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 sha256:9e464bbe95cc23854803b672fb9cf729efc290b58ff545522f92b64c7016eecd
|
| 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 sha256:822935ab8fc15c5f806a5134daaf8d3b53f885d21a92bd3e33b1903ae914170d
|
| 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:e63ec6d4f8a228ccb62beccd67d6f6e9016931e6b30ae5344fdba8c30cfe35a8
|
| 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 sha256:33825bed89719638e2e37f21656a413d50d0dc8cf99d86b4d7152f50e5bbcd6e
|
| 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 sha256:5f3682dfb503773bdb3a7d4868d8abf3b6eed45d692e22e2299624c46632a667
|
| 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 sha256:7cb07e5da45d7c5643fe3179d2cfab1712f94a8c62afddc66fee2e58ea42f700
|
| 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:f024d2f6fb6610551c472834de25d1d904c6aa9a110ea61cc065fb2a17fa713f
|
| 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": 4.7536583265546775e-05,
|
| 3354 |
"loss": 9.609,
|
| 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": 0.8599700149925037,
|
| 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": 4.7536583265546775e-05,
|
| 3354 |
"loss": 9.609,
|
| 3355 |
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.5745127436281859,
|
| 3359 |
+
"grad_norm": 0.2816186845302582,
|
| 3360 |
+
"learning_rate": 4.7322569886417006e-05,
|
| 3361 |
+
"loss": 9.6101,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.5757121439280359,
|
| 3366 |
+
"grad_norm": 0.2793320417404175,
|
| 3367 |
+
"learning_rate": 4.71086056896371e-05,
|
| 3368 |
+
"loss": 9.6206,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.576911544227886,
|
| 3373 |
+
"grad_norm": 0.2865123748779297,
|
| 3374 |
+
"learning_rate": 4.689469460556626e-05,
|
| 3375 |
+
"loss": 9.6109,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.5781109445277361,
|
| 3380 |
+
"grad_norm": 0.2744526267051697,
|
| 3381 |
+
"learning_rate": 4.6680840563587966e-05,
|
| 3382 |
+
"loss": 9.6222,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.5793103448275863,
|
| 3387 |
+
"grad_norm": 0.30048128962516785,
|
| 3388 |
+
"learning_rate": 4.646704749203793e-05,
|
| 3389 |
+
"loss": 9.6182,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.5805097451274362,
|
| 3394 |
+
"grad_norm": 0.29160621762275696,
|
| 3395 |
+
"learning_rate": 4.6253319318131926e-05,
|
| 3396 |
+
"loss": 9.618,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.5817091454272864,
|
| 3401 |
+
"grad_norm": 0.31267857551574707,
|
| 3402 |
+
"learning_rate": 4.60396599678935e-05,
|
| 3403 |
+
"loss": 9.622,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.5829085457271365,
|
| 3408 |
+
"grad_norm": 0.3598839044570923,
|
| 3409 |
+
"learning_rate": 4.582607336608205e-05,
|
| 3410 |
+
"loss": 9.6176,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.5841079460269865,
|
| 3415 |
+
"grad_norm": 0.33458805084228516,
|
| 3416 |
+
"learning_rate": 4.561256343612061e-05,
|
| 3417 |
+
"loss": 9.6256,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.5853073463268366,
|
| 3422 |
+
"grad_norm": 0.27461162209510803,
|
| 3423 |
+
"learning_rate": 4.539913410002378e-05,
|
| 3424 |
+
"loss": 9.6119,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.5865067466266867,
|
| 3429 |
+
"grad_norm": 0.2723887264728546,
|
| 3430 |
+
"learning_rate": 4.518578927832577e-05,
|
| 3431 |
+
"loss": 9.6056,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.5877061469265368,
|
| 3436 |
+
"grad_norm": 0.2768537998199463,
|
| 3437 |
+
"learning_rate": 4.4972532890008313e-05,
|
| 3438 |
+
"loss": 9.6079,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.5889055472263868,
|
| 3443 |
+
"grad_norm": 0.2774599492549896,
|
| 3444 |
+
"learning_rate": 4.4759368852428625e-05,
|
| 3445 |
+
"loss": 9.6092,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.5901049475262369,
|
| 3450 |
+
"grad_norm": 0.27346640825271606,
|
| 3451 |
+
"learning_rate": 4.45463010812476e-05,
|
| 3452 |
+
"loss": 9.6143,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.591304347826087,
|
| 3457 |
+
"grad_norm": 0.2797171175479889,
|
| 3458 |
+
"learning_rate": 4.433333349035773e-05,
|
| 3459 |
+
"loss": 9.6168,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.592503748125937,
|
| 3464 |
+
"grad_norm": 0.2800818085670471,
|
| 3465 |
+
"learning_rate": 4.4120469991811296e-05,
|
| 3466 |
+
"loss": 9.6165,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 0.5937031484257871,
|
| 3471 |
+
"grad_norm": 0.280519038438797,
|
| 3472 |
+
"learning_rate": 4.390771449574846e-05,
|
| 3473 |
+
"loss": 9.6195,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 0.5949025487256372,
|
| 3478 |
+
"grad_norm": 0.2884778678417206,
|
| 3479 |
+
"learning_rate": 4.369507091032551e-05,
|
| 3480 |
+
"loss": 9.6132,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 0.5961019490254873,
|
| 3485 |
+
"grad_norm": 0.2894138693809509,
|
| 3486 |
+
"learning_rate": 4.3482543141642943e-05,
|
| 3487 |
+
"loss": 9.6147,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.5973013493253373,
|
| 3492 |
+
"grad_norm": 0.2868705093860626,
|
| 3493 |
+
"learning_rate": 4.327013509367386e-05,
|
| 3494 |
+
"loss": 9.6242,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 0.5985007496251874,
|
| 3499 |
+
"grad_norm": 0.2994021773338318,
|
| 3500 |
+
"learning_rate": 4.305785066819218e-05,
|
| 3501 |
+
"loss": 9.6189,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 0.5997001499250375,
|
| 3506 |
+
"grad_norm": 0.3168644607067108,
|
| 3507 |
+
"learning_rate": 4.2845693764700914e-05,
|
| 3508 |
+
"loss": 9.6247,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 0.6008995502248875,
|
| 3513 |
+
"grad_norm": 0.26666632294654846,
|
| 3514 |
+
"learning_rate": 4.263366828036065e-05,
|
| 3515 |
+
"loss": 9.6057,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 0.6020989505247376,
|
| 3520 |
+
"grad_norm": 0.26327091455459595,
|
| 3521 |
+
"learning_rate": 4.242177810991789e-05,
|
| 3522 |
+
"loss": 9.6115,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 0.6032983508245877,
|
| 3527 |
+
"grad_norm": 0.27538183331489563,
|
| 3528 |
+
"learning_rate": 4.221002714563347e-05,
|
| 3529 |
+
"loss": 9.6082,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.6044977511244378,
|
| 3534 |
+
"grad_norm": 0.27597832679748535,
|
| 3535 |
+
"learning_rate": 4.19984192772112e-05,
|
| 3536 |
+
"loss": 9.6075,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 0.6056971514242878,
|
| 3541 |
+
"grad_norm": 0.28365880250930786,
|
| 3542 |
+
"learning_rate": 4.1786958391726314e-05,
|
| 3543 |
+
"loss": 9.6136,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 0.6068965517241379,
|
| 3548 |
+
"grad_norm": 0.2802659273147583,
|
| 3549 |
+
"learning_rate": 4.1575648373554e-05,
|
| 3550 |
+
"loss": 9.6158,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 0.608095952023988,
|
| 3555 |
+
"grad_norm": 0.2841864228248596,
|
| 3556 |
+
"learning_rate": 4.136449310429822e-05,
|
| 3557 |
+
"loss": 9.6115,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 0.6092953523238381,
|
| 3562 |
+
"grad_norm": 0.2928536832332611,
|
| 3563 |
+
"learning_rate": 4.115349646272029e-05,
|
| 3564 |
+
"loss": 9.6156,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 0.6104947526236881,
|
| 3569 |
+
"grad_norm": 0.2854699492454529,
|
| 3570 |
+
"learning_rate": 4.0942662324667627e-05,
|
| 3571 |
+
"loss": 9.6137,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 0.6116941529235382,
|
| 3576 |
+
"grad_norm": 0.29192522168159485,
|
| 3577 |
+
"learning_rate": 4.0731994563002606e-05,
|
| 3578 |
+
"loss": 9.6136,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 0.6128935532233883,
|
| 3583 |
+
"grad_norm": 0.3441016674041748,
|
| 3584 |
+
"learning_rate": 4.052149704753142e-05,
|
| 3585 |
+
"loss": 9.6224,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 0.6140929535232383,
|
| 3590 |
+
"grad_norm": 0.3597991466522217,
|
| 3591 |
+
"learning_rate": 4.03111736449329e-05,
|
| 3592 |
+
"loss": 9.6219,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 0.6152923538230884,
|
| 3597 |
+
"grad_norm": 0.2781412899494171,
|
| 3598 |
+
"learning_rate": 4.010102821868762e-05,
|
| 3599 |
+
"loss": 9.6056,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 0.6164917541229386,
|
| 3604 |
+
"grad_norm": 0.27280357480049133,
|
| 3605 |
+
"learning_rate": 3.989106462900686e-05,
|
| 3606 |
+
"loss": 9.6063,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 0.6176911544227887,
|
| 3611 |
+
"grad_norm": 0.27366748452186584,
|
| 3612 |
+
"learning_rate": 3.968128673276165e-05,
|
| 3613 |
+
"loss": 9.6104,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 0.6188905547226387,
|
| 3618 |
+
"grad_norm": 0.27588704228401184,
|
| 3619 |
+
"learning_rate": 3.947169838341202e-05,
|
| 3620 |
+
"loss": 9.605,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 0.6200899550224888,
|
| 3625 |
+
"grad_norm": 0.27753859758377075,
|
| 3626 |
+
"learning_rate": 3.9262303430936164e-05,
|
| 3627 |
+
"loss": 9.6033,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 0.6212893553223389,
|
| 3632 |
+
"grad_norm": 0.27255064249038696,
|
| 3633 |
+
"learning_rate": 3.9053105721759696e-05,
|
| 3634 |
+
"loss": 9.6098,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 0.6224887556221889,
|
| 3639 |
+
"grad_norm": 0.2782951295375824,
|
| 3640 |
+
"learning_rate": 3.8844109098685045e-05,
|
| 3641 |
+
"loss": 9.6184,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 0.623688155922039,
|
| 3646 |
+
"grad_norm": 0.28660768270492554,
|
| 3647 |
+
"learning_rate": 3.8635317400820855e-05,
|
| 3648 |
+
"loss": 9.6113,
|
| 3649 |
+
"step": 520
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.6248875562218891,
|
| 3653 |
+
"grad_norm": 0.28494128584861755,
|
| 3654 |
+
"learning_rate": 3.842673446351138e-05,
|
| 3655 |
+
"loss": 9.6105,
|
| 3656 |
+
"step": 521
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 0.6260869565217392,
|
| 3660 |
+
"grad_norm": 0.28198301792144775,
|
| 3661 |
+
"learning_rate": 3.82183641182662e-05,
|
| 3662 |
+
"loss": 9.626,
|
| 3663 |
+
"step": 522
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 0.6272863568215892,
|
| 3667 |
+
"grad_norm": 0.2875995337963104,
|
| 3668 |
+
"learning_rate": 3.801021019268969e-05,
|
| 3669 |
+
"loss": 9.6176,
|
| 3670 |
+
"step": 523
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 0.6284857571214393,
|
| 3674 |
+
"grad_norm": 0.2956449091434479,
|
| 3675 |
+
"learning_rate": 3.780227651041073e-05,
|
| 3676 |
+
"loss": 9.6229,
|
| 3677 |
+
"step": 524
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 0.6296851574212894,
|
| 3681 |
+
"grad_norm": 0.37847524881362915,
|
| 3682 |
+
"learning_rate": 3.7594566891012546e-05,
|
| 3683 |
+
"loss": 9.6214,
|
| 3684 |
+
"step": 525
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 0.6308845577211394,
|
| 3688 |
+
"grad_norm": 0.27030348777770996,
|
| 3689 |
+
"learning_rate": 3.7387085149962507e-05,
|
| 3690 |
+
"loss": 9.6011,
|
| 3691 |
+
"step": 526
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.6320839580209895,
|
| 3695 |
+
"grad_norm": 0.274962455034256,
|
| 3696 |
+
"learning_rate": 3.717983509854198e-05,
|
| 3697 |
+
"loss": 9.6023,
|
| 3698 |
+
"step": 527
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 0.6332833583208396,
|
| 3702 |
+
"grad_norm": 0.27726373076438904,
|
| 3703 |
+
"learning_rate": 3.69728205437764e-05,
|
| 3704 |
+
"loss": 9.6102,
|
| 3705 |
+
"step": 528
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 0.6344827586206897,
|
| 3709 |
+
"grad_norm": 0.27569401264190674,
|
| 3710 |
+
"learning_rate": 3.676604528836535e-05,
|
| 3711 |
+
"loss": 9.6077,
|
| 3712 |
+
"step": 529
|
| 3713 |
+
},
|
| 3714 |
+
{
|
| 3715 |
+
"epoch": 0.6356821589205397,
|
| 3716 |
+
"grad_norm": 0.2719118893146515,
|
| 3717 |
+
"learning_rate": 3.6559513130612565e-05,
|
| 3718 |
+
"loss": 9.6078,
|
| 3719 |
+
"step": 530
|
| 3720 |
+
},
|
| 3721 |
+
{
|
| 3722 |
+
"epoch": 0.6368815592203898,
|
| 3723 |
+
"grad_norm": 0.27930060029029846,
|
| 3724 |
+
"learning_rate": 3.635322786435635e-05,
|
| 3725 |
+
"loss": 9.6099,
|
| 3726 |
+
"step": 531
|
| 3727 |
+
},
|
| 3728 |
+
{
|
| 3729 |
+
"epoch": 0.6380809595202399,
|
| 3730 |
+
"grad_norm": 0.2761722505092621,
|
| 3731 |
+
"learning_rate": 3.614719327889978e-05,
|
| 3732 |
+
"loss": 9.6161,
|
| 3733 |
+
"step": 532
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 0.6392803598200899,
|
| 3737 |
+
"grad_norm": 0.2825543284416199,
|
| 3738 |
+
"learning_rate": 3.594141315894108e-05,
|
| 3739 |
+
"loss": 9.616,
|
| 3740 |
+
"step": 533
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 0.64047976011994,
|
| 3744 |
+
"grad_norm": 0.28519946336746216,
|
| 3745 |
+
"learning_rate": 3.573589128450418e-05,
|
| 3746 |
+
"loss": 9.6134,
|
| 3747 |
+
"step": 534
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 0.6416791604197901,
|
| 3751 |
+
"grad_norm": 0.2859567105770111,
|
| 3752 |
+
"learning_rate": 3.5530631430869234e-05,
|
| 3753 |
+
"loss": 9.6181,
|
| 3754 |
+
"step": 535
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"epoch": 0.6428785607196402,
|
| 3758 |
+
"grad_norm": 0.293560653924942,
|
| 3759 |
+
"learning_rate": 3.532563736850322e-05,
|
| 3760 |
+
"loss": 9.6141,
|
| 3761 |
+
"step": 536
|
| 3762 |
+
},
|
| 3763 |
+
{
|
| 3764 |
+
"epoch": 0.6440779610194902,
|
| 3765 |
+
"grad_norm": 0.31543228030204773,
|
| 3766 |
+
"learning_rate": 3.512091286299081e-05,
|
| 3767 |
+
"loss": 9.6132,
|
| 3768 |
+
"step": 537
|
| 3769 |
+
},
|
| 3770 |
+
{
|
| 3771 |
+
"epoch": 0.6452773613193403,
|
| 3772 |
+
"grad_norm": 0.28361520171165466,
|
| 3773 |
+
"learning_rate": 3.491646167496507e-05,
|
| 3774 |
+
"loss": 9.5993,
|
| 3775 |
+
"step": 538
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 0.6464767616191904,
|
| 3779 |
+
"grad_norm": 0.2670563757419586,
|
| 3780 |
+
"learning_rate": 3.4712287560038446e-05,
|
| 3781 |
+
"loss": 9.6042,
|
| 3782 |
+
"step": 539
|
| 3783 |
+
},
|
| 3784 |
+
{
|
| 3785 |
+
"epoch": 0.6476761619190404,
|
| 3786 |
+
"grad_norm": 0.2657446265220642,
|
| 3787 |
+
"learning_rate": 3.450839426873378e-05,
|
| 3788 |
+
"loss": 9.6106,
|
| 3789 |
+
"step": 540
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 0.6488755622188905,
|
| 3793 |
+
"grad_norm": 0.271816611289978,
|
| 3794 |
+
"learning_rate": 3.4304785546415374e-05,
|
| 3795 |
+
"loss": 9.608,
|
| 3796 |
+
"step": 541
|
| 3797 |
+
},
|
| 3798 |
+
{
|
| 3799 |
+
"epoch": 0.6500749625187406,
|
| 3800 |
+
"grad_norm": 0.27191296219825745,
|
| 3801 |
+
"learning_rate": 3.41014651332202e-05,
|
| 3802 |
+
"loss": 9.6103,
|
| 3803 |
+
"step": 542
|
| 3804 |
+
},
|
| 3805 |
+
{
|
| 3806 |
+
"epoch": 0.6512743628185907,
|
| 3807 |
+
"grad_norm": 0.27644070982933044,
|
| 3808 |
+
"learning_rate": 3.3898436763989247e-05,
|
| 3809 |
+
"loss": 9.6039,
|
| 3810 |
+
"step": 543
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.6524737631184407,
|
| 3814 |
+
"grad_norm": 0.27742430567741394,
|
| 3815 |
+
"learning_rate": 3.369570416819889e-05,
|
| 3816 |
+
"loss": 9.6053,
|
| 3817 |
+
"step": 544
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 0.6536731634182908,
|
| 3821 |
+
"grad_norm": 0.2793113589286804,
|
| 3822 |
+
"learning_rate": 3.349327106989232e-05,
|
| 3823 |
+
"loss": 9.615,
|
| 3824 |
+
"step": 545
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 0.654872563718141,
|
| 3828 |
+
"grad_norm": 0.28077057003974915,
|
| 3829 |
+
"learning_rate": 3.329114118761123e-05,
|
| 3830 |
+
"loss": 9.6101,
|
| 3831 |
+
"step": 546
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 0.656071964017991,
|
| 3835 |
+
"grad_norm": 0.2894865870475769,
|
| 3836 |
+
"learning_rate": 3.308931823432744e-05,
|
| 3837 |
+
"loss": 9.6093,
|
| 3838 |
+
"step": 547
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 0.6572713643178411,
|
| 3842 |
+
"grad_norm": 0.2894723415374756,
|
| 3843 |
+
"learning_rate": 3.288780591737474e-05,
|
| 3844 |
+
"loss": 9.6141,
|
| 3845 |
+
"step": 548
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 0.6584707646176912,
|
| 3849 |
+
"grad_norm": 0.3010658323764801,
|
| 3850 |
+
"learning_rate": 3.268660793838074e-05,
|
| 3851 |
+
"loss": 9.6249,
|
| 3852 |
+
"step": 549
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.6596701649175413,
|
| 3856 |
+
"grad_norm": 0.3542385399341583,
|
| 3857 |
+
"learning_rate": 3.2485727993198945e-05,
|
| 3858 |
+
"loss": 9.6182,
|
| 3859 |
+
"step": 550
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 0.6608695652173913,
|
| 3863 |
+
"grad_norm": 0.2821604907512665,
|
| 3864 |
+
"learning_rate": 3.228516977184075e-05,
|
| 3865 |
+
"loss": 9.6229,
|
| 3866 |
+
"step": 551
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 0.6620689655172414,
|
| 3870 |
+
"grad_norm": 0.27113085985183716,
|
| 3871 |
+
"learning_rate": 3.2084936958407805e-05,
|
| 3872 |
+
"loss": 9.6041,
|
| 3873 |
+
"step": 552
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 0.6632683658170915,
|
| 3877 |
+
"grad_norm": 0.26982516050338745,
|
| 3878 |
+
"learning_rate": 3.188503323102425e-05,
|
| 3879 |
+
"loss": 9.6084,
|
| 3880 |
+
"step": 553
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"epoch": 0.6644677661169416,
|
| 3884 |
+
"grad_norm": 0.2756569981575012,
|
| 3885 |
+
"learning_rate": 3.1685462261769105e-05,
|
| 3886 |
+
"loss": 9.6126,
|
| 3887 |
+
"step": 554
|
| 3888 |
+
},
|
| 3889 |
+
{
|
| 3890 |
+
"epoch": 0.6656671664167916,
|
| 3891 |
+
"grad_norm": 0.27629488706588745,
|
| 3892 |
+
"learning_rate": 3.1486227716608946e-05,
|
| 3893 |
+
"loss": 9.6056,
|
| 3894 |
+
"step": 555
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 0.6668665667166417,
|
| 3898 |
+
"grad_norm": 0.28036460280418396,
|
| 3899 |
+
"learning_rate": 3.128733325533047e-05,
|
| 3900 |
+
"loss": 9.6054,
|
| 3901 |
+
"step": 556
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 0.6680659670164918,
|
| 3905 |
+
"grad_norm": 0.27844056487083435,
|
| 3906 |
+
"learning_rate": 3.1088782531473266e-05,
|
| 3907 |
+
"loss": 9.6111,
|
| 3908 |
+
"step": 557
|
| 3909 |
+
},
|
| 3910 |
+
{
|
| 3911 |
+
"epoch": 0.6692653673163418,
|
| 3912 |
+
"grad_norm": 0.2862386405467987,
|
| 3913 |
+
"learning_rate": 3.089057919226277e-05,
|
| 3914 |
+
"loss": 9.612,
|
| 3915 |
+
"step": 558
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 0.6704647676161919,
|
| 3919 |
+
"grad_norm": 0.2859496474266052,
|
| 3920 |
+
"learning_rate": 3.069272687854322e-05,
|
| 3921 |
+
"loss": 9.6114,
|
| 3922 |
+
"step": 559
|
| 3923 |
+
},
|
| 3924 |
+
{
|
| 3925 |
+
"epoch": 0.671664167916042,
|
| 3926 |
+
"grad_norm": 0.28554123640060425,
|
| 3927 |
+
"learning_rate": 3.049522922471075e-05,
|
| 3928 |
+
"loss": 9.6105,
|
| 3929 |
+
"step": 560
|
| 3930 |
+
},
|
| 3931 |
+
{
|
| 3932 |
+
"epoch": 0.6728635682158921,
|
| 3933 |
+
"grad_norm": 0.30089861154556274,
|
| 3934 |
+
"learning_rate": 3.02980898586467e-05,
|
| 3935 |
+
"loss": 9.6205,
|
| 3936 |
+
"step": 561
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 0.6740629685157421,
|
| 3940 |
+
"grad_norm": 0.30331140756607056,
|
| 3941 |
+
"learning_rate": 3.0101312401650937e-05,
|
| 3942 |
+
"loss": 9.6158,
|
| 3943 |
+
"step": 562
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 0.6752623688155922,
|
| 3947 |
+
"grad_norm": 0.2732248902320862,
|
| 3948 |
+
"learning_rate": 2.9904900468375297e-05,
|
| 3949 |
+
"loss": 9.6064,
|
| 3950 |
+
"step": 563
|
| 3951 |
+
},
|
| 3952 |
+
{
|
| 3953 |
+
"epoch": 0.6764617691154423,
|
| 3954 |
+
"grad_norm": 0.27510005235671997,
|
| 3955 |
+
"learning_rate": 2.9708857666757246e-05,
|
| 3956 |
+
"loss": 9.6019,
|
| 3957 |
+
"step": 564
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 0.6776611694152923,
|
| 3961 |
+
"grad_norm": 0.27365824580192566,
|
| 3962 |
+
"learning_rate": 2.9513187597953607e-05,
|
| 3963 |
+
"loss": 9.5995,
|
| 3964 |
+
"step": 565
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 0.6788605697151424,
|
| 3968 |
+
"grad_norm": 0.2792357802391052,
|
| 3969 |
+
"learning_rate": 2.931789385627433e-05,
|
| 3970 |
+
"loss": 9.606,
|
| 3971 |
+
"step": 566
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 0.6800599700149925,
|
| 3975 |
+
"grad_norm": 0.2759556770324707,
|
| 3976 |
+
"learning_rate": 2.9122980029116586e-05,
|
| 3977 |
+
"loss": 9.6039,
|
| 3978 |
+
"step": 567
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 0.6812593703148426,
|
| 3982 |
+
"grad_norm": 0.2814030647277832,
|
| 3983 |
+
"learning_rate": 2.8928449696898763e-05,
|
| 3984 |
+
"loss": 9.602,
|
| 3985 |
+
"step": 568
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 0.6824587706146926,
|
| 3989 |
+
"grad_norm": 0.2769099771976471,
|
| 3990 |
+
"learning_rate": 2.8734306432994735e-05,
|
| 3991 |
+
"loss": 9.6079,
|
| 3992 |
+
"step": 569
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 0.6836581709145427,
|
| 3996 |
+
"grad_norm": 0.2809275686740875,
|
| 3997 |
+
"learning_rate": 2.8540553803668252e-05,
|
| 3998 |
+
"loss": 9.613,
|
| 3999 |
+
"step": 570
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 0.6848575712143928,
|
| 4003 |
+
"grad_norm": 0.275016725063324,
|
| 4004 |
+
"learning_rate": 2.8347195368007418e-05,
|
| 4005 |
+
"loss": 9.6097,
|
| 4006 |
+
"step": 571
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 0.6860569715142428,
|
| 4010 |
+
"grad_norm": 0.2964610755443573,
|
| 4011 |
+
"learning_rate": 2.815423467785925e-05,
|
| 4012 |
+
"loss": 9.6111,
|
| 4013 |
+
"step": 572
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 0.6872563718140929,
|
| 4017 |
+
"grad_norm": 0.2884480059146881,
|
| 4018 |
+
"learning_rate": 2.7961675277764498e-05,
|
| 4019 |
+
"loss": 9.6089,
|
| 4020 |
+
"step": 573
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 0.688455772113943,
|
| 4024 |
+
"grad_norm": 0.30310893058776855,
|
| 4025 |
+
"learning_rate": 2.7769520704892566e-05,
|
| 4026 |
+
"loss": 9.6102,
|
| 4027 |
+
"step": 574
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 0.6896551724137931,
|
| 4031 |
+
"grad_norm": 0.4733683466911316,
|
| 4032 |
+
"learning_rate": 2.757777448897646e-05,
|
| 4033 |
+
"loss": 9.6083,
|
| 4034 |
+
"step": 575
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 0.6908545727136431,
|
| 4038 |
+
"grad_norm": 0.272512823343277,
|
| 4039 |
+
"learning_rate": 2.7386440152247933e-05,
|
| 4040 |
+
"loss": 9.5963,
|
| 4041 |
+
"step": 576
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 0.6920539730134933,
|
| 4045 |
+
"grad_norm": 0.2810138165950775,
|
| 4046 |
+
"learning_rate": 2.71955212093729e-05,
|
| 4047 |
+
"loss": 9.6012,
|
| 4048 |
+
"step": 577
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 0.6932533733133434,
|
| 4052 |
+
"grad_norm": 0.2755623161792755,
|
| 4053 |
+
"learning_rate": 2.7005021167386803e-05,
|
| 4054 |
+
"loss": 9.6022,
|
| 4055 |
+
"step": 578
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 0.6944527736131934,
|
| 4059 |
+
"grad_norm": 0.2718299329280853,
|
| 4060 |
+
"learning_rate": 2.681494352563013e-05,
|
| 4061 |
+
"loss": 9.6096,
|
| 4062 |
+
"step": 579
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 0.6956521739130435,
|
| 4066 |
+
"grad_norm": 0.2746315896511078,
|
| 4067 |
+
"learning_rate": 2.6625291775684292e-05,
|
| 4068 |
+
"loss": 9.6124,
|
| 4069 |
+
"step": 580
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 0.6968515742128936,
|
| 4073 |
+
"grad_norm": 0.2844776511192322,
|
| 4074 |
+
"learning_rate": 2.6436069401307284e-05,
|
| 4075 |
+
"loss": 9.6054,
|
| 4076 |
+
"step": 581
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 0.6980509745127437,
|
| 4080 |
+
"grad_norm": 0.2785060703754425,
|
| 4081 |
+
"learning_rate": 2.624727987836991e-05,
|
| 4082 |
+
"loss": 9.6112,
|
| 4083 |
+
"step": 582
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 0.6992503748125937,
|
| 4087 |
+
"grad_norm": 0.2840147316455841,
|
| 4088 |
+
"learning_rate": 2.6058926674791728e-05,
|
| 4089 |
+
"loss": 9.6061,
|
| 4090 |
+
"step": 583
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 0.7004497751124438,
|
| 4094 |
+
"grad_norm": 0.28523436188697815,
|
| 4095 |
+
"learning_rate": 2.5871013250477528e-05,
|
| 4096 |
+
"loss": 9.6057,
|
| 4097 |
+
"step": 584
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.7016491754122939,
|
| 4101 |
+
"grad_norm": 0.29284006357192993,
|
| 4102 |
+
"learning_rate": 2.56835430572536e-05,
|
| 4103 |
+
"loss": 9.6091,
|
| 4104 |
+
"step": 585
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 0.7028485757121439,
|
| 4108 |
+
"grad_norm": 0.29574641585350037,
|
| 4109 |
+
"learning_rate": 2.5496519538804486e-05,
|
| 4110 |
+
"loss": 9.6155,
|
| 4111 |
+
"step": 586
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 0.704047976011994,
|
| 4115 |
+
"grad_norm": 0.3032572269439697,
|
| 4116 |
+
"learning_rate": 2.530994613060965e-05,
|
| 4117 |
+
"loss": 9.6162,
|
| 4118 |
+
"step": 587
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 0.7052473763118441,
|
| 4122 |
+
"grad_norm": 0.2718828320503235,
|
| 4123 |
+
"learning_rate": 2.5123826259880323e-05,
|
| 4124 |
+
"loss": 9.6001,
|
| 4125 |
+
"step": 588
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 0.7064467766116942,
|
| 4129 |
+
"grad_norm": 0.27074381709098816,
|
| 4130 |
+
"learning_rate": 2.493816334549664e-05,
|
| 4131 |
+
"loss": 9.6014,
|
| 4132 |
+
"step": 589
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 0.7076461769115442,
|
| 4136 |
+
"grad_norm": 0.2791549265384674,
|
| 4137 |
+
"learning_rate": 2.4752960797944802e-05,
|
| 4138 |
+
"loss": 9.5998,
|
| 4139 |
+
"step": 590
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.7088455772113943,
|
| 4143 |
+
"grad_norm": 0.28340011835098267,
|
| 4144 |
+
"learning_rate": 2.4568222019254377e-05,
|
| 4145 |
+
"loss": 9.5979,
|
| 4146 |
+
"step": 591
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 0.7100449775112444,
|
| 4150 |
+
"grad_norm": 0.2762751579284668,
|
| 4151 |
+
"learning_rate": 2.43839504029359e-05,
|
| 4152 |
+
"loss": 9.6032,
|
| 4153 |
+
"step": 592
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 0.7112443778110945,
|
| 4157 |
+
"grad_norm": 0.2753763198852539,
|
| 4158 |
+
"learning_rate": 2.4200149333918487e-05,
|
| 4159 |
+
"loss": 9.6089,
|
| 4160 |
+
"step": 593
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 0.7124437781109445,
|
| 4164 |
+
"grad_norm": 0.27482444047927856,
|
| 4165 |
+
"learning_rate": 2.4016822188487603e-05,
|
| 4166 |
+
"loss": 9.6081,
|
| 4167 |
+
"step": 594
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 0.7136431784107946,
|
| 4171 |
+
"grad_norm": 0.28210797905921936,
|
| 4172 |
+
"learning_rate": 2.383397233422318e-05,
|
| 4173 |
+
"loss": 9.6041,
|
| 4174 |
+
"step": 595
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 0.7148425787106447,
|
| 4178 |
+
"grad_norm": 0.2853706479072571,
|
| 4179 |
+
"learning_rate": 2.3651603129937592e-05,
|
| 4180 |
+
"loss": 9.6042,
|
| 4181 |
+
"step": 596
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 0.7160419790104947,
|
| 4185 |
+
"grad_norm": 0.3066234886646271,
|
| 4186 |
+
"learning_rate": 2.346971792561413e-05,
|
| 4187 |
+
"loss": 9.6053,
|
| 4188 |
+
"step": 597
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 0.7172413793103448,
|
| 4192 |
+
"grad_norm": 0.2879929542541504,
|
| 4193 |
+
"learning_rate": 2.3288320062345277e-05,
|
| 4194 |
+
"loss": 9.6069,
|
| 4195 |
+
"step": 598
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 0.7184407796101949,
|
| 4199 |
+
"grad_norm": 0.35332369804382324,
|
| 4200 |
+
"learning_rate": 2.3107412872271518e-05,
|
| 4201 |
+
"loss": 9.6162,
|
| 4202 |
+
"step": 599
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 0.719640179910045,
|
| 4206 |
+
"grad_norm": 0.5152252316474915,
|
| 4207 |
+
"learning_rate": 2.2926999678519974e-05,
|
| 4208 |
+
"loss": 9.6182,
|
| 4209 |
+
"step": 600
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 0.720839580209895,
|
| 4213 |
+
"grad_norm": 0.2663346230983734,
|
| 4214 |
+
"learning_rate": 2.274708379514348e-05,
|
| 4215 |
+
"loss": 9.5986,
|
| 4216 |
+
"step": 601
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.7220389805097451,
|
| 4220 |
+
"grad_norm": 0.27524423599243164,
|
| 4221 |
+
"learning_rate": 2.256766852705967e-05,
|
| 4222 |
+
"loss": 9.5986,
|
| 4223 |
+
"step": 602
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 0.7232383808095952,
|
| 4227 |
+
"grad_norm": 0.2814219295978546,
|
| 4228 |
+
"learning_rate": 2.238875716999019e-05,
|
| 4229 |
+
"loss": 9.6037,
|
| 4230 |
+
"step": 603
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 0.7244377811094452,
|
| 4234 |
+
"grad_norm": 0.2859136760234833,
|
| 4235 |
+
"learning_rate": 2.221035301040027e-05,
|
| 4236 |
+
"loss": 9.6002,
|
| 4237 |
+
"step": 604
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 0.7256371814092953,
|
| 4241 |
+
"grad_norm": 0.27460747957229614,
|
| 4242 |
+
"learning_rate": 2.2032459325438336e-05,
|
| 4243 |
+
"loss": 9.6031,
|
| 4244 |
+
"step": 605
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 0.7268365817091454,
|
| 4248 |
+
"grad_norm": 0.2745445966720581,
|
| 4249 |
+
"learning_rate": 2.185507938287572e-05,
|
| 4250 |
+
"loss": 9.6072,
|
| 4251 |
+
"step": 606
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 0.7280359820089956,
|
| 4255 |
+
"grad_norm": 0.2816024124622345,
|
| 4256 |
+
"learning_rate": 2.1678216441046734e-05,
|
| 4257 |
+
"loss": 9.6128,
|
| 4258 |
+
"step": 607
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.7292353823088455,
|
| 4262 |
+
"grad_norm": 0.28734058141708374,
|
| 4263 |
+
"learning_rate": 2.1501873748788802e-05,
|
| 4264 |
+
"loss": 9.6127,
|
| 4265 |
+
"step": 608
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 0.7304347826086957,
|
| 4269 |
+
"grad_norm": 0.28445249795913696,
|
| 4270 |
+
"learning_rate": 2.1326054545382695e-05,
|
| 4271 |
+
"loss": 9.6118,
|
| 4272 |
+
"step": 609
|
| 4273 |
+
},
|
| 4274 |
+
{
|
| 4275 |
+
"epoch": 0.7316341829085458,
|
| 4276 |
+
"grad_norm": 0.2825443148612976,
|
| 4277 |
+
"learning_rate": 2.1150762060493155e-05,
|
| 4278 |
+
"loss": 9.6182,
|
| 4279 |
+
"step": 610
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 0.7328335832083958,
|
| 4283 |
+
"grad_norm": 0.29409319162368774,
|
| 4284 |
+
"learning_rate": 2.09759995141095e-05,
|
| 4285 |
+
"loss": 9.611,
|
| 4286 |
+
"step": 611
|
| 4287 |
+
},
|
| 4288 |
+
{
|
| 4289 |
+
"epoch": 0.7340329835082459,
|
| 4290 |
+
"grad_norm": 0.30348506569862366,
|
| 4291 |
+
"learning_rate": 2.0801770116486447e-05,
|
| 4292 |
+
"loss": 9.6193,
|
| 4293 |
+
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 0.735232383808096,
|
| 4297 |
+
"grad_norm": 0.2586905360221863,
|
| 4298 |
+
"learning_rate": 2.0628077068085173e-05,
|
| 4299 |
+
"loss": 9.6146,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 0.7364317841079461,
|
| 4304 |
+
"grad_norm": 0.27243587374687195,
|
| 4305 |
+
"learning_rate": 2.0454923559514595e-05,
|
| 4306 |
+
"loss": 9.6025,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 0.7376311844077961,
|
| 4311 |
+
"grad_norm": 0.27491042017936707,
|
| 4312 |
+
"learning_rate": 2.028231277147261e-05,
|
| 4313 |
+
"loss": 9.6013,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 0.7388305847076462,
|
| 4318 |
+
"grad_norm": 0.279153048992157,
|
| 4319 |
+
"learning_rate": 2.0110247874687815e-05,
|
| 4320 |
+
"loss": 9.5937,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 0.7400299850074963,
|
| 4325 |
+
"grad_norm": 0.27780649065971375,
|
| 4326 |
+
"learning_rate": 1.993873202986119e-05,
|
| 4327 |
+
"loss": 9.6022,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 0.7412293853073463,
|
| 4332 |
+
"grad_norm": 0.2798539698123932,
|
| 4333 |
+
"learning_rate": 1.976776838760801e-05,
|
| 4334 |
+
"loss": 9.6022,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 0.7424287856071964,
|
| 4339 |
+
"grad_norm": 0.27843162417411804,
|
| 4340 |
+
"learning_rate": 1.9597360088400052e-05,
|
| 4341 |
+
"loss": 9.6062,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 0.7436281859070465,
|
| 4346 |
+
"grad_norm": 0.27371302247047424,
|
| 4347 |
+
"learning_rate": 1.9427510262507864e-05,
|
| 4348 |
+
"loss": 9.6119,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 0.7448275862068966,
|
| 4353 |
+
"grad_norm": 0.2873663604259491,
|
| 4354 |
+
"learning_rate": 1.925822202994323e-05,
|
| 4355 |
+
"loss": 9.6004,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 0.7460269865067466,
|
| 4360 |
+
"grad_norm": 0.2875591218471527,
|
| 4361 |
+
"learning_rate": 1.9089498500401914e-05,
|
| 4362 |
+
"loss": 9.6119,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 0.7472263868065967,
|
| 4367 |
+
"grad_norm": 0.2853778004646301,
|
| 4368 |
+
"learning_rate": 1.892134277320655e-05,
|
| 4369 |
+
"loss": 9.6091,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 0.7484257871064468,
|
| 4374 |
+
"grad_norm": 0.2952004075050354,
|
| 4375 |
+
"learning_rate": 1.87537579372496e-05,
|
| 4376 |
+
"loss": 9.6182,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 0.7496251874062968,
|
| 4381 |
+
"grad_norm": 0.3686712980270386,
|
| 4382 |
+
"learning_rate": 1.858674707093675e-05,
|
| 4383 |
+
"loss": 9.614,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 0.7508245877061469,
|
| 4388 |
+
"grad_norm": 0.2664184868335724,
|
| 4389 |
+
"learning_rate": 1.8420313242130293e-05,
|
| 4390 |
+
"loss": 9.6005,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 0.752023988005997,
|
| 4395 |
+
"grad_norm": 0.2688407301902771,
|
| 4396 |
+
"learning_rate": 1.8254459508092768e-05,
|
| 4397 |
+
"loss": 9.5988,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 0.7532233883058471,
|
| 4402 |
+
"grad_norm": 0.2794104516506195,
|
| 4403 |
+
"learning_rate": 1.8089188915430793e-05,
|
| 4404 |
+
"loss": 9.5987,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 0.7544227886056971,
|
| 4409 |
+
"grad_norm": 0.26486334204673767,
|
| 4410 |
+
"learning_rate": 1.792450450003919e-05,
|
| 4411 |
+
"loss": 9.6129,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 0.7556221889055472,
|
| 4416 |
+
"grad_norm": 0.2762359082698822,
|
| 4417 |
+
"learning_rate": 1.7760409287045078e-05,
|
| 4418 |
+
"loss": 9.6052,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 0.7568215892053973,
|
| 4423 |
+
"grad_norm": 0.27764591574668884,
|
| 4424 |
+
"learning_rate": 1.7596906290752425e-05,
|
| 4425 |
+
"loss": 9.6056,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 0.7580209895052473,
|
| 4430 |
+
"grad_norm": 0.276153028011322,
|
| 4431 |
+
"learning_rate": 1.743399851458663e-05,
|
| 4432 |
+
"loss": 9.609,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 0.7592203898050974,
|
| 4437 |
+
"grad_norm": 0.2780199646949768,
|
| 4438 |
+
"learning_rate": 1.727168895103931e-05,
|
| 4439 |
+
"loss": 9.6081,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 0.7604197901049475,
|
| 4444 |
+
"grad_norm": 0.276457816362381,
|
| 4445 |
+
"learning_rate": 1.7109980581613417e-05,
|
| 4446 |
+
"loss": 9.6062,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 0.7616191904047976,
|
| 4451 |
+
"grad_norm": 0.2808220088481903,
|
| 4452 |
+
"learning_rate": 1.6948876376768418e-05,
|
| 4453 |
+
"loss": 9.6123,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 0.7628185907046476,
|
| 4458 |
+
"grad_norm": 0.29566583037376404,
|
| 4459 |
+
"learning_rate": 1.6788379295865704e-05,
|
| 4460 |
+
"loss": 9.6094,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 0.7640179910044977,
|
| 4465 |
+
"grad_norm": 0.33136534690856934,
|
| 4466 |
+
"learning_rate": 1.6628492287114296e-05,
|
| 4467 |
+
"loss": 9.614,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 0.7652173913043478,
|
| 4472 |
+
"grad_norm": 0.27251994609832764,
|
| 4473 |
+
"learning_rate": 1.6469218287516664e-05,
|
| 4474 |
+
"loss": 9.6011,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 0.766416791604198,
|
| 4479 |
+
"grad_norm": 0.2670121490955353,
|
| 4480 |
+
"learning_rate": 1.6310560222814714e-05,
|
| 4481 |
+
"loss": 9.6037,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 0.767616191904048,
|
| 4486 |
+
"grad_norm": 0.2792399227619171,
|
| 4487 |
+
"learning_rate": 1.6152521007436145e-05,
|
| 4488 |
+
"loss": 9.6036,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 0.7688155922038981,
|
| 4493 |
+
"grad_norm": 0.275511234998703,
|
| 4494 |
+
"learning_rate": 1.599510354444087e-05,
|
| 4495 |
+
"loss": 9.5973,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 0.7700149925037482,
|
| 4500 |
+
"grad_norm": 0.2751782536506653,
|
| 4501 |
+
"learning_rate": 1.5838310725467644e-05,
|
| 4502 |
+
"loss": 9.6005,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.7712143928035982,
|
| 4507 |
+
"grad_norm": 0.28111734986305237,
|
| 4508 |
+
"learning_rate": 1.5682145430681027e-05,
|
| 4509 |
+
"loss": 9.6015,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 0.7724137931034483,
|
| 4514 |
+
"grad_norm": 0.2826797068119049,
|
| 4515 |
+
"learning_rate": 1.5526610528718415e-05,
|
| 4516 |
+
"loss": 9.6054,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 0.7736131934032984,
|
| 4521 |
+
"grad_norm": 0.28505128622055054,
|
| 4522 |
+
"learning_rate": 1.5371708876637354e-05,
|
| 4523 |
+
"loss": 9.6042,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 0.7748125937031485,
|
| 4528 |
+
"grad_norm": 0.28200674057006836,
|
| 4529 |
+
"learning_rate": 1.5217443319863112e-05,
|
| 4530 |
+
"loss": 9.6051,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 0.7760119940029985,
|
| 4535 |
+
"grad_norm": 0.2859637439250946,
|
| 4536 |
+
"learning_rate": 1.5063816692136373e-05,
|
| 4537 |
+
"loss": 9.6004,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 0.7772113943028486,
|
| 4542 |
+
"grad_norm": 0.28504401445388794,
|
| 4543 |
+
"learning_rate": 1.4910831815461123e-05,
|
| 4544 |
+
"loss": 9.6177,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 0.7784107946026987,
|
| 4549 |
+
"grad_norm": 0.2949487268924713,
|
| 4550 |
+
"learning_rate": 1.4758491500052924e-05,
|
| 4551 |
+
"loss": 9.6204,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 0.7796101949025487,
|
| 4556 |
+
"grad_norm": 0.3952041268348694,
|
| 4557 |
+
"learning_rate": 1.4606798544287243e-05,
|
| 4558 |
+
"loss": 9.62,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 0.7808095952023988,
|
| 4563 |
+
"grad_norm": 0.2684868574142456,
|
| 4564 |
+
"learning_rate": 1.445575573464799e-05,
|
| 4565 |
+
"loss": 9.5986,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 0.7820089955022489,
|
| 4570 |
+
"grad_norm": 0.2751760184764862,
|
| 4571 |
+
"learning_rate": 1.4305365845676439e-05,
|
| 4572 |
+
"loss": 9.5993,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 0.783208395802099,
|
| 4577 |
+
"grad_norm": 0.27565452456474304,
|
| 4578 |
+
"learning_rate": 1.4155631639920209e-05,
|
| 4579 |
+
"loss": 9.5939,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 0.784407796101949,
|
| 4584 |
+
"grad_norm": 0.27967387437820435,
|
| 4585 |
+
"learning_rate": 1.4006555867882464e-05,
|
| 4586 |
+
"loss": 9.6024,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 0.7856071964017991,
|
| 4591 |
+
"grad_norm": 0.28178393840789795,
|
| 4592 |
+
"learning_rate": 1.3858141267971491e-05,
|
| 4593 |
+
"loss": 9.6057,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 0.7868065967016492,
|
| 4598 |
+
"grad_norm": 0.27983683347702026,
|
| 4599 |
+
"learning_rate": 1.3710390566450366e-05,
|
| 4600 |
+
"loss": 9.6059,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 0.7880059970014992,
|
| 4605 |
+
"grad_norm": 0.286726713180542,
|
| 4606 |
+
"learning_rate": 1.3563306477386784e-05,
|
| 4607 |
+
"loss": 9.6032,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 0.7892053973013493,
|
| 4612 |
+
"grad_norm": 0.2814926505088806,
|
| 4613 |
+
"learning_rate": 1.3416891702603358e-05,
|
| 4614 |
+
"loss": 9.6077,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 0.7904047976011994,
|
| 4619 |
+
"grad_norm": 0.291660875082016,
|
| 4620 |
+
"learning_rate": 1.3271148931627858e-05,
|
| 4621 |
+
"loss": 9.6055,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 0.7916041979010495,
|
| 4626 |
+
"grad_norm": 0.2863795757293701,
|
| 4627 |
+
"learning_rate": 1.3126080841643856e-05,
|
| 4628 |
+
"loss": 9.6111,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 0.7928035982008995,
|
| 4633 |
+
"grad_norm": 0.2854698896408081,
|
| 4634 |
+
"learning_rate": 1.2981690097441573e-05,
|
| 4635 |
+
"loss": 9.6172,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 0.7940029985007496,
|
| 4640 |
+
"grad_norm": 0.3119170367717743,
|
| 4641 |
+
"learning_rate": 1.2837979351368912e-05,
|
| 4642 |
+
"loss": 9.6102,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 0.7952023988005997,
|
| 4647 |
+
"grad_norm": 0.27526015043258667,
|
| 4648 |
+
"learning_rate": 1.2694951243282683e-05,
|
| 4649 |
+
"loss": 9.6006,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 0.7964017991004497,
|
| 4654 |
+
"grad_norm": 0.27086350321769714,
|
| 4655 |
+
"learning_rate": 1.2552608400500199e-05,
|
| 4656 |
+
"loss": 9.6,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 0.7976011994002998,
|
| 4661 |
+
"grad_norm": 0.2674426734447479,
|
| 4662 |
+
"learning_rate": 1.2410953437750966e-05,
|
| 4663 |
+
"loss": 9.599,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 0.7988005997001499,
|
| 4668 |
+
"grad_norm": 0.26960834860801697,
|
| 4669 |
+
"learning_rate": 1.2269988957128636e-05,
|
| 4670 |
+
"loss": 9.6059,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 0.8,
|
| 4675 |
+
"grad_norm": 0.27745890617370605,
|
| 4676 |
+
"learning_rate": 1.212971754804324e-05,
|
| 4677 |
+
"loss": 9.6046,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 0.80119940029985,
|
| 4682 |
+
"grad_norm": 0.2803892493247986,
|
| 4683 |
+
"learning_rate": 1.1990141787173648e-05,
|
| 4684 |
+
"loss": 9.6036,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 0.8023988005997001,
|
| 4689 |
+
"grad_norm": 0.2826705574989319,
|
| 4690 |
+
"learning_rate": 1.1851264238420135e-05,
|
| 4691 |
+
"loss": 9.6031,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 0.8035982008995503,
|
| 4696 |
+
"grad_norm": 0.28543218970298767,
|
| 4697 |
+
"learning_rate": 1.1713087452857408e-05,
|
| 4698 |
+
"loss": 9.6047,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 0.8047976011994002,
|
| 4703 |
+
"grad_norm": 0.2749161124229431,
|
| 4704 |
+
"learning_rate": 1.1575613968687682e-05,
|
| 4705 |
+
"loss": 9.6061,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 0.8059970014992504,
|
| 4710 |
+
"grad_norm": 0.2880239486694336,
|
| 4711 |
+
"learning_rate": 1.1438846311194024e-05,
|
| 4712 |
+
"loss": 9.607,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 0.8071964017991005,
|
| 4717 |
+
"grad_norm": 0.2794909179210663,
|
| 4718 |
+
"learning_rate": 1.1302786992694048e-05,
|
| 4719 |
+
"loss": 9.6098,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 0.8083958020989506,
|
| 4724 |
+
"grad_norm": 0.3027547299861908,
|
| 4725 |
+
"learning_rate": 1.1167438512493683e-05,
|
| 4726 |
+
"loss": 9.6116,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 0.8095952023988006,
|
| 4731 |
+
"grad_norm": 0.34445720911026,
|
| 4732 |
+
"learning_rate": 1.1032803356841342e-05,
|
| 4733 |
+
"loss": 9.6171,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 0.8107946026986507,
|
| 4738 |
+
"grad_norm": 0.2722347378730774,
|
| 4739 |
+
"learning_rate": 1.0898883998882158e-05,
|
| 4740 |
+
"loss": 9.601,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 0.8119940029985008,
|
| 4745 |
+
"grad_norm": 0.27299922704696655,
|
| 4746 |
+
"learning_rate": 1.0765682898612656e-05,
|
| 4747 |
+
"loss": 9.5976,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 0.8131934032983508,
|
| 4752 |
+
"grad_norm": 0.2737182080745697,
|
| 4753 |
+
"learning_rate": 1.0633202502835494e-05,
|
| 4754 |
+
"loss": 9.5965,
|
| 4755 |
+
"step": 678
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 0.8143928035982009,
|
| 4759 |
+
"grad_norm": 0.2752780020236969,
|
| 4760 |
+
"learning_rate": 1.0501445245114522e-05,
|
| 4761 |
+
"loss": 9.6009,
|
| 4762 |
+
"step": 679
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 0.815592203898051,
|
| 4766 |
+
"grad_norm": 0.2721465826034546,
|
| 4767 |
+
"learning_rate": 1.0370413545730118e-05,
|
| 4768 |
+
"loss": 9.6064,
|
| 4769 |
+
"step": 680
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 0.8167916041979011,
|
| 4773 |
+
"grad_norm": 0.2846396267414093,
|
| 4774 |
+
"learning_rate": 1.0240109811634712e-05,
|
| 4775 |
+
"loss": 9.5995,
|
| 4776 |
+
"step": 681
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 0.8179910044977511,
|
| 4780 |
+
"grad_norm": 0.28411293029785156,
|
| 4781 |
+
"learning_rate": 1.0110536436408535e-05,
|
| 4782 |
+
"loss": 9.5975,
|
| 4783 |
+
"step": 682
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 0.8191904047976012,
|
| 4787 |
+
"grad_norm": 0.2815098762512207,
|
| 4788 |
+
"learning_rate": 9.9816958002157e-06,
|
| 4789 |
+
"loss": 9.6078,
|
| 4790 |
+
"step": 683
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 0.8203898050974513,
|
| 4794 |
+
"grad_norm": 0.278131902217865,
|
| 4795 |
+
"learning_rate": 9.853590269760493e-06,
|
| 4796 |
+
"loss": 9.6143,
|
| 4797 |
+
"step": 684
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 0.8215892053973014,
|
| 4801 |
+
"grad_norm": 0.2930939197540283,
|
| 4802 |
+
"learning_rate": 9.726222198243806e-06,
|
| 4803 |
+
"loss": 9.6042,
|
| 4804 |
+
"step": 685
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 0.8227886056971514,
|
| 4808 |
+
"grad_norm": 0.2876308560371399,
|
| 4809 |
+
"learning_rate": 9.599593925320016e-06,
|
| 4810 |
+
"loss": 9.6187,
|
| 4811 |
+
"step": 686
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 0.8239880059970015,
|
| 4815 |
+
"grad_norm": 0.3398456573486328,
|
| 4816 |
+
"learning_rate": 9.47370777705397e-06,
|
| 4817 |
+
"loss": 9.6115,
|
| 4818 |
+
"step": 687
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 0.8251874062968516,
|
| 4822 |
+
"grad_norm": 0.28324592113494873,
|
| 4823 |
+
"learning_rate": 9.348566065878217e-06,
|
| 4824 |
+
"loss": 9.5972,
|
| 4825 |
+
"step": 688
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 0.8263868065967016,
|
| 4829 |
+
"grad_norm": 0.271178662776947,
|
| 4830 |
+
"learning_rate": 9.224171090550571e-06,
|
| 4831 |
+
"loss": 9.6004,
|
| 4832 |
+
"step": 689
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 0.8275862068965517,
|
| 4836 |
+
"grad_norm": 0.26743438839912415,
|
| 4837 |
+
"learning_rate": 9.100525136111915e-06,
|
| 4838 |
+
"loss": 9.604,
|
| 4839 |
+
"step": 690
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 0.8287856071964018,
|
| 4843 |
+
"grad_norm": 0.2741158604621887,
|
| 4844 |
+
"learning_rate": 8.97763047384414e-06,
|
| 4845 |
+
"loss": 9.6024,
|
| 4846 |
+
"step": 691
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 0.8299850074962519,
|
| 4850 |
+
"grad_norm": 0.2776412069797516,
|
| 4851 |
+
"learning_rate": 8.855489361228496e-06,
|
| 4852 |
+
"loss": 9.5996,
|
| 4853 |
+
"step": 692
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 0.8311844077961019,
|
| 4857 |
+
"grad_norm": 0.2762274742126465,
|
| 4858 |
+
"learning_rate": 8.734104041904129e-06,
|
| 4859 |
+
"loss": 9.6041,
|
| 4860 |
+
"step": 693
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 0.832383808095952,
|
| 4864 |
+
"grad_norm": 0.2758176624774933,
|
| 4865 |
+
"learning_rate": 8.61347674562677e-06,
|
| 4866 |
+
"loss": 9.6084,
|
| 4867 |
+
"step": 694
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 0.8335832083958021,
|
| 4871 |
+
"grad_norm": 0.28230342268943787,
|
| 4872 |
+
"learning_rate": 8.4936096882279e-06,
|
| 4873 |
+
"loss": 9.6047,
|
| 4874 |
+
"step": 695
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 0.8347826086956521,
|
| 4878 |
+
"grad_norm": 0.28801631927490234,
|
| 4879 |
+
"learning_rate": 8.37450507157399e-06,
|
| 4880 |
+
"loss": 9.6084,
|
| 4881 |
+
"step": 696
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 0.8359820089955022,
|
| 4885 |
+
"grad_norm": 0.289760559797287,
|
| 4886 |
+
"learning_rate": 8.256165083526019e-06,
|
| 4887 |
+
"loss": 9.6033,
|
| 4888 |
+
"step": 697
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 0.8371814092953523,
|
| 4892 |
+
"grad_norm": 0.29011571407318115,
|
| 4893 |
+
"learning_rate": 8.138591897899345e-06,
|
| 4894 |
+
"loss": 9.6161,
|
| 4895 |
+
"step": 698
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 0.8383808095952024,
|
| 4899 |
+
"grad_norm": 0.3083633780479431,
|
| 4900 |
+
"learning_rate": 8.021787674423775e-06,
|
| 4901 |
+
"loss": 9.6152,
|
| 4902 |
+
"step": 699
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 0.8395802098950524,
|
| 4906 |
+
"grad_norm": 0.36470380425453186,
|
| 4907 |
+
"learning_rate": 7.905754558703803e-06,
|
| 4908 |
+
"loss": 9.6132,
|
| 4909 |
+
"step": 700
|
| 4910 |
+
},
|
| 4911 |
+
{
|
| 4912 |
+
"epoch": 0.8407796101949025,
|
| 4913 |
+
"grad_norm": 0.26850220561027527,
|
| 4914 |
+
"learning_rate": 7.790494682179317e-06,
|
| 4915 |
+
"loss": 9.5949,
|
| 4916 |
+
"step": 701
|
| 4917 |
+
},
|
| 4918 |
+
{
|
| 4919 |
+
"epoch": 0.8419790104947527,
|
| 4920 |
+
"grad_norm": 0.2714633643627167,
|
| 4921 |
+
"learning_rate": 7.676010162086388e-06,
|
| 4922 |
+
"loss": 9.604,
|
| 4923 |
+
"step": 702
|
| 4924 |
+
},
|
| 4925 |
+
{
|
| 4926 |
+
"epoch": 0.8431784107946027,
|
| 4927 |
+
"grad_norm": 0.2753824293613434,
|
| 4928 |
+
"learning_rate": 7.56230310141835e-06,
|
| 4929 |
+
"loss": 9.5993,
|
| 4930 |
+
"step": 703
|
| 4931 |
+
},
|
| 4932 |
+
{
|
| 4933 |
+
"epoch": 0.8443778110944528,
|
| 4934 |
+
"grad_norm": 0.2757047116756439,
|
| 4935 |
+
"learning_rate": 7.449375588887203e-06,
|
| 4936 |
+
"loss": 9.5993,
|
| 4937 |
+
"step": 704
|
| 4938 |
+
},
|
| 4939 |
+
{
|
| 4940 |
+
"epoch": 0.8455772113943029,
|
| 4941 |
+
"grad_norm": 0.27331098914146423,
|
| 4942 |
+
"learning_rate": 7.337229698885279e-06,
|
| 4943 |
+
"loss": 9.6088,
|
| 4944 |
+
"step": 705
|
| 4945 |
+
},
|
| 4946 |
+
{
|
| 4947 |
+
"epoch": 0.846776611694153,
|
| 4948 |
+
"grad_norm": 0.2818980813026428,
|
| 4949 |
+
"learning_rate": 7.225867491447053e-06,
|
| 4950 |
+
"loss": 9.6,
|
| 4951 |
+
"step": 706
|
| 4952 |
+
},
|
| 4953 |
+
{
|
| 4954 |
+
"epoch": 0.847976011994003,
|
| 4955 |
+
"grad_norm": 0.2784759998321533,
|
| 4956 |
+
"learning_rate": 7.115291012211383e-06,
|
| 4957 |
+
"loss": 9.6056,
|
| 4958 |
+
"step": 707
|
| 4959 |
+
},
|
| 4960 |
+
{
|
| 4961 |
+
"epoch": 0.8491754122938531,
|
| 4962 |
+
"grad_norm": 0.2809768319129944,
|
| 4963 |
+
"learning_rate": 7.005502292383898e-06,
|
| 4964 |
+
"loss": 9.6092,
|
| 4965 |
+
"step": 708
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.8503748125937032,
|
| 4969 |
+
"grad_norm": 0.29430076479911804,
|
| 4970 |
+
"learning_rate": 6.896503348699657e-06,
|
| 4971 |
+
"loss": 9.6031,
|
| 4972 |
+
"step": 709
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.8515742128935532,
|
| 4976 |
+
"grad_norm": 0.28350192308425903,
|
| 4977 |
+
"learning_rate": 6.788296183386162e-06,
|
| 4978 |
+
"loss": 9.6105,
|
| 4979 |
+
"step": 710
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.8527736131934033,
|
| 4983 |
+
"grad_norm": 0.29121461510658264,
|
| 4984 |
+
"learning_rate": 6.680882784126552e-06,
|
| 4985 |
+
"loss": 9.6108,
|
| 4986 |
+
"step": 711
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.8539730134932534,
|
| 4990 |
+
"grad_norm": 0.3215639889240265,
|
| 4991 |
+
"learning_rate": 6.5742651240230545e-06,
|
| 4992 |
+
"loss": 9.6104,
|
| 4993 |
+
"step": 712
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.8551724137931035,
|
| 4997 |
+
"grad_norm": 0.27074047923088074,
|
| 4998 |
+
"learning_rate": 6.46844516156081e-06,
|
| 4999 |
+
"loss": 9.598,
|
| 5000 |
+
"step": 713
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.8563718140929535,
|
| 5004 |
+
"grad_norm": 0.2728975713253021,
|
| 5005 |
+
"learning_rate": 6.363424840571869e-06,
|
| 5006 |
+
"loss": 9.5965,
|
| 5007 |
+
"step": 714
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.8575712143928036,
|
| 5011 |
+
"grad_norm": 0.2756417393684387,
|
| 5012 |
+
"learning_rate": 6.259206090199426e-06,
|
| 5013 |
+
"loss": 9.6021,
|
| 5014 |
+
"step": 715
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.8587706146926537,
|
| 5018 |
+
"grad_norm": 0.28334730863571167,
|
| 5019 |
+
"learning_rate": 6.155790824862484e-06,
|
| 5020 |
+
"loss": 9.5923,
|
| 5021 |
+
"step": 716
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.8599700149925037,
|
| 5025 |
+
"grad_norm": 0.2780725359916687,
|
| 5026 |
+
"learning_rate": 6.053180944220627e-06,
|
| 5027 |
+
"loss": 9.5977,
|
| 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
|