Training in progress, step 717, checkpoint
Browse files- last-checkpoint/adapter_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/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +1676 -3
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2145944
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9515a7002cfa7f67e6c746a7e59b868de77da444db2b95a7c9cb039f0922f08
|
| 3 |
size 2145944
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4310020
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7971e0a79864672f8071d25876b2985e088af6d6ac063d1237778d2eaf4fd19b
|
| 3 |
size 4310020
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b753a7aacbc56b237a570162811e1a71ee10b02e1bee93e6daab70cd4680802
|
| 3 |
size 14960
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c1fb5420c0f25e5f6b2392bd3436352035eb0debaec1745bf48bdc44f8008fc2
|
| 3 |
size 15024
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2c82e31e1ac17c78c3d52136bae19d297a1c91d03f90b220e02fe92fb6cac45
|
| 3 |
size 15024
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1705ef3a08286669695bd3c71729dc6c292514e0dbba93fac60bc2587d6740cc
|
| 3 |
size 15024
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27ebf93c877c54673a5776fbfe0e2f8ca99926be562e6e5f55038ce1f0e80b42
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3353,6 +3353,1679 @@
|
|
| 3353 |
"learning_rate": 3.368986053895655e-05,
|
| 3354 |
"loss": 9.7648,
|
| 3355 |
"step": 478
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3356 |
}
|
| 3357 |
],
|
| 3358 |
"logging_steps": 1,
|
|
@@ -3372,7 +5045,7 @@
|
|
| 3372 |
"attributes": {}
|
| 3373 |
}
|
| 3374 |
},
|
| 3375 |
-
"total_flos":
|
| 3376 |
"train_batch_size": 4,
|
| 3377 |
"trial_name": null,
|
| 3378 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9893066574680924,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 717,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3353 |
"learning_rate": 3.368986053895655e-05,
|
| 3354 |
"loss": 9.7648,
|
| 3355 |
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.6609175577785443,
|
| 3359 |
+
"grad_norm": 1.1496732234954834,
|
| 3360 |
+
"learning_rate": 3.345210775911166e-05,
|
| 3361 |
+
"loss": 9.7602,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.6622973439116937,
|
| 3366 |
+
"grad_norm": 1.1272746324539185,
|
| 3367 |
+
"learning_rate": 3.3214774421669774e-05,
|
| 3368 |
+
"loss": 9.7809,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.6636771300448431,
|
| 3373 |
+
"grad_norm": 1.2115309238433838,
|
| 3374 |
+
"learning_rate": 3.297786654236169e-05,
|
| 3375 |
+
"loss": 9.7554,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.6650569161779925,
|
| 3380 |
+
"grad_norm": 1.1891155242919922,
|
| 3381 |
+
"learning_rate": 3.2741390126134106e-05,
|
| 3382 |
+
"loss": 9.766,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.6664367023111418,
|
| 3387 |
+
"grad_norm": 1.2283670902252197,
|
| 3388 |
+
"learning_rate": 3.250535116699729e-05,
|
| 3389 |
+
"loss": 9.7619,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.6678164884442911,
|
| 3394 |
+
"grad_norm": 1.259109377861023,
|
| 3395 |
+
"learning_rate": 3.226975564787322e-05,
|
| 3396 |
+
"loss": 9.7414,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.6691962745774405,
|
| 3401 |
+
"grad_norm": 1.322930932044983,
|
| 3402 |
+
"learning_rate": 3.2034609540443915e-05,
|
| 3403 |
+
"loss": 9.7465,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.6705760607105898,
|
| 3408 |
+
"grad_norm": 1.3455944061279297,
|
| 3409 |
+
"learning_rate": 3.179991880500009e-05,
|
| 3410 |
+
"loss": 9.7221,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.6719558468437392,
|
| 3415 |
+
"grad_norm": 1.5582841634750366,
|
| 3416 |
+
"learning_rate": 3.1565689390290066e-05,
|
| 3417 |
+
"loss": 9.6821,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.6733356329768886,
|
| 3422 |
+
"grad_norm": 1.1807150840759277,
|
| 3423 |
+
"learning_rate": 3.133192723336895e-05,
|
| 3424 |
+
"loss": 9.7562,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.674715419110038,
|
| 3429 |
+
"grad_norm": 0.9975462555885315,
|
| 3430 |
+
"learning_rate": 3.109863825944823e-05,
|
| 3431 |
+
"loss": 9.7919,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.6760952052431873,
|
| 3436 |
+
"grad_norm": 1.0711045265197754,
|
| 3437 |
+
"learning_rate": 3.086582838174551e-05,
|
| 3438 |
+
"loss": 9.7813,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.6774749913763367,
|
| 3443 |
+
"grad_norm": 1.1586861610412598,
|
| 3444 |
+
"learning_rate": 3.063350350133467e-05,
|
| 3445 |
+
"loss": 9.7722,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.6788547775094861,
|
| 3450 |
+
"grad_norm": 1.228026270866394,
|
| 3451 |
+
"learning_rate": 3.0401669506996256e-05,
|
| 3452 |
+
"loss": 9.7511,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.6802345636426353,
|
| 3457 |
+
"grad_norm": 1.205870270729065,
|
| 3458 |
+
"learning_rate": 3.0170332275068246e-05,
|
| 3459 |
+
"loss": 9.7468,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.6816143497757847,
|
| 3464 |
+
"grad_norm": 1.283031702041626,
|
| 3465 |
+
"learning_rate": 2.9939497669297112e-05,
|
| 3466 |
+
"loss": 9.7567,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 0.6829941359089341,
|
| 3471 |
+
"grad_norm": 1.2276543378829956,
|
| 3472 |
+
"learning_rate": 2.970917154068914e-05,
|
| 3473 |
+
"loss": 9.7648,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 0.6843739220420835,
|
| 3478 |
+
"grad_norm": 1.2356555461883545,
|
| 3479 |
+
"learning_rate": 2.9479359727362173e-05,
|
| 3480 |
+
"loss": 9.7521,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 0.6857537081752328,
|
| 3485 |
+
"grad_norm": 1.301045298576355,
|
| 3486 |
+
"learning_rate": 2.92500680543976e-05,
|
| 3487 |
+
"loss": 9.7518,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.6871334943083822,
|
| 3492 |
+
"grad_norm": 1.424596905708313,
|
| 3493 |
+
"learning_rate": 2.9021302333692734e-05,
|
| 3494 |
+
"loss": 9.7215,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 0.6885132804415316,
|
| 3499 |
+
"grad_norm": 1.409056305885315,
|
| 3500 |
+
"learning_rate": 2.879306836381345e-05,
|
| 3501 |
+
"loss": 9.7073,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 0.689893066574681,
|
| 3506 |
+
"grad_norm": 1.6761122941970825,
|
| 3507 |
+
"learning_rate": 2.8565371929847284e-05,
|
| 3508 |
+
"loss": 9.6887,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 0.6912728527078302,
|
| 3513 |
+
"grad_norm": 0.9244851469993591,
|
| 3514 |
+
"learning_rate": 2.8338218803256745e-05,
|
| 3515 |
+
"loss": 9.8076,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 0.6926526388409796,
|
| 3520 |
+
"grad_norm": 1.1031217575073242,
|
| 3521 |
+
"learning_rate": 2.811161474173297e-05,
|
| 3522 |
+
"loss": 9.7844,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 0.694032424974129,
|
| 3527 |
+
"grad_norm": 1.1576576232910156,
|
| 3528 |
+
"learning_rate": 2.7885565489049946e-05,
|
| 3529 |
+
"loss": 9.7647,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.6954122111072784,
|
| 3534 |
+
"grad_norm": 1.1267688274383545,
|
| 3535 |
+
"learning_rate": 2.7660076774918708e-05,
|
| 3536 |
+
"loss": 9.786,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 0.6967919972404277,
|
| 3541 |
+
"grad_norm": 1.165797472000122,
|
| 3542 |
+
"learning_rate": 2.7435154314842338e-05,
|
| 3543 |
+
"loss": 9.7736,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 0.6981717833735771,
|
| 3548 |
+
"grad_norm": 1.1887449026107788,
|
| 3549 |
+
"learning_rate": 2.7210803809970853e-05,
|
| 3550 |
+
"loss": 9.7515,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 0.6995515695067265,
|
| 3555 |
+
"grad_norm": 1.2904964685440063,
|
| 3556 |
+
"learning_rate": 2.6987030946956936e-05,
|
| 3557 |
+
"loss": 9.7602,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 0.7009313556398759,
|
| 3562 |
+
"grad_norm": 1.2434381246566772,
|
| 3563 |
+
"learning_rate": 2.6763841397811573e-05,
|
| 3564 |
+
"loss": 9.7476,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 0.7023111417730252,
|
| 3569 |
+
"grad_norm": 1.282375454902649,
|
| 3570 |
+
"learning_rate": 2.654124081976048e-05,
|
| 3571 |
+
"loss": 9.7474,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 0.7036909279061745,
|
| 3576 |
+
"grad_norm": 1.3851263523101807,
|
| 3577 |
+
"learning_rate": 2.631923485510054e-05,
|
| 3578 |
+
"loss": 9.7208,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 0.7050707140393239,
|
| 3583 |
+
"grad_norm": 1.4085386991500854,
|
| 3584 |
+
"learning_rate": 2.6097829131056907e-05,
|
| 3585 |
+
"loss": 9.7317,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 0.7064505001724732,
|
| 3590 |
+
"grad_norm": 1.4174658060073853,
|
| 3591 |
+
"learning_rate": 2.587702925964034e-05,
|
| 3592 |
+
"loss": 9.7054,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 0.7078302863056226,
|
| 3597 |
+
"grad_norm": 1.1571439504623413,
|
| 3598 |
+
"learning_rate": 2.565684083750488e-05,
|
| 3599 |
+
"loss": 9.7617,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 0.709210072438772,
|
| 3604 |
+
"grad_norm": 1.0106278657913208,
|
| 3605 |
+
"learning_rate": 2.5437269445806145e-05,
|
| 3606 |
+
"loss": 9.7957,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 0.7105898585719214,
|
| 3611 |
+
"grad_norm": 1.1029208898544312,
|
| 3612 |
+
"learning_rate": 2.5218320650059657e-05,
|
| 3613 |
+
"loss": 9.7729,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 0.7119696447050707,
|
| 3618 |
+
"grad_norm": 1.134572148323059,
|
| 3619 |
+
"learning_rate": 2.500000000000001e-05,
|
| 3620 |
+
"loss": 9.7702,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 0.7133494308382201,
|
| 3625 |
+
"grad_norm": 1.1734071969985962,
|
| 3626 |
+
"learning_rate": 2.478231302943997e-05,
|
| 3627 |
+
"loss": 9.782,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 0.7147292169713695,
|
| 3632 |
+
"grad_norm": 1.220053791999817,
|
| 3633 |
+
"learning_rate": 2.4565265256130394e-05,
|
| 3634 |
+
"loss": 9.7545,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 0.7161090031045187,
|
| 3639 |
+
"grad_norm": 1.2671356201171875,
|
| 3640 |
+
"learning_rate": 2.4348862181620224e-05,
|
| 3641 |
+
"loss": 9.7482,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 0.7174887892376681,
|
| 3646 |
+
"grad_norm": 1.2364708185195923,
|
| 3647 |
+
"learning_rate": 2.4133109291117156e-05,
|
| 3648 |
+
"loss": 9.7619,
|
| 3649 |
+
"step": 520
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.7188685753708175,
|
| 3653 |
+
"grad_norm": 1.251885175704956,
|
| 3654 |
+
"learning_rate": 2.391801205334856e-05,
|
| 3655 |
+
"loss": 9.7535,
|
| 3656 |
+
"step": 521
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 0.7202483615039669,
|
| 3660 |
+
"grad_norm": 1.345314621925354,
|
| 3661 |
+
"learning_rate": 2.3703575920422795e-05,
|
| 3662 |
+
"loss": 9.7491,
|
| 3663 |
+
"step": 522
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 0.7216281476371162,
|
| 3667 |
+
"grad_norm": 1.4312492609024048,
|
| 3668 |
+
"learning_rate": 2.3489806327691155e-05,
|
| 3669 |
+
"loss": 9.7244,
|
| 3670 |
+
"step": 523
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 0.7230079337702656,
|
| 3674 |
+
"grad_norm": 1.4245972633361816,
|
| 3675 |
+
"learning_rate": 2.3276708693609943e-05,
|
| 3676 |
+
"loss": 9.7038,
|
| 3677 |
+
"step": 524
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 0.724387719903415,
|
| 3681 |
+
"grad_norm": 1.5328452587127686,
|
| 3682 |
+
"learning_rate": 2.3064288419603292e-05,
|
| 3683 |
+
"loss": 9.6869,
|
| 3684 |
+
"step": 525
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 0.7257675060365644,
|
| 3688 |
+
"grad_norm": 0.9019702672958374,
|
| 3689 |
+
"learning_rate": 2.2852550889926067e-05,
|
| 3690 |
+
"loss": 9.8176,
|
| 3691 |
+
"step": 526
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.7271472921697137,
|
| 3695 |
+
"grad_norm": 1.0406774282455444,
|
| 3696 |
+
"learning_rate": 2.2641501471527595e-05,
|
| 3697 |
+
"loss": 9.7709,
|
| 3698 |
+
"step": 527
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 0.728527078302863,
|
| 3702 |
+
"grad_norm": 1.0880190134048462,
|
| 3703 |
+
"learning_rate": 2.243114551391542e-05,
|
| 3704 |
+
"loss": 9.7687,
|
| 3705 |
+
"step": 528
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 0.7299068644360124,
|
| 3709 |
+
"grad_norm": 1.1651344299316406,
|
| 3710 |
+
"learning_rate": 2.2221488349019903e-05,
|
| 3711 |
+
"loss": 9.7504,
|
| 3712 |
+
"step": 529
|
| 3713 |
+
},
|
| 3714 |
+
{
|
| 3715 |
+
"epoch": 0.7312866505691618,
|
| 3716 |
+
"grad_norm": 1.1736459732055664,
|
| 3717 |
+
"learning_rate": 2.20125352910589e-05,
|
| 3718 |
+
"loss": 9.7507,
|
| 3719 |
+
"step": 530
|
| 3720 |
+
},
|
| 3721 |
+
{
|
| 3722 |
+
"epoch": 0.7326664367023111,
|
| 3723 |
+
"grad_norm": 1.196784257888794,
|
| 3724 |
+
"learning_rate": 2.18042916364032e-05,
|
| 3725 |
+
"loss": 9.7769,
|
| 3726 |
+
"step": 531
|
| 3727 |
+
},
|
| 3728 |
+
{
|
| 3729 |
+
"epoch": 0.7340462228354605,
|
| 3730 |
+
"grad_norm": 1.2587947845458984,
|
| 3731 |
+
"learning_rate": 2.1596762663442218e-05,
|
| 3732 |
+
"loss": 9.7598,
|
| 3733 |
+
"step": 532
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 0.7354260089686099,
|
| 3737 |
+
"grad_norm": 1.1886099576950073,
|
| 3738 |
+
"learning_rate": 2.1389953632450145e-05,
|
| 3739 |
+
"loss": 9.7561,
|
| 3740 |
+
"step": 533
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 0.7368057951017593,
|
| 3744 |
+
"grad_norm": 1.3150951862335205,
|
| 3745 |
+
"learning_rate": 2.118386978545274e-05,
|
| 3746 |
+
"loss": 9.7332,
|
| 3747 |
+
"step": 534
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 0.7381855812349086,
|
| 3751 |
+
"grad_norm": 1.3374465703964233,
|
| 3752 |
+
"learning_rate": 2.097851634609434e-05,
|
| 3753 |
+
"loss": 9.7546,
|
| 3754 |
+
"step": 535
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"epoch": 0.739565367368058,
|
| 3758 |
+
"grad_norm": 1.3926987648010254,
|
| 3759 |
+
"learning_rate": 2.077389851950557e-05,
|
| 3760 |
+
"loss": 9.7265,
|
| 3761 |
+
"step": 536
|
| 3762 |
+
},
|
| 3763 |
+
{
|
| 3764 |
+
"epoch": 0.7409451535012073,
|
| 3765 |
+
"grad_norm": 1.4018597602844238,
|
| 3766 |
+
"learning_rate": 2.0570021492171264e-05,
|
| 3767 |
+
"loss": 9.7042,
|
| 3768 |
+
"step": 537
|
| 3769 |
+
},
|
| 3770 |
+
{
|
| 3771 |
+
"epoch": 0.7423249396343566,
|
| 3772 |
+
"grad_norm": 1.165271520614624,
|
| 3773 |
+
"learning_rate": 2.0366890431799167e-05,
|
| 3774 |
+
"loss": 9.7577,
|
| 3775 |
+
"step": 538
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 0.743704725767506,
|
| 3779 |
+
"grad_norm": 1.0279182195663452,
|
| 3780 |
+
"learning_rate": 2.0164510487188798e-05,
|
| 3781 |
+
"loss": 9.7801,
|
| 3782 |
+
"step": 539
|
| 3783 |
+
},
|
| 3784 |
+
{
|
| 3785 |
+
"epoch": 0.7450845119006554,
|
| 3786 |
+
"grad_norm": 1.0791254043579102,
|
| 3787 |
+
"learning_rate": 1.996288678810105e-05,
|
| 3788 |
+
"loss": 9.7826,
|
| 3789 |
+
"step": 540
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 0.7464642980338048,
|
| 3793 |
+
"grad_norm": 1.1344528198242188,
|
| 3794 |
+
"learning_rate": 1.976202444512813e-05,
|
| 3795 |
+
"loss": 9.7708,
|
| 3796 |
+
"step": 541
|
| 3797 |
+
},
|
| 3798 |
+
{
|
| 3799 |
+
"epoch": 0.7478440841669541,
|
| 3800 |
+
"grad_norm": 1.1656285524368286,
|
| 3801 |
+
"learning_rate": 1.9561928549563968e-05,
|
| 3802 |
+
"loss": 9.7673,
|
| 3803 |
+
"step": 542
|
| 3804 |
+
},
|
| 3805 |
+
{
|
| 3806 |
+
"epoch": 0.7492238703001035,
|
| 3807 |
+
"grad_norm": 1.1736102104187012,
|
| 3808 |
+
"learning_rate": 1.9362604173275268e-05,
|
| 3809 |
+
"loss": 9.7814,
|
| 3810 |
+
"step": 543
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.7506036564332529,
|
| 3814 |
+
"grad_norm": 1.2199715375900269,
|
| 3815 |
+
"learning_rate": 1.9164056368572846e-05,
|
| 3816 |
+
"loss": 9.7803,
|
| 3817 |
+
"step": 544
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 0.7519834425664023,
|
| 3821 |
+
"grad_norm": 1.196496844291687,
|
| 3822 |
+
"learning_rate": 1.8966290168083677e-05,
|
| 3823 |
+
"loss": 9.7406,
|
| 3824 |
+
"step": 545
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 0.7533632286995515,
|
| 3828 |
+
"grad_norm": 1.2514899969100952,
|
| 3829 |
+
"learning_rate": 1.87693105846232e-05,
|
| 3830 |
+
"loss": 9.7567,
|
| 3831 |
+
"step": 546
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 0.7547430148327009,
|
| 3835 |
+
"grad_norm": 1.3064500093460083,
|
| 3836 |
+
"learning_rate": 1.8573122611068406e-05,
|
| 3837 |
+
"loss": 9.7289,
|
| 3838 |
+
"step": 547
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 0.7561228009658503,
|
| 3842 |
+
"grad_norm": 1.400439739227295,
|
| 3843 |
+
"learning_rate": 1.837773122023114e-05,
|
| 3844 |
+
"loss": 9.7237,
|
| 3845 |
+
"step": 548
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 0.7575025870989996,
|
| 3849 |
+
"grad_norm": 1.3692952394485474,
|
| 3850 |
+
"learning_rate": 1.818314136473217e-05,
|
| 3851 |
+
"loss": 9.7226,
|
| 3852 |
+
"step": 549
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.758882373232149,
|
| 3856 |
+
"grad_norm": 1.5424178838729858,
|
| 3857 |
+
"learning_rate": 1.7989357976875603e-05,
|
| 3858 |
+
"loss": 9.6992,
|
| 3859 |
+
"step": 550
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 0.7602621593652984,
|
| 3863 |
+
"grad_norm": 0.8980839848518372,
|
| 3864 |
+
"learning_rate": 1.7796385968523815e-05,
|
| 3865 |
+
"loss": 9.811,
|
| 3866 |
+
"step": 551
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 0.7616419454984478,
|
| 3870 |
+
"grad_norm": 0.9639766216278076,
|
| 3871 |
+
"learning_rate": 1.760423023097307e-05,
|
| 3872 |
+
"loss": 9.7805,
|
| 3873 |
+
"step": 552
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 0.7630217316315971,
|
| 3877 |
+
"grad_norm": 1.0791096687316895,
|
| 3878 |
+
"learning_rate": 1.741289563482939e-05,
|
| 3879 |
+
"loss": 9.7779,
|
| 3880 |
+
"step": 553
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"epoch": 0.7644015177647465,
|
| 3884 |
+
"grad_norm": 1.1886143684387207,
|
| 3885 |
+
"learning_rate": 1.7222387029885268e-05,
|
| 3886 |
+
"loss": 9.7533,
|
| 3887 |
+
"step": 554
|
| 3888 |
+
},
|
| 3889 |
+
{
|
| 3890 |
+
"epoch": 0.7657813038978958,
|
| 3891 |
+
"grad_norm": 1.2121399641036987,
|
| 3892 |
+
"learning_rate": 1.703270924499656e-05,
|
| 3893 |
+
"loss": 9.756,
|
| 3894 |
+
"step": 555
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 0.7671610900310452,
|
| 3898 |
+
"grad_norm": 1.2668524980545044,
|
| 3899 |
+
"learning_rate": 1.684386708796025e-05,
|
| 3900 |
+
"loss": 9.7412,
|
| 3901 |
+
"step": 556
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 0.7685408761641945,
|
| 3905 |
+
"grad_norm": 1.212928056716919,
|
| 3906 |
+
"learning_rate": 1.665586534539246e-05,
|
| 3907 |
+
"loss": 9.7641,
|
| 3908 |
+
"step": 557
|
| 3909 |
+
},
|
| 3910 |
+
{
|
| 3911 |
+
"epoch": 0.7699206622973439,
|
| 3912 |
+
"grad_norm": 1.2839542627334595,
|
| 3913 |
+
"learning_rate": 1.646870878260721e-05,
|
| 3914 |
+
"loss": 9.7659,
|
| 3915 |
+
"step": 558
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 0.7713004484304933,
|
| 3919 |
+
"grad_norm": 1.2510404586791992,
|
| 3920 |
+
"learning_rate": 1.6282402143495568e-05,
|
| 3921 |
+
"loss": 9.7529,
|
| 3922 |
+
"step": 559
|
| 3923 |
+
},
|
| 3924 |
+
{
|
| 3925 |
+
"epoch": 0.7726802345636427,
|
| 3926 |
+
"grad_norm": 1.3847991228103638,
|
| 3927 |
+
"learning_rate": 1.6096950150405454e-05,
|
| 3928 |
+
"loss": 9.7254,
|
| 3929 |
+
"step": 560
|
| 3930 |
+
},
|
| 3931 |
+
{
|
| 3932 |
+
"epoch": 0.774060020696792,
|
| 3933 |
+
"grad_norm": 1.3537993431091309,
|
| 3934 |
+
"learning_rate": 1.5912357504021937e-05,
|
| 3935 |
+
"loss": 9.7168,
|
| 3936 |
+
"step": 561
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 0.7754398068299414,
|
| 3940 |
+
"grad_norm": 1.5678364038467407,
|
| 3941 |
+
"learning_rate": 1.5728628883248007e-05,
|
| 3942 |
+
"loss": 9.6773,
|
| 3943 |
+
"step": 562
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 0.7768195929630907,
|
| 3947 |
+
"grad_norm": 1.173730731010437,
|
| 3948 |
+
"learning_rate": 1.554576894508613e-05,
|
| 3949 |
+
"loss": 9.7439,
|
| 3950 |
+
"step": 563
|
| 3951 |
+
},
|
| 3952 |
+
{
|
| 3953 |
+
"epoch": 0.77819937909624,
|
| 3954 |
+
"grad_norm": 1.0499186515808105,
|
| 3955 |
+
"learning_rate": 1.536378232452003e-05,
|
| 3956 |
+
"loss": 9.7698,
|
| 3957 |
+
"step": 564
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 0.7795791652293894,
|
| 3961 |
+
"grad_norm": 1.048542857170105,
|
| 3962 |
+
"learning_rate": 1.5182673634397365e-05,
|
| 3963 |
+
"loss": 9.7965,
|
| 3964 |
+
"step": 565
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 0.7809589513625388,
|
| 3968 |
+
"grad_norm": 1.149909496307373,
|
| 3969 |
+
"learning_rate": 1.5002447465312675e-05,
|
| 3970 |
+
"loss": 9.7774,
|
| 3971 |
+
"step": 566
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 0.7823387374956882,
|
| 3975 |
+
"grad_norm": 1.1968557834625244,
|
| 3976 |
+
"learning_rate": 1.4823108385491124e-05,
|
| 3977 |
+
"loss": 9.7471,
|
| 3978 |
+
"step": 567
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 0.7837185236288375,
|
| 3982 |
+
"grad_norm": 1.135711908340454,
|
| 3983 |
+
"learning_rate": 1.4644660940672627e-05,
|
| 3984 |
+
"loss": 9.7686,
|
| 3985 |
+
"step": 568
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 0.7850983097619869,
|
| 3989 |
+
"grad_norm": 1.3147321939468384,
|
| 3990 |
+
"learning_rate": 1.4467109653996697e-05,
|
| 3991 |
+
"loss": 9.7424,
|
| 3992 |
+
"step": 569
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 0.7864780958951363,
|
| 3996 |
+
"grad_norm": 1.1964170932769775,
|
| 3997 |
+
"learning_rate": 1.429045902588777e-05,
|
| 3998 |
+
"loss": 9.7611,
|
| 3999 |
+
"step": 570
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 0.7878578820282857,
|
| 4003 |
+
"grad_norm": 1.376891016960144,
|
| 4004 |
+
"learning_rate": 1.4114713533941082e-05,
|
| 4005 |
+
"loss": 9.7149,
|
| 4006 |
+
"step": 571
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 0.7892376681614349,
|
| 4010 |
+
"grad_norm": 1.3214612007141113,
|
| 4011 |
+
"learning_rate": 1.3939877632809278e-05,
|
| 4012 |
+
"loss": 9.7467,
|
| 4013 |
+
"step": 572
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 0.7906174542945843,
|
| 4017 |
+
"grad_norm": 1.346970796585083,
|
| 4018 |
+
"learning_rate": 1.3765955754089383e-05,
|
| 4019 |
+
"loss": 9.7122,
|
| 4020 |
+
"step": 573
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 0.7919972404277337,
|
| 4024 |
+
"grad_norm": 1.4399006366729736,
|
| 4025 |
+
"learning_rate": 1.3592952306210588e-05,
|
| 4026 |
+
"loss": 9.7232,
|
| 4027 |
+
"step": 574
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 0.793377026560883,
|
| 4031 |
+
"grad_norm": 1.645202875137329,
|
| 4032 |
+
"learning_rate": 1.3420871674322394e-05,
|
| 4033 |
+
"loss": 9.6723,
|
| 4034 |
+
"step": 575
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 0.7947568126940324,
|
| 4038 |
+
"grad_norm": 0.9395657181739807,
|
| 4039 |
+
"learning_rate": 1.3249718220183583e-05,
|
| 4040 |
+
"loss": 9.8069,
|
| 4041 |
+
"step": 576
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 0.7961365988271818,
|
| 4045 |
+
"grad_norm": 1.003074049949646,
|
| 4046 |
+
"learning_rate": 1.3079496282051529e-05,
|
| 4047 |
+
"loss": 9.7872,
|
| 4048 |
+
"step": 577
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 0.7975163849603312,
|
| 4052 |
+
"grad_norm": 1.1187260150909424,
|
| 4053 |
+
"learning_rate": 1.2910210174572346e-05,
|
| 4054 |
+
"loss": 9.7854,
|
| 4055 |
+
"step": 578
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 0.7988961710934805,
|
| 4059 |
+
"grad_norm": 1.1820377111434937,
|
| 4060 |
+
"learning_rate": 1.2741864188671492e-05,
|
| 4061 |
+
"loss": 9.7445,
|
| 4062 |
+
"step": 579
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 0.8002759572266299,
|
| 4066 |
+
"grad_norm": 1.135115146636963,
|
| 4067 |
+
"learning_rate": 1.257446259144494e-05,
|
| 4068 |
+
"loss": 9.743,
|
| 4069 |
+
"step": 580
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 0.8016557433597792,
|
| 4073 |
+
"grad_norm": 1.2748380899429321,
|
| 4074 |
+
"learning_rate": 1.2408009626051137e-05,
|
| 4075 |
+
"loss": 9.7491,
|
| 4076 |
+
"step": 581
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 0.8030355294929286,
|
| 4080 |
+
"grad_norm": 1.2434227466583252,
|
| 4081 |
+
"learning_rate": 1.2242509511603317e-05,
|
| 4082 |
+
"loss": 9.7696,
|
| 4083 |
+
"step": 582
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 0.8044153156260779,
|
| 4087 |
+
"grad_norm": 1.251225233078003,
|
| 4088 |
+
"learning_rate": 1.2077966443062705e-05,
|
| 4089 |
+
"loss": 9.7418,
|
| 4090 |
+
"step": 583
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 0.8057951017592273,
|
| 4094 |
+
"grad_norm": 1.2516130208969116,
|
| 4095 |
+
"learning_rate": 1.1914384591132044e-05,
|
| 4096 |
+
"loss": 9.7266,
|
| 4097 |
+
"step": 584
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.8071748878923767,
|
| 4101 |
+
"grad_norm": 1.392945647239685,
|
| 4102 |
+
"learning_rate": 1.1751768102150007e-05,
|
| 4103 |
+
"loss": 9.7269,
|
| 4104 |
+
"step": 585
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 0.808554674025526,
|
| 4108 |
+
"grad_norm": 1.510158896446228,
|
| 4109 |
+
"learning_rate": 1.159012109798598e-05,
|
| 4110 |
+
"loss": 9.7113,
|
| 4111 |
+
"step": 586
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 0.8099344601586754,
|
| 4115 |
+
"grad_norm": 1.5267751216888428,
|
| 4116 |
+
"learning_rate": 1.1429447675935695e-05,
|
| 4117 |
+
"loss": 9.704,
|
| 4118 |
+
"step": 587
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 0.8113142462918248,
|
| 4122 |
+
"grad_norm": 1.1726975440979004,
|
| 4123 |
+
"learning_rate": 1.1269751908617277e-05,
|
| 4124 |
+
"loss": 9.7488,
|
| 4125 |
+
"step": 588
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 0.8126940324249742,
|
| 4129 |
+
"grad_norm": 0.9718952775001526,
|
| 4130 |
+
"learning_rate": 1.1111037843868094e-05,
|
| 4131 |
+
"loss": 9.7921,
|
| 4132 |
+
"step": 589
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 0.8140738185581234,
|
| 4136 |
+
"grad_norm": 1.0413875579833984,
|
| 4137 |
+
"learning_rate": 1.0953309504642128e-05,
|
| 4138 |
+
"loss": 9.7741,
|
| 4139 |
+
"step": 590
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.8154536046912728,
|
| 4143 |
+
"grad_norm": 1.1728605031967163,
|
| 4144 |
+
"learning_rate": 1.0796570888907953e-05,
|
| 4145 |
+
"loss": 9.7641,
|
| 4146 |
+
"step": 591
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 0.8168333908244222,
|
| 4150 |
+
"grad_norm": 1.157538652420044,
|
| 4151 |
+
"learning_rate": 1.0640825969547496e-05,
|
| 4152 |
+
"loss": 9.7696,
|
| 4153 |
+
"step": 592
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 0.8182131769575716,
|
| 4157 |
+
"grad_norm": 1.2358477115631104,
|
| 4158 |
+
"learning_rate": 1.0486078694255242e-05,
|
| 4159 |
+
"loss": 9.7469,
|
| 4160 |
+
"step": 593
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 0.8195929630907209,
|
| 4164 |
+
"grad_norm": 1.2600133419036865,
|
| 4165 |
+
"learning_rate": 1.0332332985438248e-05,
|
| 4166 |
+
"loss": 9.747,
|
| 4167 |
+
"step": 594
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 0.8209727492238703,
|
| 4171 |
+
"grad_norm": 1.2197083234786987,
|
| 4172 |
+
"learning_rate": 1.017959274011665e-05,
|
| 4173 |
+
"loss": 9.7532,
|
| 4174 |
+
"step": 595
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 0.8223525353570197,
|
| 4178 |
+
"grad_norm": 1.3774051666259766,
|
| 4179 |
+
"learning_rate": 1.0027861829824952e-05,
|
| 4180 |
+
"loss": 9.7325,
|
| 4181 |
+
"step": 596
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 0.8237323214901691,
|
| 4185 |
+
"grad_norm": 1.38478422164917,
|
| 4186 |
+
"learning_rate": 9.87714410051383e-06,
|
| 4187 |
+
"loss": 9.7306,
|
| 4188 |
+
"step": 597
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 0.8251121076233184,
|
| 4192 |
+
"grad_norm": 1.432771921157837,
|
| 4193 |
+
"learning_rate": 9.7274433724527e-06,
|
| 4194 |
+
"loss": 9.7276,
|
| 4195 |
+
"step": 598
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 0.8264918937564677,
|
| 4199 |
+
"grad_norm": 1.5427794456481934,
|
| 4200 |
+
"learning_rate": 9.578763440132883e-06,
|
| 4201 |
+
"loss": 9.7113,
|
| 4202 |
+
"step": 599
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 0.8278716798896171,
|
| 4206 |
+
"grad_norm": 1.5743299722671509,
|
| 4207 |
+
"learning_rate": 9.431108072171346e-06,
|
| 4208 |
+
"loss": 9.6803,
|
| 4209 |
+
"step": 600
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 0.8292514660227664,
|
| 4213 |
+
"grad_norm": 0.921704113483429,
|
| 4214 |
+
"learning_rate": 9.284481011215318e-06,
|
| 4215 |
+
"loss": 9.7834,
|
| 4216 |
+
"step": 601
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.8306312521559158,
|
| 4220 |
+
"grad_norm": 1.0345510244369507,
|
| 4221 |
+
"learning_rate": 9.138885973847261e-06,
|
| 4222 |
+
"loss": 9.79,
|
| 4223 |
+
"step": 602
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 0.8320110382890652,
|
| 4227 |
+
"grad_norm": 1.182210087776184,
|
| 4228 |
+
"learning_rate": 8.994326650490831e-06,
|
| 4229 |
+
"loss": 9.7632,
|
| 4230 |
+
"step": 603
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 0.8333908244222146,
|
| 4234 |
+
"grad_norm": 1.1576429605484009,
|
| 4235 |
+
"learning_rate": 8.850806705317183e-06,
|
| 4236 |
+
"loss": 9.7689,
|
| 4237 |
+
"step": 604
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 0.834770610555364,
|
| 4241 |
+
"grad_norm": 1.2053581476211548,
|
| 4242 |
+
"learning_rate": 8.708329776152224e-06,
|
| 4243 |
+
"loss": 9.7558,
|
| 4244 |
+
"step": 605
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 0.8361503966885133,
|
| 4248 |
+
"grad_norm": 1.2201980352401733,
|
| 4249 |
+
"learning_rate": 8.566899474384299e-06,
|
| 4250 |
+
"loss": 9.7571,
|
| 4251 |
+
"step": 606
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 0.8375301828216627,
|
| 4255 |
+
"grad_norm": 1.1964430809020996,
|
| 4256 |
+
"learning_rate": 8.426519384872733e-06,
|
| 4257 |
+
"loss": 9.7458,
|
| 4258 |
+
"step": 607
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.838909968954812,
|
| 4262 |
+
"grad_norm": 1.321816086769104,
|
| 4263 |
+
"learning_rate": 8.287193065856935e-06,
|
| 4264 |
+
"loss": 9.7307,
|
| 4265 |
+
"step": 608
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 0.8402897550879613,
|
| 4269 |
+
"grad_norm": 1.290749430656433,
|
| 4270 |
+
"learning_rate": 8.148924048866169e-06,
|
| 4271 |
+
"loss": 9.7321,
|
| 4272 |
+
"step": 609
|
| 4273 |
+
},
|
| 4274 |
+
{
|
| 4275 |
+
"epoch": 0.8416695412211107,
|
| 4276 |
+
"grad_norm": 1.353271722793579,
|
| 4277 |
+
"learning_rate": 8.011715838630107e-06,
|
| 4278 |
+
"loss": 9.7231,
|
| 4279 |
+
"step": 610
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 0.8430493273542601,
|
| 4283 |
+
"grad_norm": 1.4000368118286133,
|
| 4284 |
+
"learning_rate": 7.875571912989938e-06,
|
| 4285 |
+
"loss": 9.7083,
|
| 4286 |
+
"step": 611
|
| 4287 |
+
},
|
| 4288 |
+
{
|
| 4289 |
+
"epoch": 0.8444291134874095,
|
| 4290 |
+
"grad_norm": 1.4400883913040161,
|
| 4291 |
+
"learning_rate": 7.740495722810271e-06,
|
| 4292 |
+
"loss": 9.7085,
|
| 4293 |
+
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 0.8458088996205588,
|
| 4297 |
+
"grad_norm": 1.134522795677185,
|
| 4298 |
+
"learning_rate": 7.606490691891577e-06,
|
| 4299 |
+
"loss": 9.7551,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 0.8471886857537082,
|
| 4304 |
+
"grad_norm": 1.0418236255645752,
|
| 4305 |
+
"learning_rate": 7.4735602168835236e-06,
|
| 4306 |
+
"loss": 9.7693,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 0.8485684718868576,
|
| 4311 |
+
"grad_norm": 1.080651879310608,
|
| 4312 |
+
"learning_rate": 7.341707667198766e-06,
|
| 4313 |
+
"loss": 9.7787,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 0.849948258020007,
|
| 4318 |
+
"grad_norm": 1.1420639753341675,
|
| 4319 |
+
"learning_rate": 7.21093638492763e-06,
|
| 4320 |
+
"loss": 9.7674,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 0.8513280441531562,
|
| 4325 |
+
"grad_norm": 1.221085786819458,
|
| 4326 |
+
"learning_rate": 7.081249684753327e-06,
|
| 4327 |
+
"loss": 9.7556,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 0.8527078302863056,
|
| 4332 |
+
"grad_norm": 1.2512654066085815,
|
| 4333 |
+
"learning_rate": 6.952650853867993e-06,
|
| 4334 |
+
"loss": 9.7305,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 0.854087616419455,
|
| 4339 |
+
"grad_norm": 1.2443575859069824,
|
| 4340 |
+
"learning_rate": 6.825143151889357e-06,
|
| 4341 |
+
"loss": 9.7577,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 0.8554674025526043,
|
| 4346 |
+
"grad_norm": 1.2522335052490234,
|
| 4347 |
+
"learning_rate": 6.698729810778065e-06,
|
| 4348 |
+
"loss": 9.7628,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 0.8568471886857537,
|
| 4353 |
+
"grad_norm": 1.360615611076355,
|
| 4354 |
+
"learning_rate": 6.5734140347558536e-06,
|
| 4355 |
+
"loss": 9.7367,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 0.8582269748189031,
|
| 4360 |
+
"grad_norm": 1.3139582872390747,
|
| 4361 |
+
"learning_rate": 6.449199000224221e-06,
|
| 4362 |
+
"loss": 9.732,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 0.8596067609520525,
|
| 4367 |
+
"grad_norm": 1.3527382612228394,
|
| 4368 |
+
"learning_rate": 6.3260878556840375e-06,
|
| 4369 |
+
"loss": 9.7361,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 0.8609865470852018,
|
| 4374 |
+
"grad_norm": 1.4311867952346802,
|
| 4375 |
+
"learning_rate": 6.204083721655607e-06,
|
| 4376 |
+
"loss": 9.706,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 0.8623663332183512,
|
| 4381 |
+
"grad_norm": 1.6295082569122314,
|
| 4382 |
+
"learning_rate": 6.083189690599711e-06,
|
| 4383 |
+
"loss": 9.6803,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 0.8637461193515005,
|
| 4388 |
+
"grad_norm": 0.9568197131156921,
|
| 4389 |
+
"learning_rate": 5.9634088268390784e-06,
|
| 4390 |
+
"loss": 9.7839,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 0.8651259054846498,
|
| 4395 |
+
"grad_norm": 1.080190896987915,
|
| 4396 |
+
"learning_rate": 5.8447441664808424e-06,
|
| 4397 |
+
"loss": 9.7847,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 0.8665056916177992,
|
| 4402 |
+
"grad_norm": 1.0810461044311523,
|
| 4403 |
+
"learning_rate": 5.727198717339511e-06,
|
| 4404 |
+
"loss": 9.7586,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 0.8678854777509486,
|
| 4409 |
+
"grad_norm": 1.1114495992660522,
|
| 4410 |
+
"learning_rate": 5.610775458860718e-06,
|
| 4411 |
+
"loss": 9.7832,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 0.869265263884098,
|
| 4416 |
+
"grad_norm": 1.2362908124923706,
|
| 4417 |
+
"learning_rate": 5.495477342045779e-06,
|
| 4418 |
+
"loss": 9.7565,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 0.8706450500172473,
|
| 4423 |
+
"grad_norm": 1.2435539960861206,
|
| 4424 |
+
"learning_rate": 5.381307289376786e-06,
|
| 4425 |
+
"loss": 9.7561,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 0.8720248361503967,
|
| 4430 |
+
"grad_norm": 1.3071742057800293,
|
| 4431 |
+
"learning_rate": 5.2682681947426375e-06,
|
| 4432 |
+
"loss": 9.7482,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 0.8734046222835461,
|
| 4437 |
+
"grad_norm": 1.2518724203109741,
|
| 4438 |
+
"learning_rate": 5.156362923365588e-06,
|
| 4439 |
+
"loss": 9.7569,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 0.8747844084166954,
|
| 4444 |
+
"grad_norm": 1.3225029706954956,
|
| 4445 |
+
"learning_rate": 5.045594311728707e-06,
|
| 4446 |
+
"loss": 9.7413,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 0.8761641945498447,
|
| 4451 |
+
"grad_norm": 1.3300327062606812,
|
| 4452 |
+
"learning_rate": 4.93596516750392e-06,
|
| 4453 |
+
"loss": 9.7283,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 0.8775439806829941,
|
| 4458 |
+
"grad_norm": 1.3841952085494995,
|
| 4459 |
+
"learning_rate": 4.827478269480895e-06,
|
| 4460 |
+
"loss": 9.7273,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 0.8789237668161435,
|
| 4465 |
+
"grad_norm": 1.5817680358886719,
|
| 4466 |
+
"learning_rate": 4.720136367496536e-06,
|
| 4467 |
+
"loss": 9.671,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 0.8803035529492929,
|
| 4472 |
+
"grad_norm": 1.1424212455749512,
|
| 4473 |
+
"learning_rate": 4.613942182365372e-06,
|
| 4474 |
+
"loss": 9.7741,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 0.8816833390824422,
|
| 4479 |
+
"grad_norm": 1.0098145008087158,
|
| 4480 |
+
"learning_rate": 4.508898405810524e-06,
|
| 4481 |
+
"loss": 9.7891,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 0.8830631252155916,
|
| 4486 |
+
"grad_norm": 1.0495284795761108,
|
| 4487 |
+
"learning_rate": 4.405007700395497e-06,
|
| 4488 |
+
"loss": 9.7944,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 0.884442911348741,
|
| 4493 |
+
"grad_norm": 1.165906310081482,
|
| 4494 |
+
"learning_rate": 4.3022726994567105e-06,
|
| 4495 |
+
"loss": 9.773,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 0.8858226974818904,
|
| 4500 |
+
"grad_norm": 1.1741012334823608,
|
| 4501 |
+
"learning_rate": 4.200696007036703e-06,
|
| 4502 |
+
"loss": 9.767,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.8872024836150396,
|
| 4507 |
+
"grad_norm": 1.2205920219421387,
|
| 4508 |
+
"learning_rate": 4.100280197818207e-06,
|
| 4509 |
+
"loss": 9.7574,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 0.888582269748189,
|
| 4514 |
+
"grad_norm": 1.2672699689865112,
|
| 4515 |
+
"learning_rate": 4.001027817058789e-06,
|
| 4516 |
+
"loss": 9.7688,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 0.8899620558813384,
|
| 4521 |
+
"grad_norm": 1.2361886501312256,
|
| 4522 |
+
"learning_rate": 3.902941380526426e-06,
|
| 4523 |
+
"loss": 9.7636,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 0.8913418420144877,
|
| 4528 |
+
"grad_norm": 1.322023868560791,
|
| 4529 |
+
"learning_rate": 3.8060233744356633e-06,
|
| 4530 |
+
"loss": 9.7413,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 0.8927216281476371,
|
| 4535 |
+
"grad_norm": 1.3382331132888794,
|
| 4536 |
+
"learning_rate": 3.7102762553846536e-06,
|
| 4537 |
+
"loss": 9.7279,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 0.8941014142807865,
|
| 4542 |
+
"grad_norm": 1.345215082168579,
|
| 4543 |
+
"learning_rate": 3.615702450292857e-06,
|
| 4544 |
+
"loss": 9.745,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 0.8954812004139359,
|
| 4549 |
+
"grad_norm": 1.3454034328460693,
|
| 4550 |
+
"learning_rate": 3.5223043563395288e-06,
|
| 4551 |
+
"loss": 9.721,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 0.8968609865470852,
|
| 4556 |
+
"grad_norm": 1.4947688579559326,
|
| 4557 |
+
"learning_rate": 3.4300843409029726e-06,
|
| 4558 |
+
"loss": 9.7119,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 0.8982407726802346,
|
| 4563 |
+
"grad_norm": 0.9093929529190063,
|
| 4564 |
+
"learning_rate": 3.3390447415005077e-06,
|
| 4565 |
+
"loss": 9.8062,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 0.8996205588133839,
|
| 4570 |
+
"grad_norm": 1.040328860282898,
|
| 4571 |
+
"learning_rate": 3.249187865729264e-06,
|
| 4572 |
+
"loss": 9.7664,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 0.9010003449465332,
|
| 4577 |
+
"grad_norm": 1.1193190813064575,
|
| 4578 |
+
"learning_rate": 3.160515991207619e-06,
|
| 4579 |
+
"loss": 9.7838,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 0.9023801310796826,
|
| 4584 |
+
"grad_norm": 1.14997136592865,
|
| 4585 |
+
"learning_rate": 3.0730313655175645e-06,
|
| 4586 |
+
"loss": 9.768,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 0.903759917212832,
|
| 4591 |
+
"grad_norm": 1.1968692541122437,
|
| 4592 |
+
"learning_rate": 2.9867362061476276e-06,
|
| 4593 |
+
"loss": 9.7622,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 0.9051397033459814,
|
| 4598 |
+
"grad_norm": 1.1504422426223755,
|
| 4599 |
+
"learning_rate": 2.901632700436757e-06,
|
| 4600 |
+
"loss": 9.7521,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 0.9065194894791307,
|
| 4605 |
+
"grad_norm": 1.2824842929840088,
|
| 4606 |
+
"learning_rate": 2.8177230055188353e-06,
|
| 4607 |
+
"loss": 9.7633,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 0.9078992756122801,
|
| 4612 |
+
"grad_norm": 1.2600338459014893,
|
| 4613 |
+
"learning_rate": 2.7350092482679836e-06,
|
| 4614 |
+
"loss": 9.7494,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 0.9092790617454295,
|
| 4619 |
+
"grad_norm": 1.2912122011184692,
|
| 4620 |
+
"learning_rate": 2.653493525244721e-06,
|
| 4621 |
+
"loss": 9.7263,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 0.9106588478785789,
|
| 4626 |
+
"grad_norm": 1.415069580078125,
|
| 4627 |
+
"learning_rate": 2.573177902642726e-06,
|
| 4628 |
+
"loss": 9.7187,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 0.9120386340117281,
|
| 4633 |
+
"grad_norm": 1.3458380699157715,
|
| 4634 |
+
"learning_rate": 2.4940644162365522e-06,
|
| 4635 |
+
"loss": 9.7266,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 0.9134184201448775,
|
| 4640 |
+
"grad_norm": 1.5353480577468872,
|
| 4641 |
+
"learning_rate": 2.416155071329973e-06,
|
| 4642 |
+
"loss": 9.6964,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 0.9147982062780269,
|
| 4647 |
+
"grad_norm": 1.0634307861328125,
|
| 4648 |
+
"learning_rate": 2.339451842705187e-06,
|
| 4649 |
+
"loss": 9.7851,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 0.9161779924111763,
|
| 4654 |
+
"grad_norm": 1.0400745868682861,
|
| 4655 |
+
"learning_rate": 2.2639566745727205e-06,
|
| 4656 |
+
"loss": 9.7713,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 0.9175577785443256,
|
| 4661 |
+
"grad_norm": 1.134487509727478,
|
| 4662 |
+
"learning_rate": 2.189671480522204e-06,
|
| 4663 |
+
"loss": 9.7789,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 0.918937564677475,
|
| 4668 |
+
"grad_norm": 1.1036391258239746,
|
| 4669 |
+
"learning_rate": 2.1165981434738026e-06,
|
| 4670 |
+
"loss": 9.7722,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 0.9203173508106244,
|
| 4675 |
+
"grad_norm": 1.1182793378829956,
|
| 4676 |
+
"learning_rate": 2.0447385156305565e-06,
|
| 4677 |
+
"loss": 9.7933,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 0.9216971369437738,
|
| 4682 |
+
"grad_norm": 1.2758280038833618,
|
| 4683 |
+
"learning_rate": 1.974094418431388e-06,
|
| 4684 |
+
"loss": 9.7561,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 0.9230769230769231,
|
| 4689 |
+
"grad_norm": 1.2668101787567139,
|
| 4690 |
+
"learning_rate": 1.9046676425049315e-06,
|
| 4691 |
+
"loss": 9.7434,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 0.9244567092100724,
|
| 4696 |
+
"grad_norm": 1.2824689149856567,
|
| 4697 |
+
"learning_rate": 1.8364599476241862e-06,
|
| 4698 |
+
"loss": 9.7397,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 0.9258364953432218,
|
| 4703 |
+
"grad_norm": 1.2909704446792603,
|
| 4704 |
+
"learning_rate": 1.7694730626618583e-06,
|
| 4705 |
+
"loss": 9.7306,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 0.9272162814763711,
|
| 4710 |
+
"grad_norm": 1.3211469650268555,
|
| 4711 |
+
"learning_rate": 1.70370868554659e-06,
|
| 4712 |
+
"loss": 9.7482,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 0.9285960676095205,
|
| 4717 |
+
"grad_norm": 1.3531687259674072,
|
| 4718 |
+
"learning_rate": 1.6391684832198718e-06,
|
| 4719 |
+
"loss": 9.7382,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 0.9299758537426699,
|
| 4724 |
+
"grad_norm": 1.4724624156951904,
|
| 4725 |
+
"learning_rate": 1.5758540915938368e-06,
|
| 4726 |
+
"loss": 9.7051,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 0.9313556398758193,
|
| 4731 |
+
"grad_norm": 1.6450470685958862,
|
| 4732 |
+
"learning_rate": 1.5137671155097565e-06,
|
| 4733 |
+
"loss": 9.6571,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 0.9327354260089686,
|
| 4738 |
+
"grad_norm": 0.8748622536659241,
|
| 4739 |
+
"learning_rate": 1.4529091286973995e-06,
|
| 4740 |
+
"loss": 9.8211,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 0.934115212142118,
|
| 4745 |
+
"grad_norm": 1.065389633178711,
|
| 4746 |
+
"learning_rate": 1.3932816737351085e-06,
|
| 4747 |
+
"loss": 9.7792,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 0.9354949982752674,
|
| 4752 |
+
"grad_norm": 1.1271086931228638,
|
| 4753 |
+
"learning_rate": 1.3348862620107038e-06,
|
| 4754 |
+
"loss": 9.7635,
|
| 4755 |
+
"step": 678
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 0.9368747844084166,
|
| 4759 |
+
"grad_norm": 1.126579999923706,
|
| 4760 |
+
"learning_rate": 1.27772437368322e-06,
|
| 4761 |
+
"loss": 9.7736,
|
| 4762 |
+
"step": 679
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 0.938254570541566,
|
| 4766 |
+
"grad_norm": 1.1667218208312988,
|
| 4767 |
+
"learning_rate": 1.2217974576453073e-06,
|
| 4768 |
+
"loss": 9.7564,
|
| 4769 |
+
"step": 680
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 0.9396343566747154,
|
| 4773 |
+
"grad_norm": 1.243585228919983,
|
| 4774 |
+
"learning_rate": 1.1671069314865834e-06,
|
| 4775 |
+
"loss": 9.7614,
|
| 4776 |
+
"step": 681
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 0.9410141428078648,
|
| 4780 |
+
"grad_norm": 1.1980286836624146,
|
| 4781 |
+
"learning_rate": 1.1136541814576573e-06,
|
| 4782 |
+
"loss": 9.7782,
|
| 4783 |
+
"step": 682
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 0.9423939289410141,
|
| 4787 |
+
"grad_norm": 1.2429455518722534,
|
| 4788 |
+
"learning_rate": 1.061440562435001e-06,
|
| 4789 |
+
"loss": 9.7618,
|
| 4790 |
+
"step": 683
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 0.9437737150741635,
|
| 4794 |
+
"grad_norm": 1.306204080581665,
|
| 4795 |
+
"learning_rate": 1.0104673978866164e-06,
|
| 4796 |
+
"loss": 9.7464,
|
| 4797 |
+
"step": 684
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 0.9451535012073129,
|
| 4801 |
+
"grad_norm": 1.3777480125427246,
|
| 4802 |
+
"learning_rate": 9.607359798384785e-07,
|
| 4803 |
+
"loss": 9.7229,
|
| 4804 |
+
"step": 685
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 0.9465332873404623,
|
| 4808 |
+
"grad_norm": 1.3626405000686646,
|
| 4809 |
+
"learning_rate": 9.122475688417953e-07,
|
| 4810 |
+
"loss": 9.712,
|
| 4811 |
+
"step": 686
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 0.9479130734736116,
|
| 4815 |
+
"grad_norm": 1.4798493385314941,
|
| 4816 |
+
"learning_rate": 8.650033939410385e-07,
|
| 4817 |
+
"loss": 9.7036,
|
| 4818 |
+
"step": 687
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 0.9492928596067609,
|
| 4822 |
+
"grad_norm": 1.1905264854431152,
|
| 4823 |
+
"learning_rate": 8.190046526428242e-07,
|
| 4824 |
+
"loss": 9.7456,
|
| 4825 |
+
"step": 688
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 0.9506726457399103,
|
| 4829 |
+
"grad_norm": 0.9911084175109863,
|
| 4830 |
+
"learning_rate": 7.742525108855148e-07,
|
| 4831 |
+
"loss": 9.7855,
|
| 4832 |
+
"step": 689
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 0.9520524318730597,
|
| 4836 |
+
"grad_norm": 1.0882097482681274,
|
| 4837 |
+
"learning_rate": 7.307481030097152e-07,
|
| 4838 |
+
"loss": 9.7704,
|
| 4839 |
+
"step": 690
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 0.953432218006209,
|
| 4843 |
+
"grad_norm": 1.1580933332443237,
|
| 4844 |
+
"learning_rate": 6.884925317294677e-07,
|
| 4845 |
+
"loss": 9.7733,
|
| 4846 |
+
"step": 691
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 0.9548120041393584,
|
| 4850 |
+
"grad_norm": 1.1494512557983398,
|
| 4851 |
+
"learning_rate": 6.474868681043578e-07,
|
| 4852 |
+
"loss": 9.7834,
|
| 4853 |
+
"step": 692
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 0.9561917902725078,
|
| 4857 |
+
"grad_norm": 1.1650831699371338,
|
| 4858 |
+
"learning_rate": 6.077321515123136e-07,
|
| 4859 |
+
"loss": 9.7708,
|
| 4860 |
+
"step": 693
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 0.9575715764056572,
|
| 4864 |
+
"grad_norm": 1.2284369468688965,
|
| 4865 |
+
"learning_rate": 5.692293896232936e-07,
|
| 4866 |
+
"loss": 9.7564,
|
| 4867 |
+
"step": 694
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 0.9589513625388065,
|
| 4871 |
+
"grad_norm": 1.2123266458511353,
|
| 4872 |
+
"learning_rate": 5.319795583737242e-07,
|
| 4873 |
+
"loss": 9.7493,
|
| 4874 |
+
"step": 695
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 0.9603311486719558,
|
| 4878 |
+
"grad_norm": 1.2671525478363037,
|
| 4879 |
+
"learning_rate": 4.959836019417963e-07,
|
| 4880 |
+
"loss": 9.7491,
|
| 4881 |
+
"step": 696
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 0.9617109348051052,
|
| 4885 |
+
"grad_norm": 1.2901546955108643,
|
| 4886 |
+
"learning_rate": 4.6124243272349653e-07,
|
| 4887 |
+
"loss": 9.745,
|
| 4888 |
+
"step": 697
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 0.9630907209382545,
|
| 4892 |
+
"grad_norm": 1.3228528499603271,
|
| 4893 |
+
"learning_rate": 4.277569313094809e-07,
|
| 4894 |
+
"loss": 9.7245,
|
| 4895 |
+
"step": 698
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 0.9644705070714039,
|
| 4899 |
+
"grad_norm": 1.4477964639663696,
|
| 4900 |
+
"learning_rate": 3.9552794646279255e-07,
|
| 4901 |
+
"loss": 9.7099,
|
| 4902 |
+
"step": 699
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 0.9658502932045533,
|
| 4906 |
+
"grad_norm": 1.4955841302871704,
|
| 4907 |
+
"learning_rate": 3.6455629509730136e-07,
|
| 4908 |
+
"loss": 9.6912,
|
| 4909 |
+
"step": 700
|
| 4910 |
+
},
|
| 4911 |
+
{
|
| 4912 |
+
"epoch": 0.9672300793377027,
|
| 4913 |
+
"grad_norm": 0.8777285218238831,
|
| 4914 |
+
"learning_rate": 3.3484276225703714e-07,
|
| 4915 |
+
"loss": 9.8028,
|
| 4916 |
+
"step": 701
|
| 4917 |
+
},
|
| 4918 |
+
{
|
| 4919 |
+
"epoch": 0.968609865470852,
|
| 4920 |
+
"grad_norm": 1.04829740524292,
|
| 4921 |
+
"learning_rate": 3.0638810109626103e-07,
|
| 4922 |
+
"loss": 9.7746,
|
| 4923 |
+
"step": 702
|
| 4924 |
+
},
|
| 4925 |
+
{
|
| 4926 |
+
"epoch": 0.9699896516040014,
|
| 4927 |
+
"grad_norm": 1.18967604637146,
|
| 4928 |
+
"learning_rate": 2.79193032860392e-07,
|
| 4929 |
+
"loss": 9.7497,
|
| 4930 |
+
"step": 703
|
| 4931 |
+
},
|
| 4932 |
+
{
|
| 4933 |
+
"epoch": 0.9713694377371508,
|
| 4934 |
+
"grad_norm": 1.118967890739441,
|
| 4935 |
+
"learning_rate": 2.532582468677214e-07,
|
| 4936 |
+
"loss": 9.7728,
|
| 4937 |
+
"step": 704
|
| 4938 |
+
},
|
| 4939 |
+
{
|
| 4940 |
+
"epoch": 0.9727492238703,
|
| 4941 |
+
"grad_norm": 1.229331612586975,
|
| 4942 |
+
"learning_rate": 2.285844004919324e-07,
|
| 4943 |
+
"loss": 9.7464,
|
| 4944 |
+
"step": 705
|
| 4945 |
+
},
|
| 4946 |
+
{
|
| 4947 |
+
"epoch": 0.9741290100034494,
|
| 4948 |
+
"grad_norm": 1.1736466884613037,
|
| 4949 |
+
"learning_rate": 2.0517211914545254e-07,
|
| 4950 |
+
"loss": 9.7577,
|
| 4951 |
+
"step": 706
|
| 4952 |
+
},
|
| 4953 |
+
{
|
| 4954 |
+
"epoch": 0.9755087961365988,
|
| 4955 |
+
"grad_norm": 1.2047674655914307,
|
| 4956 |
+
"learning_rate": 1.8302199626358818e-07,
|
| 4957 |
+
"loss": 9.752,
|
| 4958 |
+
"step": 707
|
| 4959 |
+
},
|
| 4960 |
+
{
|
| 4961 |
+
"epoch": 0.9768885822697482,
|
| 4962 |
+
"grad_norm": 1.251903772354126,
|
| 4963 |
+
"learning_rate": 1.6213459328950352e-07,
|
| 4964 |
+
"loss": 9.7461,
|
| 4965 |
+
"step": 708
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.9782683684028975,
|
| 4969 |
+
"grad_norm": 1.369354486465454,
|
| 4970 |
+
"learning_rate": 1.4251043965994304e-07,
|
| 4971 |
+
"loss": 9.7395,
|
| 4972 |
+
"step": 709
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.9796481545360469,
|
| 4976 |
+
"grad_norm": 1.3757730722427368,
|
| 4977 |
+
"learning_rate": 1.2415003279186987e-07,
|
| 4978 |
+
"loss": 9.7186,
|
| 4979 |
+
"step": 710
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.9810279406691963,
|
| 4983 |
+
"grad_norm": 1.3921722173690796,
|
| 4984 |
+
"learning_rate": 1.0705383806982606e-07,
|
| 4985 |
+
"loss": 9.7241,
|
| 4986 |
+
"step": 711
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.9824077268023457,
|
| 4990 |
+
"grad_norm": 1.4798989295959473,
|
| 4991 |
+
"learning_rate": 9.12222888341252e-08,
|
| 4992 |
+
"loss": 9.7088,
|
| 4993 |
+
"step": 712
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.983787512935495,
|
| 4997 |
+
"grad_norm": 1.149962067604065,
|
| 4998 |
+
"learning_rate": 7.665578636990579e-08,
|
| 4999 |
+
"loss": 9.7613,
|
| 5000 |
+
"step": 713
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.9851672990686443,
|
| 5004 |
+
"grad_norm": 0.952008843421936,
|
| 5005 |
+
"learning_rate": 6.335469989692256e-08,
|
| 5006 |
+
"loss": 9.7976,
|
| 5007 |
+
"step": 714
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.9865470852017937,
|
| 5011 |
+
"grad_norm": 1.064375400543213,
|
| 5012 |
+
"learning_rate": 5.131936656020409e-08,
|
| 5013 |
+
"loss": 9.7691,
|
| 5014 |
+
"step": 715
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.9879268713349431,
|
| 5018 |
+
"grad_norm": 1.1418269872665405,
|
| 5019 |
+
"learning_rate": 4.055009142152067e-08,
|
| 5020 |
+
"loss": 9.7665,
|
| 5021 |
+
"step": 716
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.9893066574680924,
|
| 5025 |
+
"grad_norm": 1.1035135984420776,
|
| 5026 |
+
"learning_rate": 3.10471474516183e-08,
|
| 5027 |
+
"loss": 9.782,
|
| 5028 |
+
"step": 717
|
| 5029 |
}
|
| 5030 |
],
|
| 5031 |
"logging_steps": 1,
|
|
|
|
| 5045 |
"attributes": {}
|
| 5046 |
}
|
| 5047 |
},
|
| 5048 |
+
"total_flos": 595349320237056.0,
|
| 5049 |
"train_batch_size": 4,
|
| 5050 |
"trial_name": null,
|
| 5051 |
"trial_params": null
|