 0.21987162101075852,
|
2142 |
+
"grad_norm": 16.734375,
|
2143 |
+
"learning_rate": 9.982822597550976e-07,
|
2144 |
+
"loss": 24.8429,
|
2145 |
+
"step": 3040
|
2146 |
+
},
|
2147 |
+
{
|
2148 |
+
"epoch": 0.22059488292197812,
|
2149 |
+
"grad_norm": 15.96875,
|
2150 |
+
"learning_rate": 9.982766092937658e-07,
|
2151 |
+
"loss": 25.369,
|
2152 |
+
"step": 3050
|
2153 |
+
},
|
2154 |
+
{
|
2155 |
+
"epoch": 0.22131814483319773,
|
2156 |
+
"grad_norm": 15.46875,
|
2157 |
+
"learning_rate": 9.982709588324338e-07,
|
2158 |
+
"loss": 25.6081,
|
2159 |
+
"step": 3060
|
2160 |
+
},
|
2161 |
+
{
|
2162 |
+
"epoch": 0.22204140674441733,
|
2163 |
+
"grad_norm": 16.59375,
|
2164 |
+
"learning_rate": 9.982653083711018e-07,
|
2165 |
+
"loss": 25.4421,
|
2166 |
+
"step": 3070
|
2167 |
+
},
|
2168 |
+
{
|
2169 |
+
"epoch": 0.22276466865563693,
|
2170 |
+
"grad_norm": 15.734375,
|
2171 |
+
"learning_rate": 9.9825965790977e-07,
|
2172 |
+
"loss": 24.9267,
|
2173 |
+
"step": 3080
|
2174 |
+
},
|
2175 |
+
{
|
2176 |
+
"epoch": 0.22348793056685654,
|
2177 |
+
"grad_norm": 15.7421875,
|
2178 |
+
"learning_rate": 9.98254007448438e-07,
|
2179 |
+
"loss": 25.1702,
|
2180 |
+
"step": 3090
|
2181 |
+
},
|
2182 |
+
{
|
2183 |
+
"epoch": 0.2242111924780761,
|
2184 |
+
"grad_norm": 16.34375,
|
2185 |
+
"learning_rate": 9.982483569871063e-07,
|
2186 |
+
"loss": 24.97,
|
2187 |
+
"step": 3100
|
2188 |
+
},
|
2189 |
+
{
|
2190 |
+
"epoch": 0.2249344543892957,
|
2191 |
+
"grad_norm": 16.328125,
|
2192 |
+
"learning_rate": 9.982427065257743e-07,
|
2193 |
+
"loss": 25.1489,
|
2194 |
+
"step": 3110
|
2195 |
+
},
|
2196 |
+
{
|
2197 |
+
"epoch": 0.22565771630051532,
|
2198 |
+
"grad_norm": 16.46875,
|
2199 |
+
"learning_rate": 9.982370560644423e-07,
|
2200 |
+
"loss": 25.291,
|
2201 |
+
"step": 3120
|
2202 |
+
},
|
2203 |
+
{
|
2204 |
+
"epoch": 0.22638097821173492,
|
2205 |
+
"grad_norm": 16.859375,
|
2206 |
+
"learning_rate": 9.982314056031105e-07,
|
2207 |
+
"loss": 25.1147,
|
2208 |
+
"step": 3130
|
2209 |
+
},
|
2210 |
+
{
|
2211 |
+
"epoch": 0.22710424012295452,
|
2212 |
+
"grad_norm": 16.1875,
|
2213 |
+
"learning_rate": 9.982257551417785e-07,
|
2214 |
+
"loss": 25.0294,
|
2215 |
+
"step": 3140
|
2216 |
+
},
|
2217 |
+
{
|
2218 |
+
"epoch": 0.22782750203417412,
|
2219 |
+
"grad_norm": 16.609375,
|
2220 |
+
"learning_rate": 9.982201046804467e-07,
|
2221 |
+
"loss": 24.9066,
|
2222 |
+
"step": 3150
|
2223 |
+
},
|
2224 |
+
{
|
2225 |
+
"epoch": 0.22855076394539373,
|
2226 |
+
"grad_norm": 16.296875,
|
2227 |
+
"learning_rate": 9.982144542191147e-07,
|
2228 |
+
"loss": 25.5522,
|
2229 |
+
"step": 3160
|
2230 |
+
},
|
2231 |
+
{
|
2232 |
+
"epoch": 0.22927402585661333,
|
2233 |
+
"grad_norm": 16.875,
|
2234 |
+
"learning_rate": 9.982088037577827e-07,
|
2235 |
+
"loss": 25.046,
|
2236 |
+
"step": 3170
|
2237 |
+
},
|
2238 |
+
{
|
2239 |
+
"epoch": 0.22999728776783293,
|
2240 |
+
"grad_norm": 17.34375,
|
2241 |
+
"learning_rate": 9.982031532964509e-07,
|
2242 |
+
"loss": 25.0408,
|
2243 |
+
"step": 3180
|
2244 |
+
},
|
2245 |
+
{
|
2246 |
+
"epoch": 0.23072054967905253,
|
2247 |
+
"grad_norm": 16.078125,
|
2248 |
+
"learning_rate": 9.981975028351189e-07,
|
2249 |
+
"loss": 25.6239,
|
2250 |
+
"step": 3190
|
2251 |
+
},
|
2252 |
+
{
|
2253 |
+
"epoch": 0.23144381159027214,
|
2254 |
+
"grad_norm": 14.75,
|
2255 |
+
"learning_rate": 9.98191852373787e-07,
|
2256 |
+
"loss": 24.7332,
|
2257 |
+
"step": 3200
|
2258 |
+
},
|
2259 |
+
{
|
2260 |
+
"epoch": 0.23216707350149174,
|
2261 |
+
"grad_norm": 16.265625,
|
2262 |
+
"learning_rate": 9.98186201912455e-07,
|
2263 |
+
"loss": 25.0991,
|
2264 |
+
"step": 3210
|
2265 |
+
},
|
2266 |
+
{
|
2267 |
+
"epoch": 0.23289033541271134,
|
2268 |
+
"grad_norm": 16.609375,
|
2269 |
+
"learning_rate": 9.981805514511233e-07,
|
2270 |
+
"loss": 24.9936,
|
2271 |
+
"step": 3220
|
2272 |
+
},
|
2273 |
+
{
|
2274 |
+
"epoch": 0.23361359732393092,
|
2275 |
+
"grad_norm": 17.03125,
|
2276 |
+
"learning_rate": 9.981749009897913e-07,
|
2277 |
+
"loss": 25.4223,
|
2278 |
+
"step": 3230
|
2279 |
+
},
|
2280 |
+
{
|
2281 |
+
"epoch": 0.23433685923515052,
|
2282 |
+
"grad_norm": 16.625,
|
2283 |
+
"learning_rate": 9.981692505284593e-07,
|
2284 |
+
"loss": 25.1375,
|
2285 |
+
"step": 3240
|
2286 |
+
},
|
2287 |
+
{
|
2288 |
+
"epoch": 0.23506012114637012,
|
2289 |
+
"grad_norm": 16.71875,
|
2290 |
+
"learning_rate": 9.981636000671275e-07,
|
2291 |
+
"loss": 25.1016,
|
2292 |
+
"step": 3250
|
2293 |
+
},
|
2294 |
+
{
|
2295 |
+
"epoch": 0.23578338305758972,
|
2296 |
+
"grad_norm": 16.15625,
|
2297 |
+
"learning_rate": 9.981579496057955e-07,
|
2298 |
+
"loss": 24.9417,
|
2299 |
+
"step": 3260
|
2300 |
+
},
|
2301 |
+
{
|
2302 |
+
"epoch": 0.23650664496880933,
|
2303 |
+
"grad_norm": 16.84375,
|
2304 |
+
"learning_rate": 9.981522991444637e-07,
|
2305 |
+
"loss": 25.2421,
|
2306 |
+
"step": 3270
|
2307 |
+
},
|
2308 |
+
{
|
2309 |
+
"epoch": 0.23722990688002893,
|
2310 |
+
"grad_norm": 15.8359375,
|
2311 |
+
"learning_rate": 9.981466486831317e-07,
|
2312 |
+
"loss": 25.1771,
|
2313 |
+
"step": 3280
|
2314 |
+
},
|
2315 |
+
{
|
2316 |
+
"epoch": 0.23795316879124853,
|
2317 |
+
"grad_norm": 16.65625,
|
2318 |
+
"learning_rate": 9.981409982217997e-07,
|
2319 |
+
"loss": 25.3145,
|
2320 |
+
"step": 3290
|
2321 |
+
},
|
2322 |
+
{
|
2323 |
+
"epoch": 0.23867643070246813,
|
2324 |
+
"grad_norm": 16.984375,
|
2325 |
+
"learning_rate": 9.98135347760468e-07,
|
2326 |
+
"loss": 24.93,
|
2327 |
+
"step": 3300
|
2328 |
+
},
|
2329 |
+
{
|
2330 |
+
"epoch": 0.23939969261368774,
|
2331 |
+
"grad_norm": 16.6875,
|
2332 |
+
"learning_rate": 9.98129697299136e-07,
|
2333 |
+
"loss": 25.5047,
|
2334 |
+
"step": 3310
|
2335 |
+
},
|
2336 |
+
{
|
2337 |
+
"epoch": 0.24012295452490734,
|
2338 |
+
"grad_norm": 16.28125,
|
2339 |
+
"learning_rate": 9.98124046837804e-07,
|
2340 |
+
"loss": 25.0701,
|
2341 |
+
"step": 3320
|
2342 |
+
},
|
2343 |
+
{
|
2344 |
+
"epoch": 0.24084621643612694,
|
2345 |
+
"grad_norm": 15.265625,
|
2346 |
+
"learning_rate": 9.98118396376472e-07,
|
2347 |
+
"loss": 25.1109,
|
2348 |
+
"step": 3330
|
2349 |
+
},
|
2350 |
+
{
|
2351 |
+
"epoch": 0.24156947834734654,
|
2352 |
+
"grad_norm": 16.5625,
|
2353 |
+
"learning_rate": 9.9811274591514e-07,
|
2354 |
+
"loss": 25.2658,
|
2355 |
+
"step": 3340
|
2356 |
+
},
|
2357 |
+
{
|
2358 |
+
"epoch": 0.24229274025856612,
|
2359 |
+
"grad_norm": 16.140625,
|
2360 |
+
"learning_rate": 9.981070954538083e-07,
|
2361 |
+
"loss": 24.9092,
|
2362 |
+
"step": 3350
|
2363 |
+
},
|
2364 |
+
{
|
2365 |
+
"epoch": 0.24301600216978572,
|
2366 |
+
"grad_norm": 16.71875,
|
2367 |
+
"learning_rate": 9.981014449924763e-07,
|
2368 |
+
"loss": 25.0737,
|
2369 |
+
"step": 3360
|
2370 |
+
},
|
2371 |
+
{
|
2372 |
+
"epoch": 0.24373926408100532,
|
2373 |
+
"grad_norm": 17.578125,
|
2374 |
+
"learning_rate": 9.980957945311445e-07,
|
2375 |
+
"loss": 24.9847,
|
2376 |
+
"step": 3370
|
2377 |
+
},
|
2378 |
+
{
|
2379 |
+
"epoch": 0.24446252599222493,
|
2380 |
+
"grad_norm": 16.125,
|
2381 |
+
"learning_rate": 9.980901440698125e-07,
|
2382 |
+
"loss": 25.1988,
|
2383 |
+
"step": 3380
|
2384 |
+
},
|
2385 |
+
{
|
2386 |
+
"epoch": 0.24518578790344453,
|
2387 |
+
"grad_norm": 17.296875,
|
2388 |
+
"learning_rate": 9.980844936084805e-07,
|
2389 |
+
"loss": 25.0352,
|
2390 |
+
"step": 3390
|
2391 |
+
},
|
2392 |
+
{
|
2393 |
+
"epoch": 0.24590904981466413,
|
2394 |
+
"grad_norm": 15.0078125,
|
2395 |
+
"learning_rate": 9.980788431471487e-07,
|
2396 |
+
"loss": 25.0363,
|
2397 |
+
"step": 3400
|
2398 |
+
},
|
2399 |
+
{
|
2400 |
+
"epoch": 0.24663231172588373,
|
2401 |
+
"grad_norm": 17.265625,
|
2402 |
+
"learning_rate": 9.980731926858167e-07,
|
2403 |
+
"loss": 24.8165,
|
2404 |
+
"step": 3410
|
2405 |
+
},
|
2406 |
+
{
|
2407 |
+
"epoch": 0.24735557363710334,
|
2408 |
+
"grad_norm": 18.09375,
|
2409 |
+
"learning_rate": 9.98067542224485e-07,
|
2410 |
+
"loss": 25.1422,
|
2411 |
+
"step": 3420
|
2412 |
+
},
|
2413 |
+
{
|
2414 |
+
"epoch": 0.24807883554832294,
|
2415 |
+
"grad_norm": 16.203125,
|
2416 |
+
"learning_rate": 9.98061891763153e-07,
|
2417 |
+
"loss": 24.8054,
|
2418 |
+
"step": 3430
|
2419 |
+
},
|
2420 |
+
{
|
2421 |
+
"epoch": 0.24880209745954254,
|
2422 |
+
"grad_norm": 16.703125,
|
2423 |
+
"learning_rate": 9.980562413018211e-07,
|
2424 |
+
"loss": 25.1454,
|
2425 |
+
"step": 3440
|
2426 |
+
},
|
2427 |
+
{
|
2428 |
+
"epoch": 0.24952535937076215,
|
2429 |
+
"grad_norm": 15.359375,
|
2430 |
+
"learning_rate": 9.980505908404891e-07,
|
2431 |
+
"loss": 25.3514,
|
2432 |
+
"step": 3450
|
2433 |
+
},
|
2434 |
+
{
|
2435 |
+
"epoch": 0.2502486212819817,
|
2436 |
+
"grad_norm": 15.3984375,
|
2437 |
+
"learning_rate": 9.980449403791571e-07,
|
2438 |
+
"loss": 25.141,
|
2439 |
+
"step": 3460
|
2440 |
+
},
|
2441 |
+
{
|
2442 |
+
"epoch": 0.25097188319320135,
|
2443 |
+
"grad_norm": 16.078125,
|
2444 |
+
"learning_rate": 9.980392899178253e-07,
|
2445 |
+
"loss": 24.8799,
|
2446 |
+
"step": 3470
|
2447 |
+
},
|
2448 |
+
{
|
2449 |
+
"epoch": 0.2516951451044209,
|
2450 |
+
"grad_norm": 16.515625,
|
2451 |
+
"learning_rate": 9.980336394564933e-07,
|
2452 |
+
"loss": 24.9967,
|
2453 |
+
"step": 3480
|
2454 |
+
},
|
2455 |
+
{
|
2456 |
+
"epoch": 0.25241840701564056,
|
2457 |
+
"grad_norm": 16.46875,
|
2458 |
+
"learning_rate": 9.980279889951615e-07,
|
2459 |
+
"loss": 25.1666,
|
2460 |
+
"step": 3490
|
2461 |
+
},
|
2462 |
+
{
|
2463 |
+
"epoch": 0.25314166892686013,
|
2464 |
+
"grad_norm": 16.25,
|
2465 |
+
"learning_rate": 9.980223385338295e-07,
|
2466 |
+
"loss": 25.3913,
|
2467 |
+
"step": 3500
|
2468 |
+
},
|
2469 |
+
{
|
2470 |
+
"epoch": 0.25386493083807976,
|
2471 |
+
"grad_norm": 15.6640625,
|
2472 |
+
"learning_rate": 9.980166880724975e-07,
|
2473 |
+
"loss": 25.0079,
|
2474 |
+
"step": 3510
|
2475 |
+
},
|
2476 |
+
{
|
2477 |
+
"epoch": 0.25458819274929934,
|
2478 |
+
"grad_norm": 16.015625,
|
2479 |
+
"learning_rate": 9.980110376111657e-07,
|
2480 |
+
"loss": 25.0063,
|
2481 |
+
"step": 3520
|
2482 |
+
},
|
2483 |
+
{
|
2484 |
+
"epoch": 0.25531145466051897,
|
2485 |
+
"grad_norm": 15.8515625,
|
2486 |
+
"learning_rate": 9.980053871498337e-07,
|
2487 |
+
"loss": 25.5363,
|
2488 |
+
"step": 3530
|
2489 |
+
},
|
2490 |
+
{
|
2491 |
+
"epoch": 0.25603471657173854,
|
2492 |
+
"grad_norm": 16.984375,
|
2493 |
+
"learning_rate": 9.97999736688502e-07,
|
2494 |
+
"loss": 25.1561,
|
2495 |
+
"step": 3540
|
2496 |
+
},
|
2497 |
+
{
|
2498 |
+
"epoch": 0.2567579784829581,
|
2499 |
+
"grad_norm": 16.640625,
|
2500 |
+
"learning_rate": 9.9799408622717e-07,
|
2501 |
+
"loss": 25.1949,
|
2502 |
+
"step": 3550
|
2503 |
+
},
|
2504 |
+
{
|
2505 |
+
"epoch": 0.25748124039417775,
|
2506 |
+
"grad_norm": 16.0,
|
2507 |
+
"learning_rate": 9.97988435765838e-07,
|
2508 |
+
"loss": 24.992,
|
2509 |
+
"step": 3560
|
2510 |
+
},
|
2511 |
+
{
|
2512 |
+
"epoch": 0.2582045023053973,
|
2513 |
+
"grad_norm": 16.375,
|
2514 |
+
"learning_rate": 9.979827853045062e-07,
|
2515 |
+
"loss": 25.1319,
|
2516 |
+
"step": 3570
|
2517 |
+
},
|
2518 |
+
{
|
2519 |
+
"epoch": 0.25892776421661695,
|
2520 |
+
"grad_norm": 15.9296875,
|
2521 |
+
"learning_rate": 9.979771348431742e-07,
|
2522 |
+
"loss": 25.0064,
|
2523 |
+
"step": 3580
|
2524 |
+
},
|
2525 |
+
{
|
2526 |
+
"epoch": 0.2596510261278365,
|
2527 |
+
"grad_norm": 16.265625,
|
2528 |
+
"learning_rate": 9.979714843818424e-07,
|
2529 |
+
"loss": 24.9227,
|
2530 |
+
"step": 3590
|
2531 |
+
},
|
2532 |
+
{
|
2533 |
+
"epoch": 0.26037428803905616,
|
2534 |
+
"grad_norm": 16.578125,
|
2535 |
+
"learning_rate": 9.979658339205104e-07,
|
2536 |
+
"loss": 25.1148,
|
2537 |
+
"step": 3600
|
2538 |
+
},
|
2539 |
+
{
|
2540 |
+
"epoch": 0.26109754995027573,
|
2541 |
+
"grad_norm": 18.4375,
|
2542 |
+
"learning_rate": 9.979601834591784e-07,
|
2543 |
+
"loss": 24.9109,
|
2544 |
+
"step": 3610
|
2545 |
+
},
|
2546 |
+
{
|
2547 |
+
"epoch": 0.26182081186149536,
|
2548 |
+
"grad_norm": 17.640625,
|
2549 |
+
"learning_rate": 9.979545329978466e-07,
|
2550 |
+
"loss": 25.1381,
|
2551 |
+
"step": 3620
|
2552 |
+
},
|
2553 |
+
{
|
2554 |
+
"epoch": 0.26254407377271494,
|
2555 |
+
"grad_norm": 15.78125,
|
2556 |
+
"learning_rate": 9.979488825365146e-07,
|
2557 |
+
"loss": 25.0138,
|
2558 |
+
"step": 3630
|
2559 |
+
},
|
2560 |
+
{
|
2561 |
+
"epoch": 0.26326733568393457,
|
2562 |
+
"grad_norm": 16.65625,
|
2563 |
+
"learning_rate": 9.979432320751828e-07,
|
2564 |
+
"loss": 25.0708,
|
2565 |
+
"step": 3640
|
2566 |
+
},
|
2567 |
+
{
|
2568 |
+
"epoch": 0.26399059759515414,
|
2569 |
+
"grad_norm": 15.875,
|
2570 |
+
"learning_rate": 9.979375816138508e-07,
|
2571 |
+
"loss": 25.0401,
|
2572 |
+
"step": 3650
|
2573 |
+
},
|
2574 |
+
{
|
2575 |
+
"epoch": 0.26471385950637377,
|
2576 |
+
"grad_norm": 16.359375,
|
2577 |
+
"learning_rate": 9.97931931152519e-07,
|
2578 |
+
"loss": 25.1882,
|
2579 |
+
"step": 3660
|
2580 |
+
},
|
2581 |
+
{
|
2582 |
+
"epoch": 0.26543712141759335,
|
2583 |
+
"grad_norm": 15.5625,
|
2584 |
+
"learning_rate": 9.97926280691187e-07,
|
2585 |
+
"loss": 25.2344,
|
2586 |
+
"step": 3670
|
2587 |
+
},
|
2588 |
+
{
|
2589 |
+
"epoch": 0.2661603833288129,
|
2590 |
+
"grad_norm": 15.921875,
|
2591 |
+
"learning_rate": 9.97920630229855e-07,
|
2592 |
+
"loss": 25.0361,
|
2593 |
+
"step": 3680
|
2594 |
+
},
|
2595 |
+
{
|
2596 |
+
"epoch": 0.26688364524003255,
|
2597 |
+
"grad_norm": 15.7890625,
|
2598 |
+
"learning_rate": 9.979149797685232e-07,
|
2599 |
+
"loss": 25.2095,
|
2600 |
+
"step": 3690
|
2601 |
+
},
|
2602 |
+
{
|
2603 |
+
"epoch": 0.2676069071512521,
|
2604 |
+
"grad_norm": 16.25,
|
2605 |
+
"learning_rate": 9.979093293071914e-07,
|
2606 |
+
"loss": 25.4166,
|
2607 |
+
"step": 3700
|
2608 |
+
},
|
2609 |
+
{
|
2610 |
+
"epoch": 0.26833016906247176,
|
2611 |
+
"grad_norm": 17.03125,
|
2612 |
+
"learning_rate": 9.979036788458594e-07,
|
2613 |
+
"loss": 25.1349,
|
2614 |
+
"step": 3710
|
2615 |
+
},
|
2616 |
+
{
|
2617 |
+
"epoch": 0.26905343097369133,
|
2618 |
+
"grad_norm": 15.703125,
|
2619 |
+
"learning_rate": 9.978980283845274e-07,
|
2620 |
+
"loss": 24.6449,
|
2621 |
+
"step": 3720
|
2622 |
+
},
|
2623 |
+
{
|
2624 |
+
"epoch": 0.26977669288491096,
|
2625 |
+
"grad_norm": 15.625,
|
2626 |
+
"learning_rate": 9.978923779231954e-07,
|
2627 |
+
"loss": 25.4002,
|
2628 |
+
"step": 3730
|
2629 |
+
},
|
2630 |
+
{
|
2631 |
+
"epoch": 0.27049995479613054,
|
2632 |
+
"grad_norm": 15.8671875,
|
2633 |
+
"learning_rate": 9.978867274618636e-07,
|
2634 |
+
"loss": 24.9894,
|
2635 |
+
"step": 3740
|
2636 |
+
},
|
2637 |
+
{
|
2638 |
+
"epoch": 0.27122321670735017,
|
2639 |
+
"grad_norm": 17.234375,
|
2640 |
+
"learning_rate": 9.978810770005318e-07,
|
2641 |
+
"loss": 25.3439,
|
2642 |
+
"step": 3750
|
2643 |
+
},
|
2644 |
+
{
|
2645 |
+
"epoch": 0.27194647861856974,
|
2646 |
+
"grad_norm": 16.171875,
|
2647 |
+
"learning_rate": 9.978754265391998e-07,
|
2648 |
+
"loss": 25.1633,
|
2649 |
+
"step": 3760
|
2650 |
+
},
|
2651 |
+
{
|
2652 |
+
"epoch": 0.2726697405297894,
|
2653 |
+
"grad_norm": 16.875,
|
2654 |
+
"learning_rate": 9.978697760778678e-07,
|
2655 |
+
"loss": 25.0982,
|
2656 |
+
"step": 3770
|
2657 |
+
},
|
2658 |
+
{
|
2659 |
+
"epoch": 0.27339300244100895,
|
2660 |
+
"grad_norm": 16.71875,
|
2661 |
+
"learning_rate": 9.978641256165358e-07,
|
2662 |
+
"loss": 25.2635,
|
2663 |
+
"step": 3780
|
2664 |
+
},
|
2665 |
+
{
|
2666 |
+
"epoch": 0.2741162643522286,
|
2667 |
+
"grad_norm": 16.59375,
|
2668 |
+
"learning_rate": 9.97858475155204e-07,
|
2669 |
+
"loss": 25.4958,
|
2670 |
+
"step": 3790
|
2671 |
+
},
|
2672 |
+
{
|
2673 |
+
"epoch": 0.27483952626344815,
|
2674 |
+
"grad_norm": 15.9921875,
|
2675 |
+
"learning_rate": 9.97852824693872e-07,
|
2676 |
+
"loss": 24.7305,
|
2677 |
+
"step": 3800
|
2678 |
+
},
|
2679 |
+
{
|
2680 |
+
"epoch": 0.2755627881746677,
|
2681 |
+
"grad_norm": 16.46875,
|
2682 |
+
"learning_rate": 9.978471742325402e-07,
|
2683 |
+
"loss": 25.187,
|
2684 |
+
"step": 3810
|
2685 |
+
},
|
2686 |
+
{
|
2687 |
+
"epoch": 0.27628605008588736,
|
2688 |
+
"grad_norm": 15.8984375,
|
2689 |
+
"learning_rate": 9.978415237712082e-07,
|
2690 |
+
"loss": 25.1958,
|
2691 |
+
"step": 3820
|
2692 |
+
},
|
2693 |
+
{
|
2694 |
+
"epoch": 0.27700931199710693,
|
2695 |
+
"grad_norm": 16.109375,
|
2696 |
+
"learning_rate": 9.978358733098762e-07,
|
2697 |
+
"loss": 25.5358,
|
2698 |
+
"step": 3830
|
2699 |
+
},
|
2700 |
+
{
|
2701 |
+
"epoch": 0.27773257390832656,
|
2702 |
+
"grad_norm": 15.5,
|
2703 |
+
"learning_rate": 9.978302228485444e-07,
|
2704 |
+
"loss": 24.9236,
|
2705 |
+
"step": 3840
|
2706 |
+
},
|
2707 |
+
{
|
2708 |
+
"epoch": 0.27845583581954614,
|
2709 |
+
"grad_norm": 16.375,
|
2710 |
+
"learning_rate": 9.978245723872124e-07,
|
2711 |
+
"loss": 25.3469,
|
2712 |
+
"step": 3850
|
2713 |
+
},
|
2714 |
+
{
|
2715 |
+
"epoch": 0.27917909773076577,
|
2716 |
+
"grad_norm": 16.15625,
|
2717 |
+
"learning_rate": 9.978189219258806e-07,
|
2718 |
+
"loss": 24.9141,
|
2719 |
+
"step": 3860
|
2720 |
+
},
|
2721 |
+
{
|
2722 |
+
"epoch": 0.27990235964198534,
|
2723 |
+
"grad_norm": 17.125,
|
2724 |
+
"learning_rate": 9.978132714645486e-07,
|
2725 |
+
"loss": 25.0336,
|
2726 |
+
"step": 3870
|
2727 |
+
},
|
2728 |
+
{
|
2729 |
+
"epoch": 0.280625621553205,
|
2730 |
+
"grad_norm": 15.640625,
|
2731 |
+
"learning_rate": 9.978076210032168e-07,
|
2732 |
+
"loss": 25.2943,
|
2733 |
+
"step": 3880
|
2734 |
+
},
|
2735 |
+
{
|
2736 |
+
"epoch": 0.28134888346442455,
|
2737 |
+
"grad_norm": 15.8984375,
|
2738 |
+
"learning_rate": 9.978019705418848e-07,
|
2739 |
+
"loss": 24.9711,
|
2740 |
+
"step": 3890
|
2741 |
+
},
|
2742 |
+
{
|
2743 |
+
"epoch": 0.2820721453756442,
|
2744 |
+
"grad_norm": 16.09375,
|
2745 |
+
"learning_rate": 9.977963200805528e-07,
|
2746 |
+
"loss": 25.2311,
|
2747 |
+
"step": 3900
|
2748 |
+
},
|
2749 |
+
{
|
2750 |
+
"epoch": 0.28279540728686375,
|
2751 |
+
"grad_norm": 16.578125,
|
2752 |
+
"learning_rate": 9.97790669619221e-07,
|
2753 |
+
"loss": 25.4184,
|
2754 |
+
"step": 3910
|
2755 |
+
},
|
2756 |
+
{
|
2757 |
+
"epoch": 0.28351866919808333,
|
2758 |
+
"grad_norm": 15.890625,
|
2759 |
+
"learning_rate": 9.977850191578892e-07,
|
2760 |
+
"loss": 25.2003,
|
2761 |
+
"step": 3920
|
2762 |
+
},
|
2763 |
+
{
|
2764 |
+
"epoch": 0.28424193110930296,
|
2765 |
+
"grad_norm": 17.75,
|
2766 |
+
"learning_rate": 9.977793686965572e-07,
|
2767 |
+
"loss": 25.3082,
|
2768 |
+
"step": 3930
|
2769 |
+
},
|
2770 |
+
{
|
2771 |
+
"epoch": 0.28496519302052253,
|
2772 |
+
"grad_norm": 16.03125,
|
2773 |
+
"learning_rate": 9.977737182352252e-07,
|
2774 |
+
"loss": 25.265,
|
2775 |
+
"step": 3940
|
2776 |
+
},
|
2777 |
+
{
|
2778 |
+
"epoch": 0.28568845493174216,
|
2779 |
+
"grad_norm": 16.65625,
|
2780 |
+
"learning_rate": 9.977680677738932e-07,
|
2781 |
+
"loss": 25.6086,
|
2782 |
+
"step": 3950
|
2783 |
+
},
|
2784 |
+
{
|
2785 |
+
"epoch": 0.28641171684296174,
|
2786 |
+
"grad_norm": 15.8671875,
|
2787 |
+
"learning_rate": 9.977624173125614e-07,
|
2788 |
+
"loss": 25.6748,
|
2789 |
+
"step": 3960
|
2790 |
+
},
|
2791 |
+
{
|
2792 |
+
"epoch": 0.28713497875418137,
|
2793 |
+
"grad_norm": 16.5625,
|
2794 |
+
"learning_rate": 9.977567668512297e-07,
|
2795 |
+
"loss": 25.3439,
|
2796 |
+
"step": 3970
|
2797 |
+
},
|
2798 |
+
{
|
2799 |
+
"epoch": 0.28785824066540094,
|
2800 |
+
"grad_norm": 16.78125,
|
2801 |
+
"learning_rate": 9.977511163898977e-07,
|
2802 |
+
"loss": 24.9551,
|
2803 |
+
"step": 3980
|
2804 |
+
},
|
2805 |
+
{
|
2806 |
+
"epoch": 0.2885815025766206,
|
2807 |
+
"grad_norm": 16.546875,
|
2808 |
+
"learning_rate": 9.977454659285656e-07,
|
2809 |
+
"loss": 25.1372,
|
2810 |
+
"step": 3990
|
2811 |
+
},
|
2812 |
+
{
|
2813 |
+
"epoch": 0.28930476448784015,
|
2814 |
+
"grad_norm": 15.8515625,
|
2815 |
+
"learning_rate": 9.977398154672336e-07,
|
2816 |
+
"loss": 25.5988,
|
2817 |
+
"step": 4000
|
2818 |
+
},
|
2819 |
+
{
|
2820 |
+
"epoch": 0.2900280263990598,
|
2821 |
+
"grad_norm": 17.25,
|
2822 |
+
"learning_rate": 9.977341650059019e-07,
|
2823 |
+
"loss": 25.5525,
|
2824 |
+
"step": 4010
|
2825 |
+
},
|
2826 |
+
{
|
2827 |
+
"epoch": 0.29075128831027935,
|
2828 |
+
"grad_norm": 16.734375,
|
2829 |
+
"learning_rate": 9.9772851454457e-07,
|
2830 |
+
"loss": 24.7127,
|
2831 |
+
"step": 4020
|
2832 |
+
},
|
2833 |
+
{
|
2834 |
+
"epoch": 0.291474550221499,
|
2835 |
+
"grad_norm": 15.8125,
|
2836 |
+
"learning_rate": 9.97722864083238e-07,
|
2837 |
+
"loss": 24.8364,
|
2838 |
+
"step": 4030
|
2839 |
+
},
|
2840 |
+
{
|
2841 |
+
"epoch": 0.29219781213271856,
|
2842 |
+
"grad_norm": 17.75,
|
2843 |
+
"learning_rate": 9.97717213621906e-07,
|
2844 |
+
"loss": 25.1328,
|
2845 |
+
"step": 4040
|
2846 |
+
},
|
2847 |
+
{
|
2848 |
+
"epoch": 0.29292107404393813,
|
2849 |
+
"grad_norm": 15.7890625,
|
2850 |
+
"learning_rate": 9.977115631605743e-07,
|
2851 |
+
"loss": 24.932,
|
2852 |
+
"step": 4050
|
2853 |
+
},
|
2854 |
+
{
|
2855 |
+
"epoch": 0.29364433595515776,
|
2856 |
+
"grad_norm": 16.71875,
|
2857 |
+
"learning_rate": 9.977059126992423e-07,
|
2858 |
+
"loss": 25.0942,
|
2859 |
+
"step": 4060
|
2860 |
+
},
|
2861 |
+
{
|
2862 |
+
"epoch": 0.29436759786637734,
|
2863 |
+
"grad_norm": 17.0625,
|
2864 |
+
"learning_rate": 9.977002622379105e-07,
|
2865 |
+
"loss": 25.0096,
|
2866 |
+
"step": 4070
|
2867 |
+
},
|
2868 |
+
{
|
2869 |
+
"epoch": 0.29509085977759697,
|
2870 |
+
"grad_norm": 15.59375,
|
2871 |
+
"learning_rate": 9.976946117765785e-07,
|
2872 |
+
"loss": 24.9936,
|
2873 |
+
"step": 4080
|
2874 |
+
},
|
2875 |
+
{
|
2876 |
+
"epoch": 0.29581412168881654,
|
2877 |
+
"grad_norm": 16.890625,
|
2878 |
+
"learning_rate": 9.976889613152465e-07,
|
2879 |
+
"loss": 25.1561,
|
2880 |
+
"step": 4090
|
2881 |
+
},
|
2882 |
+
{
|
2883 |
+
"epoch": 0.2965373836000362,
|
2884 |
+
"grad_norm": 16.5625,
|
2885 |
+
"learning_rate": 9.976833108539147e-07,
|
2886 |
+
"loss": 25.2981,
|
2887 |
+
"step": 4100
|
2888 |
+
},
|
2889 |
+
{
|
2890 |
+
"epoch": 0.29726064551125575,
|
2891 |
+
"grad_norm": 16.03125,
|
2892 |
+
"learning_rate": 9.976776603925827e-07,
|
2893 |
+
"loss": 24.9605,
|
2894 |
+
"step": 4110
|
2895 |
+
},
|
2896 |
+
{
|
2897 |
+
"epoch": 0.2979839074224754,
|
2898 |
+
"grad_norm": 16.09375,
|
2899 |
+
"learning_rate": 9.976720099312507e-07,
|
2900 |
+
"loss": 25.254,
|
2901 |
+
"step": 4120
|
2902 |
+
},
|
2903 |
+
{
|
2904 |
+
"epoch": 0.29870716933369496,
|
2905 |
+
"grad_norm": 15.390625,
|
2906 |
+
"learning_rate": 9.976663594699189e-07,
|
2907 |
+
"loss": 25.2773,
|
2908 |
+
"step": 4130
|
2909 |
+
},
|
2910 |
+
{
|
2911 |
+
"epoch": 0.2994304312449146,
|
2912 |
+
"grad_norm": 16.96875,
|
2913 |
+
"learning_rate": 9.97660709008587e-07,
|
2914 |
+
"loss": 25.3654,
|
2915 |
+
"step": 4140
|
2916 |
+
},
|
2917 |
+
{
|
2918 |
+
"epoch": 0.30015369315613416,
|
2919 |
+
"grad_norm": 16.328125,
|
2920 |
+
"learning_rate": 9.97655058547255e-07,
|
2921 |
+
"loss": 24.5768,
|
2922 |
+
"step": 4150
|
2923 |
+
},
|
2924 |
+
{
|
2925 |
+
"epoch": 0.3008769550673538,
|
2926 |
+
"grad_norm": 16.640625,
|
2927 |
+
"learning_rate": 9.97649408085923e-07,
|
2928 |
+
"loss": 25.0535,
|
2929 |
+
"step": 4160
|
2930 |
+
},
|
2931 |
+
{
|
2932 |
+
"epoch": 0.30160021697857337,
|
2933 |
+
"grad_norm": 15.484375,
|
2934 |
+
"learning_rate": 9.97643757624591e-07,
|
2935 |
+
"loss": 25.2813,
|
2936 |
+
"step": 4170
|
2937 |
+
},
|
2938 |
+
{
|
2939 |
+
"epoch": 0.30232347888979294,
|
2940 |
+
"grad_norm": 16.734375,
|
2941 |
+
"learning_rate": 9.976381071632593e-07,
|
2942 |
+
"loss": 25.172,
|
2943 |
+
"step": 4180
|
2944 |
+
},
|
2945 |
+
{
|
2946 |
+
"epoch": 0.30304674080101257,
|
2947 |
+
"grad_norm": 16.140625,
|
2948 |
+
"learning_rate": 9.976324567019275e-07,
|
2949 |
+
"loss": 25.2808,
|
2950 |
+
"step": 4190
|
2951 |
+
},
|
2952 |
+
{
|
2953 |
+
"epoch": 0.30377000271223215,
|
2954 |
+
"grad_norm": 14.953125,
|
2955 |
+
"learning_rate": 9.976268062405955e-07,
|
2956 |
+
"loss": 25.1555,
|
2957 |
+
"step": 4200
|
2958 |
+
},
|
2959 |
+
{
|
2960 |
+
"epoch": 0.3044932646234518,
|
2961 |
+
"grad_norm": 16.0,
|
2962 |
+
"learning_rate": 9.976211557792635e-07,
|
2963 |
+
"loss": 25.4662,
|
2964 |
+
"step": 4210
|
2965 |
+
},
|
2966 |
+
{
|
2967 |
+
"epoch": 0.30521652653467135,
|
2968 |
+
"grad_norm": 16.640625,
|
2969 |
+
"learning_rate": 9.976155053179315e-07,
|
2970 |
+
"loss": 25.3153,
|
2971 |
+
"step": 4220
|
2972 |
+
},
|
2973 |
+
{
|
2974 |
+
"epoch": 0.305939788445891,
|
2975 |
+
"grad_norm": 15.9609375,
|
2976 |
+
"learning_rate": 9.976098548565997e-07,
|
2977 |
+
"loss": 25.1885,
|
2978 |
+
"step": 4230
|
2979 |
+
},
|
2980 |
+
{
|
2981 |
+
"epoch": 0.30666305035711056,
|
2982 |
+
"grad_norm": 15.90625,
|
2983 |
+
"learning_rate": 9.97604204395268e-07,
|
2984 |
+
"loss": 24.9472,
|
2985 |
+
"step": 4240
|
2986 |
+
},
|
2987 |
+
{
|
2988 |
+
"epoch": 0.3073863122683302,
|
2989 |
+
"grad_norm": 15.515625,
|
2990 |
+
"learning_rate": 9.97598553933936e-07,
|
2991 |
+
"loss": 25.0862,
|
2992 |
+
"step": 4250
|
2993 |
+
},
|
2994 |
+
{
|
2995 |
+
"epoch": 0.30810957417954976,
|
2996 |
+
"grad_norm": 16.421875,
|
2997 |
+
"learning_rate": 9.97592903472604e-07,
|
2998 |
+
"loss": 24.8722,
|
2999 |
+
"step": 4260
|
3000 |
+
},
|
3001 |
+
{
|
3002 |
+
"epoch": 0.3088328360907694,
|
3003 |
+
"grad_norm": 16.15625,
|
3004 |
+
"learning_rate": 9.975872530112721e-07,
|
3005 |
+
"loss": 25.3074,
|
3006 |
+
"step": 4270
|
3007 |
+
},
|
3008 |
+
{
|
3009 |
+
"epoch": 0.30955609800198897,
|
3010 |
+
"grad_norm": 15.7109375,
|
3011 |
+
"learning_rate": 9.975816025499401e-07,
|
3012 |
+
"loss": 25.0129,
|
3013 |
+
"step": 4280
|
3014 |
+
},
|
3015 |
+
{
|
3016 |
+
"epoch": 0.3102793599132086,
|
3017 |
+
"grad_norm": 15.859375,
|
3018 |
+
"learning_rate": 9.975759520886083e-07,
|
3019 |
+
"loss": 25.239,
|
3020 |
+
"step": 4290
|
3021 |
+
},
|
3022 |
+
{
|
3023 |
+
"epoch": 0.31100262182442817,
|
3024 |
+
"grad_norm": 15.859375,
|
3025 |
+
"learning_rate": 9.975703016272763e-07,
|
3026 |
+
"loss": 24.9693,
|
3027 |
+
"step": 4300
|
3028 |
+
},
|
3029 |
+
{
|
3030 |
+
"epoch": 0.31172588373564775,
|
3031 |
+
"grad_norm": 16.0625,
|
3032 |
+
"learning_rate": 9.975646511659443e-07,
|
3033 |
+
"loss": 25.1788,
|
3034 |
+
"step": 4310
|
3035 |
+
},
|
3036 |
+
{
|
3037 |
+
"epoch": 0.3124491456468674,
|
3038 |
+
"grad_norm": 16.1875,
|
3039 |
+
"learning_rate": 9.975590007046125e-07,
|
3040 |
+
"loss": 25.3888,
|
3041 |
+
"step": 4320
|
3042 |
+
},
|
3043 |
+
{
|
3044 |
+
"epoch": 0.31317240755808695,
|
3045 |
+
"grad_norm": 16.359375,
|
3046 |
+
"learning_rate": 9.975533502432805e-07,
|
3047 |
+
"loss": 25.269,
|
3048 |
+
"step": 4330
|
3049 |
+
},
|
3050 |
+
{
|
3051 |
+
"epoch": 0.3138956694693066,
|
3052 |
+
"grad_norm": 16.796875,
|
3053 |
+
"learning_rate": 9.975476997819487e-07,
|
3054 |
+
"loss": 25.0877,
|
3055 |
+
"step": 4340
|
3056 |
+
},
|
3057 |
+
{
|
3058 |
+
"epoch": 0.31461893138052616,
|
3059 |
+
"grad_norm": 16.453125,
|
3060 |
+
"learning_rate": 9.975420493206167e-07,
|
3061 |
+
"loss": 24.8608,
|
3062 |
+
"step": 4350
|
3063 |
+
},
|
3064 |
+
{
|
3065 |
+
"epoch": 0.3153421932917458,
|
3066 |
+
"grad_norm": 16.71875,
|
3067 |
+
"learning_rate": 9.97536398859285e-07,
|
3068 |
+
"loss": 25.3031,
|
3069 |
+
"step": 4360
|
3070 |
+
},
|
3071 |
+
{
|
3072 |
+
"epoch": 0.31606545520296536,
|
3073 |
+
"grad_norm": 16.03125,
|
3074 |
+
"learning_rate": 9.97530748397953e-07,
|
3075 |
+
"loss": 25.1261,
|
3076 |
+
"step": 4370
|
3077 |
+
},
|
3078 |
+
{
|
3079 |
+
"epoch": 0.316788717114185,
|
3080 |
+
"grad_norm": 16.25,
|
3081 |
+
"learning_rate": 9.97525097936621e-07,
|
3082 |
+
"loss": 25.1505,
|
3083 |
+
"step": 4380
|
3084 |
+
},
|
3085 |
+
{
|
3086 |
+
"epoch": 0.31751197902540457,
|
3087 |
+
"grad_norm": 15.640625,
|
3088 |
+
"learning_rate": 9.975194474752891e-07,
|
3089 |
+
"loss": 25.0906,
|
3090 |
+
"step": 4390
|
3091 |
+
},
|
3092 |
+
{
|
3093 |
+
"epoch": 0.3182352409366242,
|
3094 |
+
"grad_norm": 16.109375,
|
3095 |
+
"learning_rate": 9.975137970139571e-07,
|
3096 |
+
"loss": 25.1493,
|
3097 |
+
"step": 4400
|
3098 |
+
},
|
3099 |
+
{
|
3100 |
+
"epoch": 0.3189585028478438,
|
3101 |
+
"grad_norm": 16.25,
|
3102 |
+
"learning_rate": 9.975081465526254e-07,
|
3103 |
+
"loss": 25.2401,
|
3104 |
+
"step": 4410
|
3105 |
+
},
|
3106 |
+
{
|
3107 |
+
"epoch": 0.3196817647590634,
|
3108 |
+
"grad_norm": 16.921875,
|
3109 |
+
"learning_rate": 9.975024960912934e-07,
|
3110 |
+
"loss": 25.1871,
|
3111 |
+
"step": 4420
|
3112 |
+
},
|
3113 |
+
{
|
3114 |
+
"epoch": 0.320405026670283,
|
3115 |
+
"grad_norm": 16.328125,
|
3116 |
+
"learning_rate": 9.974968456299613e-07,
|
3117 |
+
"loss": 25.03,
|
3118 |
+
"step": 4430
|
3119 |
+
},
|
3120 |
+
{
|
3121 |
+
"epoch": 0.32112828858150255,
|
3122 |
+
"grad_norm": 16.546875,
|
3123 |
+
"learning_rate": 9.974911951686296e-07,
|
3124 |
+
"loss": 25.1426,
|
3125 |
+
"step": 4440
|
3126 |
+
},
|
3127 |
+
{
|
3128 |
+
"epoch": 0.3218515504927222,
|
3129 |
+
"grad_norm": 16.328125,
|
3130 |
+
"learning_rate": 9.974855447072976e-07,
|
3131 |
+
"loss": 25.0376,
|
3132 |
+
"step": 4450
|
3133 |
+
},
|
3134 |
+
{
|
3135 |
+
"epoch": 0.32257481240394176,
|
3136 |
+
"grad_norm": 15.9375,
|
3137 |
+
"learning_rate": 9.974798942459658e-07,
|
3138 |
+
"loss": 24.6728,
|
3139 |
+
"step": 4460
|
3140 |
+
},
|
3141 |
+
{
|
3142 |
+
"epoch": 0.3232980743151614,
|
3143 |
+
"grad_norm": 16.546875,
|
3144 |
+
"learning_rate": 9.974742437846338e-07,
|
3145 |
+
"loss": 25.2818,
|
3146 |
+
"step": 4470
|
3147 |
+
},
|
3148 |
+
{
|
3149 |
+
"epoch": 0.32402133622638096,
|
3150 |
+
"grad_norm": 15.46875,
|
3151 |
+
"learning_rate": 9.974685933233018e-07,
|
3152 |
+
"loss": 25.5436,
|
3153 |
+
"step": 4480
|
3154 |
+
},
|
3155 |
+
{
|
3156 |
+
"epoch": 0.3247445981376006,
|
3157 |
+
"grad_norm": 16.703125,
|
3158 |
+
"learning_rate": 9.9746294286197e-07,
|
3159 |
+
"loss": 25.0338,
|
3160 |
+
"step": 4490
|
3161 |
+
},
|
3162 |
+
{
|
3163 |
+
"epoch": 0.32546786004882017,
|
3164 |
+
"grad_norm": 16.546875,
|
3165 |
+
"learning_rate": 9.97457292400638e-07,
|
3166 |
+
"loss": 24.9645,
|
3167 |
+
"step": 4500
|
3168 |
+
},
|
3169 |
+
{
|
3170 |
+
"epoch": 0.3261911219600398,
|
3171 |
+
"grad_norm": 15.9296875,
|
3172 |
+
"learning_rate": 9.974516419393062e-07,
|
3173 |
+
"loss": 25.2643,
|
3174 |
+
"step": 4510
|
3175 |
+
},
|
3176 |
+
{
|
3177 |
+
"epoch": 0.3269143838712594,
|
3178 |
+
"grad_norm": 15.984375,
|
3179 |
+
"learning_rate": 9.974459914779742e-07,
|
3180 |
+
"loss": 25.1678,
|
3181 |
+
"step": 4520
|
3182 |
+
},
|
3183 |
+
{
|
3184 |
+
"epoch": 0.327637645782479,
|
3185 |
+
"grad_norm": 16.046875,
|
3186 |
+
"learning_rate": 9.974403410166422e-07,
|
3187 |
+
"loss": 25.2445,
|
3188 |
+
"step": 4530
|
3189 |
+
},
|
3190 |
+
{
|
3191 |
+
"epoch": 0.3283609076936986,
|
3192 |
+
"grad_norm": 16.96875,
|
3193 |
+
"learning_rate": 9.974346905553104e-07,
|
3194 |
+
"loss": 24.8622,
|
3195 |
+
"step": 4540
|
3196 |
+
},
|
3197 |
+
{
|
3198 |
+
"epoch": 0.3290841696049182,
|
3199 |
+
"grad_norm": 16.84375,
|
3200 |
+
"learning_rate": 9.974290400939784e-07,
|
3201 |
+
"loss": 25.0903,
|
3202 |
+
"step": 4550
|
3203 |
+
},
|
3204 |
+
{
|
3205 |
+
"epoch": 0.3298074315161378,
|
3206 |
+
"grad_norm": 16.71875,
|
3207 |
+
"learning_rate": 9.974233896326466e-07,
|
3208 |
+
"loss": 24.9463,
|
3209 |
+
"step": 4560
|
3210 |
+
},
|
3211 |
+
{
|
3212 |
+
"epoch": 0.33053069342735736,
|
3213 |
+
"grad_norm": 16.765625,
|
3214 |
+
"learning_rate": 9.974177391713146e-07,
|
3215 |
+
"loss": 25.1533,
|
3216 |
+
"step": 4570
|
3217 |
+
},
|
3218 |
+
{
|
3219 |
+
"epoch": 0.331253955338577,
|
3220 |
+
"grad_norm": 16.828125,
|
3221 |
+
"learning_rate": 9.974120887099828e-07,
|
3222 |
+
"loss": 25.1094,
|
3223 |
+
"step": 4580
|
3224 |
+
},
|
3225 |
+
{
|
3226 |
+
"epoch": 0.33197721724979656,
|
3227 |
+
"grad_norm": 16.8125,
|
3228 |
+
"learning_rate": 9.974064382486508e-07,
|
3229 |
+
"loss": 25.1725,
|
3230 |
+
"step": 4590
|
3231 |
+
},
|
3232 |
+
{
|
3233 |
+
"epoch": 0.3327004791610162,
|
3234 |
+
"grad_norm": 16.65625,
|
3235 |
+
"learning_rate": 9.974007877873188e-07,
|
3236 |
+
"loss": 24.9989,
|
3237 |
+
"step": 4600
|
3238 |
+
},
|
3239 |
+
{
|
3240 |
+
"epoch": 0.33342374107223577,
|
3241 |
+
"grad_norm": 15.7109375,
|
3242 |
+
"learning_rate": 9.97395137325987e-07,
|
3243 |
+
"loss": 24.916,
|
3244 |
+
"step": 4610
|
3245 |
+
},
|
3246 |
+
{
|
3247 |
+
"epoch": 0.3341470029834554,
|
3248 |
+
"grad_norm": 16.140625,
|
3249 |
+
"learning_rate": 9.97389486864655e-07,
|
3250 |
+
"loss": 24.9188,
|
3251 |
+
"step": 4620
|
3252 |
+
},
|
3253 |
+
{
|
3254 |
+
"epoch": 0.334870264894675,
|
3255 |
+
"grad_norm": 17.515625,
|
3256 |
+
"learning_rate": 9.973838364033232e-07,
|
3257 |
+
"loss": 24.9345,
|
3258 |
+
"step": 4630
|
3259 |
+
},
|
3260 |
+
{
|
3261 |
+
"epoch": 0.3355935268058946,
|
3262 |
+
"grad_norm": 16.4375,
|
3263 |
+
"learning_rate": 9.973781859419912e-07,
|
3264 |
+
"loss": 25.1931,
|
3265 |
+
"step": 4640
|
3266 |
+
},
|
3267 |
+
{
|
3268 |
+
"epoch": 0.3363167887171142,
|
3269 |
+
"grad_norm": 17.125,
|
3270 |
+
"learning_rate": 9.973725354806592e-07,
|
3271 |
+
"loss": 25.1182,
|
3272 |
+
"step": 4650
|
3273 |
+
},
|
3274 |
+
{
|
3275 |
+
"epoch": 0.3370400506283338,
|
3276 |
+
"grad_norm": 16.515625,
|
3277 |
+
"learning_rate": 9.973668850193274e-07,
|
3278 |
+
"loss": 25.5517,
|
3279 |
+
"step": 4660
|
3280 |
+
},
|
3281 |
+
{
|
3282 |
+
"epoch": 0.3377633125395534,
|
3283 |
+
"grad_norm": 16.8125,
|
3284 |
+
"learning_rate": 9.973612345579954e-07,
|
3285 |
+
"loss": 24.8154,
|
3286 |
+
"step": 4670
|
3287 |
+
},
|
3288 |
+
{
|
3289 |
+
"epoch": 0.33848657445077296,
|
3290 |
+
"grad_norm": 16.40625,
|
3291 |
+
"learning_rate": 9.973555840966636e-07,
|
3292 |
+
"loss": 25.2684,
|
3293 |
+
"step": 4680
|
3294 |
+
},
|
3295 |
+
{
|
3296 |
+
"epoch": 0.3392098363619926,
|
3297 |
+
"grad_norm": 16.109375,
|
3298 |
+
"learning_rate": 9.973499336353316e-07,
|
3299 |
+
"loss": 25.024,
|
3300 |
+
"step": 4690
|
3301 |
+
},
|
3302 |
+
{
|
3303 |
+
"epoch": 0.33993309827321216,
|
3304 |
+
"grad_norm": 16.296875,
|
3305 |
+
"learning_rate": 9.973442831739996e-07,
|
3306 |
+
"loss": 25.1598,
|
3307 |
+
"step": 4700
|
3308 |
+
},
|
3309 |
+
{
|
3310 |
+
"epoch": 0.3406563601844318,
|
3311 |
+
"grad_norm": 16.46875,
|
3312 |
+
"learning_rate": 9.973386327126678e-07,
|
3313 |
+
"loss": 25.0369,
|
3314 |
+
"step": 4710
|
3315 |
+
},
|
3316 |
+
{
|
3317 |
+
"epoch": 0.34137962209565137,
|
3318 |
+
"grad_norm": 15.53125,
|
3319 |
+
"learning_rate": 9.973329822513358e-07,
|
3320 |
+
"loss": 25.4247,
|
3321 |
+
"step": 4720
|
3322 |
+
},
|
3323 |
+
{
|
3324 |
+
"epoch": 0.342102884006871,
|
3325 |
+
"grad_norm": 16.75,
|
3326 |
+
"learning_rate": 9.97327331790004e-07,
|
3327 |
+
"loss": 25.0935,
|
3328 |
+
"step": 4730
|
3329 |
+
},
|
3330 |
+
{
|
3331 |
+
"epoch": 0.3428261459180906,
|
3332 |
+
"grad_norm": 15.84375,
|
3333 |
+
"learning_rate": 9.97321681328672e-07,
|
3334 |
+
"loss": 24.8433,
|
3335 |
+
"step": 4740
|
3336 |
+
},
|
3337 |
+
{
|
3338 |
+
"epoch": 0.3435494078293102,
|
3339 |
+
"grad_norm": 16.953125,
|
3340 |
+
"learning_rate": 9.9731603086734e-07,
|
3341 |
+
"loss": 25.3383,
|
3342 |
+
"step": 4750
|
3343 |
+
},
|
3344 |
+
{
|
3345 |
+
"epoch": 0.3442726697405298,
|
3346 |
+
"grad_norm": 16.15625,
|
3347 |
+
"learning_rate": 9.973103804060082e-07,
|
3348 |
+
"loss": 24.9644,
|
3349 |
+
"step": 4760
|
3350 |
+
},
|
3351 |
+
{
|
3352 |
+
"epoch": 0.3449959316517494,
|
3353 |
+
"grad_norm": 17.34375,
|
3354 |
+
"learning_rate": 9.973047299446762e-07,
|
3355 |
+
"loss": 25.2677,
|
3356 |
+
"step": 4770
|
3357 |
+
},
|
3358 |
+
{
|
3359 |
+
"epoch": 0.345719193562969,
|
3360 |
+
"grad_norm": 17.125,
|
3361 |
+
"learning_rate": 9.972990794833444e-07,
|
3362 |
+
"loss": 25.1205,
|
3363 |
+
"step": 4780
|
3364 |
+
},
|
3365 |
+
{
|
3366 |
+
"epoch": 0.3464424554741886,
|
3367 |
+
"grad_norm": 16.015625,
|
3368 |
+
"learning_rate": 9.972934290220124e-07,
|
3369 |
+
"loss": 24.9001,
|
3370 |
+
"step": 4790
|
3371 |
+
},
|
3372 |
+
{
|
3373 |
+
"epoch": 0.3471657173854082,
|
3374 |
+
"grad_norm": 16.984375,
|
3375 |
+
"learning_rate": 9.972877785606806e-07,
|
3376 |
+
"loss": 24.6669,
|
3377 |
+
"step": 4800
|
3378 |
+
},
|
3379 |
+
{
|
3380 |
+
"epoch": 0.34788897929662776,
|
3381 |
+
"grad_norm": 15.9453125,
|
3382 |
+
"learning_rate": 9.972821280993486e-07,
|
3383 |
+
"loss": 24.8667,
|
3384 |
+
"step": 4810
|
3385 |
+
},
|
3386 |
+
{
|
3387 |
+
"epoch": 0.3486122412078474,
|
3388 |
+
"grad_norm": 15.609375,
|
3389 |
+
"learning_rate": 9.972764776380166e-07,
|
3390 |
+
"loss": 24.8402,
|
3391 |
+
"step": 4820
|
3392 |
+
},
|
3393 |
+
{
|
3394 |
+
"epoch": 0.34933550311906697,
|
3395 |
+
"grad_norm": 15.3203125,
|
3396 |
+
"learning_rate": 9.972708271766848e-07,
|
3397 |
+
"loss": 25.0204,
|
3398 |
+
"step": 4830
|
3399 |
+
},
|
3400 |
+
{
|
3401 |
+
"epoch": 0.3500587650302866,
|
3402 |
+
"grad_norm": 16.0625,
|
3403 |
+
"learning_rate": 9.972651767153528e-07,
|
3404 |
+
"loss": 25.1096,
|
3405 |
+
"step": 4840
|
3406 |
+
},
|
3407 |
+
{
|
3408 |
+
"epoch": 0.3507820269415062,
|
3409 |
+
"grad_norm": 15.2421875,
|
3410 |
+
"learning_rate": 9.97259526254021e-07,
|
3411 |
+
"loss": 25.4209,
|
3412 |
+
"step": 4850
|
3413 |
+
},
|
3414 |
+
{
|
3415 |
+
"epoch": 0.3515052888527258,
|
3416 |
+
"grad_norm": 18.0,
|
3417 |
+
"learning_rate": 9.97253875792689e-07,
|
3418 |
+
"loss": 25.3704,
|
3419 |
+
"step": 4860
|
3420 |
+
},
|
3421 |
+
{
|
3422 |
+
"epoch": 0.3522285507639454,
|
3423 |
+
"grad_norm": 15.8984375,
|
3424 |
+
"learning_rate": 9.97248225331357e-07,
|
3425 |
+
"loss": 25.0614,
|
3426 |
+
"step": 4870
|
3427 |
+
},
|
3428 |
+
{
|
3429 |
+
"epoch": 0.352951812675165,
|
3430 |
+
"grad_norm": 16.5,
|
3431 |
+
"learning_rate": 9.972425748700253e-07,
|
3432 |
+
"loss": 24.972,
|
3433 |
+
"step": 4880
|
3434 |
+
},
|
3435 |
+
{
|
3436 |
+
"epoch": 0.3536750745863846,
|
3437 |
+
"grad_norm": 16.765625,
|
3438 |
+
"learning_rate": 9.972369244086933e-07,
|
3439 |
+
"loss": 25.0881,
|
3440 |
+
"step": 4890
|
3441 |
+
},
|
3442 |
+
{
|
3443 |
+
"epoch": 0.3543983364976042,
|
3444 |
+
"grad_norm": 16.546875,
|
3445 |
+
"learning_rate": 9.972312739473615e-07,
|
3446 |
+
"loss": 25.3568,
|
3447 |
+
"step": 4900
|
3448 |
+
},
|
3449 |
+
{
|
3450 |
+
"epoch": 0.3551215984088238,
|
3451 |
+
"grad_norm": 17.03125,
|
3452 |
+
"learning_rate": 9.972256234860295e-07,
|
3453 |
+
"loss": 24.9578,
|
3454 |
+
"step": 4910
|
3455 |
+
},
|
3456 |
+
{
|
3457 |
+
"epoch": 0.3558448603200434,
|
3458 |
+
"grad_norm": 17.015625,
|
3459 |
+
"learning_rate": 9.972199730246975e-07,
|
3460 |
+
"loss": 25.1697,
|
3461 |
+
"step": 4920
|
3462 |
+
},
|
3463 |
+
{
|
3464 |
+
"epoch": 0.356568122231263,
|
3465 |
+
"grad_norm": 17.28125,
|
3466 |
+
"learning_rate": 9.972143225633657e-07,
|
3467 |
+
"loss": 25.3285,
|
3468 |
+
"step": 4930
|
3469 |
+
},
|
3470 |
+
{
|
3471 |
+
"epoch": 0.35729138414248257,
|
3472 |
+
"grad_norm": 16.671875,
|
3473 |
+
"learning_rate": 9.972086721020337e-07,
|
3474 |
+
"loss": 25.0862,
|
3475 |
+
"step": 4940
|
3476 |
+
},
|
3477 |
+
{
|
3478 |
+
"epoch": 0.3580146460537022,
|
3479 |
+
"grad_norm": 15.0703125,
|
3480 |
+
"learning_rate": 9.972030216407019e-07,
|
3481 |
+
"loss": 25.1828,
|
3482 |
+
"step": 4950
|
3483 |
+
},
|
3484 |
+
{
|
3485 |
+
"epoch": 0.3587379079649218,
|
3486 |
+
"grad_norm": 15.6484375,
|
3487 |
+
"learning_rate": 9.971973711793699e-07,
|
3488 |
+
"loss": 25.1832,
|
3489 |
+
"step": 4960
|
3490 |
+
},
|
3491 |
+
{
|
3492 |
+
"epoch": 0.3594611698761414,
|
3493 |
+
"grad_norm": 16.46875,
|
3494 |
+
"learning_rate": 9.97191720718038e-07,
|
3495 |
+
"loss": 25.3505,
|
3496 |
+
"step": 4970
|
3497 |
+
},
|
3498 |
+
{
|
3499 |
+
"epoch": 0.360184431787361,
|
3500 |
+
"grad_norm": 16.421875,
|
3501 |
+
"learning_rate": 9.97186070256706e-07,
|
3502 |
+
"loss": 24.9362,
|
3503 |
+
"step": 4980
|
3504 |
+
},
|
3505 |
+
{
|
3506 |
+
"epoch": 0.3609076936985806,
|
3507 |
+
"grad_norm": 16.6875,
|
3508 |
+
"learning_rate": 9.97180419795374e-07,
|
3509 |
+
"loss": 25.5243,
|
3510 |
+
"step": 4990
|
3511 |
+
},
|
3512 |
+
{
|
3513 |
+
"epoch": 0.3616309556098002,
|
3514 |
+
"grad_norm": 16.234375,
|
3515 |
+
"learning_rate": 9.971747693340423e-07,
|
3516 |
+
"loss": 25.2643,
|
3517 |
+
"step": 5000
|
3518 |
+
},
|
3519 |
+
{
|
3520 |
+
"epoch": 0.3616309556098002,
|
3521 |
+
"eval_loss": 1.5911972522735596,
|
3522 |
+
"eval_runtime": 376.056,
|
3523 |
+
"eval_samples_per_second": 1930.72,
|
3524 |
+
"eval_steps_per_second": 30.168,
|
3525 |
+
"step": 5000
|
3526 |
}
|
3527 |
],
|
3528 |
"logging_steps": 10,
|
|
|
3542 |
"attributes": {}
|
3543 |
}
|
3544 |
},
|
3545 |
+
"total_flos": 6.984642783608832e+18,
|
3546 |
"train_batch_size": 16,
|
3547 |
"trial_name": null,
|
3548 |
"trial_params": null
|