Training in progress, step 773, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 80792096
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b9b64dc8416b708564ab9d59be261d9ecadfa40201427399c2e7f5c5a4c8fc14
|
| 3 |
size 80792096
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 41460084
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3bd2e8eb393e003fbe4e09c5774f39ecf0ce15f5be5a1d82cb0adc4ea59c9c6b
|
| 3 |
size 41460084
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21dbde94f8edaf47214de2b254032cbe2979740ac38bbb264a3164210c85320d
|
| 3 |
size 14244
|
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 sha256:45719461b132f33086549160c158249554ace1a4b2303f17dcf1d7d8dd8db700
|
| 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":
|
| 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,
|
|
@@ -4081,6 +4081,1343 @@
|
|
| 4081 |
"learning_rate": 1.4500332050779386e-05,
|
| 4082 |
"loss": 1.7656,
|
| 4083 |
"step": 582
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4084 |
}
|
| 4085 |
],
|
| 4086 |
"logging_steps": 1,
|
|
@@ -4095,12 +5432,12 @@
|
|
| 4095 |
"should_evaluate": false,
|
| 4096 |
"should_log": false,
|
| 4097 |
"should_save": true,
|
| 4098 |
-
"should_training_stop":
|
| 4099 |
},
|
| 4100 |
"attributes": {}
|
| 4101 |
}
|
| 4102 |
},
|
| 4103 |
-
"total_flos":
|
| 4104 |
"train_batch_size": 4,
|
| 4105 |
"trial_name": null,
|
| 4106 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 773,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4081 |
"learning_rate": 1.4500332050779386e-05,
|
| 4082 |
"loss": 1.7656,
|
| 4083 |
"step": 582
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 0.7542043984476067,
|
| 4087 |
+
"grad_norm": 0.6377444863319397,
|
| 4088 |
+
"learning_rate": 1.4356597180781018e-05,
|
| 4089 |
+
"loss": 1.8165,
|
| 4090 |
+
"step": 583
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 0.7554980595084088,
|
| 4094 |
+
"grad_norm": 0.6441891193389893,
|
| 4095 |
+
"learning_rate": 1.4213458735809072e-05,
|
| 4096 |
+
"loss": 1.7498,
|
| 4097 |
+
"step": 584
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.7567917205692108,
|
| 4101 |
+
"grad_norm": 0.6479103565216064,
|
| 4102 |
+
"learning_rate": 1.4070919111015096e-05,
|
| 4103 |
+
"loss": 2.0357,
|
| 4104 |
+
"step": 585
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 0.7580853816300129,
|
| 4108 |
+
"grad_norm": 0.6354538798332214,
|
| 4109 |
+
"learning_rate": 1.3928980691530514e-05,
|
| 4110 |
+
"loss": 1.5854,
|
| 4111 |
+
"step": 586
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 0.759379042690815,
|
| 4115 |
+
"grad_norm": 0.697376549243927,
|
| 4116 |
+
"learning_rate": 1.3787645852426661e-05,
|
| 4117 |
+
"loss": 1.9868,
|
| 4118 |
+
"step": 587
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 0.7606727037516171,
|
| 4122 |
+
"grad_norm": 0.6565506458282471,
|
| 4123 |
+
"learning_rate": 1.364691695867515e-05,
|
| 4124 |
+
"loss": 1.8827,
|
| 4125 |
+
"step": 588
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 0.7619663648124192,
|
| 4129 |
+
"grad_norm": 0.6202589869499207,
|
| 4130 |
+
"learning_rate": 1.3506796365108232e-05,
|
| 4131 |
+
"loss": 1.8662,
|
| 4132 |
+
"step": 589
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 0.7632600258732212,
|
| 4136 |
+
"grad_norm": 0.7095828056335449,
|
| 4137 |
+
"learning_rate": 1.3367286416379365e-05,
|
| 4138 |
+
"loss": 1.9568,
|
| 4139 |
+
"step": 590
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.7645536869340233,
|
| 4143 |
+
"grad_norm": 0.7279965877532959,
|
| 4144 |
+
"learning_rate": 1.3228389446924072e-05,
|
| 4145 |
+
"loss": 1.922,
|
| 4146 |
+
"step": 591
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 0.7658473479948253,
|
| 4150 |
+
"grad_norm": 0.6959128379821777,
|
| 4151 |
+
"learning_rate": 1.3090107780920807e-05,
|
| 4152 |
+
"loss": 1.9386,
|
| 4153 |
+
"step": 592
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 0.7671410090556274,
|
| 4157 |
+
"grad_norm": 0.6932267546653748,
|
| 4158 |
+
"learning_rate": 1.2952443732252057e-05,
|
| 4159 |
+
"loss": 1.8064,
|
| 4160 |
+
"step": 593
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 0.7684346701164295,
|
| 4164 |
+
"grad_norm": 0.7042782306671143,
|
| 4165 |
+
"learning_rate": 1.2815399604465672e-05,
|
| 4166 |
+
"loss": 1.9048,
|
| 4167 |
+
"step": 594
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 0.7697283311772316,
|
| 4171 |
+
"grad_norm": 0.7183865308761597,
|
| 4172 |
+
"learning_rate": 1.267897769073631e-05,
|
| 4173 |
+
"loss": 1.8095,
|
| 4174 |
+
"step": 595
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 0.7710219922380336,
|
| 4178 |
+
"grad_norm": 0.7274515628814697,
|
| 4179 |
+
"learning_rate": 1.2543180273827043e-05,
|
| 4180 |
+
"loss": 1.8595,
|
| 4181 |
+
"step": 596
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 0.7723156532988357,
|
| 4185 |
+
"grad_norm": 0.8378881216049194,
|
| 4186 |
+
"learning_rate": 1.2408009626051137e-05,
|
| 4187 |
+
"loss": 1.9941,
|
| 4188 |
+
"step": 597
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 0.7736093143596378,
|
| 4192 |
+
"grad_norm": 0.8286713361740112,
|
| 4193 |
+
"learning_rate": 1.2273468009234095e-05,
|
| 4194 |
+
"loss": 1.9748,
|
| 4195 |
+
"step": 598
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 0.7749029754204398,
|
| 4199 |
+
"grad_norm": 0.8339402079582214,
|
| 4200 |
+
"learning_rate": 1.2139557674675772e-05,
|
| 4201 |
+
"loss": 1.8496,
|
| 4202 |
+
"step": 599
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 0.7761966364812419,
|
| 4206 |
+
"grad_norm": 1.088751196861267,
|
| 4207 |
+
"learning_rate": 1.200628086311269e-05,
|
| 4208 |
+
"loss": 2.094,
|
| 4209 |
+
"step": 600
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 0.777490297542044,
|
| 4213 |
+
"grad_norm": 0.4214520752429962,
|
| 4214 |
+
"learning_rate": 1.1873639804680598e-05,
|
| 4215 |
+
"loss": 1.4714,
|
| 4216 |
+
"step": 601
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.7787839586028461,
|
| 4220 |
+
"grad_norm": 0.450844943523407,
|
| 4221 |
+
"learning_rate": 1.1741636718877053e-05,
|
| 4222 |
+
"loss": 1.518,
|
| 4223 |
+
"step": 602
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 0.7800776196636481,
|
| 4227 |
+
"grad_norm": 0.458598256111145,
|
| 4228 |
+
"learning_rate": 1.1610273814524398e-05,
|
| 4229 |
+
"loss": 1.5158,
|
| 4230 |
+
"step": 603
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 0.7813712807244502,
|
| 4234 |
+
"grad_norm": 0.47176575660705566,
|
| 4235 |
+
"learning_rate": 1.1479553289732742e-05,
|
| 4236 |
+
"loss": 1.5977,
|
| 4237 |
+
"step": 604
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 0.7826649417852523,
|
| 4241 |
+
"grad_norm": 0.4848724603652954,
|
| 4242 |
+
"learning_rate": 1.134947733186315e-05,
|
| 4243 |
+
"loss": 1.5488,
|
| 4244 |
+
"step": 605
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 0.7839586028460543,
|
| 4248 |
+
"grad_norm": 0.4914746880531311,
|
| 4249 |
+
"learning_rate": 1.1220048117491123e-05,
|
| 4250 |
+
"loss": 1.579,
|
| 4251 |
+
"step": 606
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 0.7852522639068564,
|
| 4255 |
+
"grad_norm": 0.5144307613372803,
|
| 4256 |
+
"learning_rate": 1.1091267812370116e-05,
|
| 4257 |
+
"loss": 1.5866,
|
| 4258 |
+
"step": 607
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.7865459249676585,
|
| 4262 |
+
"grad_norm": 0.4992072284221649,
|
| 4263 |
+
"learning_rate": 1.0963138571395276e-05,
|
| 4264 |
+
"loss": 1.5532,
|
| 4265 |
+
"step": 608
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 0.7878395860284605,
|
| 4269 |
+
"grad_norm": 0.5066733360290527,
|
| 4270 |
+
"learning_rate": 1.0835662538567481e-05,
|
| 4271 |
+
"loss": 1.6586,
|
| 4272 |
+
"step": 609
|
| 4273 |
+
},
|
| 4274 |
+
{
|
| 4275 |
+
"epoch": 0.7891332470892626,
|
| 4276 |
+
"grad_norm": 0.5066677927970886,
|
| 4277 |
+
"learning_rate": 1.0708841846957374e-05,
|
| 4278 |
+
"loss": 1.7165,
|
| 4279 |
+
"step": 610
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 0.7904269081500647,
|
| 4283 |
+
"grad_norm": 0.5076642632484436,
|
| 4284 |
+
"learning_rate": 1.058267861866969e-05,
|
| 4285 |
+
"loss": 1.5526,
|
| 4286 |
+
"step": 611
|
| 4287 |
+
},
|
| 4288 |
+
{
|
| 4289 |
+
"epoch": 0.7917205692108668,
|
| 4290 |
+
"grad_norm": 0.5482640862464905,
|
| 4291 |
+
"learning_rate": 1.0457174964807775e-05,
|
| 4292 |
+
"loss": 1.6631,
|
| 4293 |
+
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 0.7930142302716688,
|
| 4297 |
+
"grad_norm": 0.534303605556488,
|
| 4298 |
+
"learning_rate": 1.0332332985438248e-05,
|
| 4299 |
+
"loss": 1.6207,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 0.7943078913324709,
|
| 4304 |
+
"grad_norm": 0.5623872876167297,
|
| 4305 |
+
"learning_rate": 1.0208154769555827e-05,
|
| 4306 |
+
"loss": 1.6125,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 0.795601552393273,
|
| 4311 |
+
"grad_norm": 0.526637852191925,
|
| 4312 |
+
"learning_rate": 1.0084642395048426e-05,
|
| 4313 |
+
"loss": 1.6224,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 0.796895213454075,
|
| 4318 |
+
"grad_norm": 0.5447207093238831,
|
| 4319 |
+
"learning_rate": 9.96179792866237e-06,
|
| 4320 |
+
"loss": 1.7249,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 0.7981888745148771,
|
| 4325 |
+
"grad_norm": 0.5318351984024048,
|
| 4326 |
+
"learning_rate": 9.83962342596776e-06,
|
| 4327 |
+
"loss": 1.6172,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 0.7994825355756792,
|
| 4332 |
+
"grad_norm": 0.5474776029586792,
|
| 4333 |
+
"learning_rate": 9.71812093132416e-06,
|
| 4334 |
+
"loss": 1.619,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 0.8007761966364813,
|
| 4339 |
+
"grad_norm": 0.5580310821533203,
|
| 4340 |
+
"learning_rate": 9.597292477846353e-06,
|
| 4341 |
+
"loss": 1.7075,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 0.8020698576972833,
|
| 4346 |
+
"grad_norm": 0.5502408146858215,
|
| 4347 |
+
"learning_rate": 9.477140087370268e-06,
|
| 4348 |
+
"loss": 1.6628,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 0.8033635187580854,
|
| 4353 |
+
"grad_norm": 0.5557919144630432,
|
| 4354 |
+
"learning_rate": 9.357665770419244e-06,
|
| 4355 |
+
"loss": 1.8303,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 0.8046571798188874,
|
| 4360 |
+
"grad_norm": 0.5612048506736755,
|
| 4361 |
+
"learning_rate": 9.238871526170334e-06,
|
| 4362 |
+
"loss": 1.672,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 0.8059508408796895,
|
| 4367 |
+
"grad_norm": 0.5757585167884827,
|
| 4368 |
+
"learning_rate": 9.12075934242082e-06,
|
| 4369 |
+
"loss": 1.6857,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 0.8072445019404916,
|
| 4374 |
+
"grad_norm": 0.5578299164772034,
|
| 4375 |
+
"learning_rate": 9.003331195555043e-06,
|
| 4376 |
+
"loss": 1.8077,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 0.8085381630012937,
|
| 4381 |
+
"grad_norm": 0.5981864929199219,
|
| 4382 |
+
"learning_rate": 8.886589050511257e-06,
|
| 4383 |
+
"loss": 1.7781,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 0.8098318240620958,
|
| 4388 |
+
"grad_norm": 0.5973100066184998,
|
| 4389 |
+
"learning_rate": 8.770534860748747e-06,
|
| 4390 |
+
"loss": 1.7344,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 0.8111254851228978,
|
| 4395 |
+
"grad_norm": 0.5700523257255554,
|
| 4396 |
+
"learning_rate": 8.655170568215193e-06,
|
| 4397 |
+
"loss": 1.7169,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 0.8124191461836999,
|
| 4402 |
+
"grad_norm": 0.6092631220817566,
|
| 4403 |
+
"learning_rate": 8.540498103314154e-06,
|
| 4404 |
+
"loss": 1.8428,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 0.8137128072445019,
|
| 4409 |
+
"grad_norm": 0.6047289967536926,
|
| 4410 |
+
"learning_rate": 8.426519384872733e-06,
|
| 4411 |
+
"loss": 1.826,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 0.815006468305304,
|
| 4416 |
+
"grad_norm": 0.6209927201271057,
|
| 4417 |
+
"learning_rate": 8.313236320109542e-06,
|
| 4418 |
+
"loss": 1.7561,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 0.8163001293661061,
|
| 4423 |
+
"grad_norm": 0.5962238311767578,
|
| 4424 |
+
"learning_rate": 8.200650804602667e-06,
|
| 4425 |
+
"loss": 1.8062,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 0.8175937904269082,
|
| 4430 |
+
"grad_norm": 0.6150239109992981,
|
| 4431 |
+
"learning_rate": 8.088764722258097e-06,
|
| 4432 |
+
"loss": 1.7115,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 0.8188874514877102,
|
| 4437 |
+
"grad_norm": 0.6377602219581604,
|
| 4438 |
+
"learning_rate": 7.97757994527809e-06,
|
| 4439 |
+
"loss": 1.7661,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 0.8201811125485123,
|
| 4444 |
+
"grad_norm": 0.6375808715820312,
|
| 4445 |
+
"learning_rate": 7.86709833412992e-06,
|
| 4446 |
+
"loss": 1.7545,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 0.8214747736093143,
|
| 4451 |
+
"grad_norm": 0.641861081123352,
|
| 4452 |
+
"learning_rate": 7.757321737514645e-06,
|
| 4453 |
+
"loss": 1.8139,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 0.8227684346701164,
|
| 4458 |
+
"grad_norm": 0.6656633019447327,
|
| 4459 |
+
"learning_rate": 7.648251992336291e-06,
|
| 4460 |
+
"loss": 1.7277,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 0.8240620957309185,
|
| 4465 |
+
"grad_norm": 0.6750869154930115,
|
| 4466 |
+
"learning_rate": 7.539890923671062e-06,
|
| 4467 |
+
"loss": 1.946,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 0.8253557567917206,
|
| 4472 |
+
"grad_norm": 0.6535232067108154,
|
| 4473 |
+
"learning_rate": 7.43224034473674e-06,
|
| 4474 |
+
"loss": 1.7854,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 0.8266494178525227,
|
| 4479 |
+
"grad_norm": 0.6922054886817932,
|
| 4480 |
+
"learning_rate": 7.325302056862477e-06,
|
| 4481 |
+
"loss": 1.7684,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 0.8279430789133247,
|
| 4486 |
+
"grad_norm": 0.6872656345367432,
|
| 4487 |
+
"learning_rate": 7.219077849458538e-06,
|
| 4488 |
+
"loss": 1.8425,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 0.8292367399741267,
|
| 4493 |
+
"grad_norm": 0.7008704543113708,
|
| 4494 |
+
"learning_rate": 7.1135694999864e-06,
|
| 4495 |
+
"loss": 1.8705,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 0.8305304010349288,
|
| 4500 |
+
"grad_norm": 0.729759156703949,
|
| 4501 |
+
"learning_rate": 7.00877877392902e-06,
|
| 4502 |
+
"loss": 1.6901,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.8318240620957309,
|
| 4507 |
+
"grad_norm": 0.7461994886398315,
|
| 4508 |
+
"learning_rate": 6.904707424761292e-06,
|
| 4509 |
+
"loss": 1.9108,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 0.833117723156533,
|
| 4514 |
+
"grad_norm": 0.769800066947937,
|
| 4515 |
+
"learning_rate": 6.801357193920665e-06,
|
| 4516 |
+
"loss": 2.1872,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 0.8344113842173351,
|
| 4521 |
+
"grad_norm": 0.7531771659851074,
|
| 4522 |
+
"learning_rate": 6.698729810778065e-06,
|
| 4523 |
+
"loss": 1.8978,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 0.8357050452781372,
|
| 4528 |
+
"grad_norm": 0.7885345220565796,
|
| 4529 |
+
"learning_rate": 6.596826992608929e-06,
|
| 4530 |
+
"loss": 1.9816,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 0.8369987063389392,
|
| 4535 |
+
"grad_norm": 0.7793993353843689,
|
| 4536 |
+
"learning_rate": 6.495650444564433e-06,
|
| 4537 |
+
"loss": 1.8709,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 0.8382923673997412,
|
| 4542 |
+
"grad_norm": 0.8853140473365784,
|
| 4543 |
+
"learning_rate": 6.3952018596430245e-06,
|
| 4544 |
+
"loss": 2.0212,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 0.8395860284605433,
|
| 4549 |
+
"grad_norm": 0.8999887108802795,
|
| 4550 |
+
"learning_rate": 6.295482918662066e-06,
|
| 4551 |
+
"loss": 2.2873,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 0.8408796895213454,
|
| 4556 |
+
"grad_norm": 0.9850187301635742,
|
| 4557 |
+
"learning_rate": 6.196495290229676e-06,
|
| 4558 |
+
"loss": 2.1436,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 0.8421733505821475,
|
| 4563 |
+
"grad_norm": 0.4259641468524933,
|
| 4564 |
+
"learning_rate": 6.098240630716867e-06,
|
| 4565 |
+
"loss": 1.2801,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 0.8434670116429496,
|
| 4570 |
+
"grad_norm": 0.46614962816238403,
|
| 4571 |
+
"learning_rate": 6.000720584229802e-06,
|
| 4572 |
+
"loss": 1.567,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 0.8447606727037517,
|
| 4577 |
+
"grad_norm": 0.4704464077949524,
|
| 4578 |
+
"learning_rate": 5.903936782582253e-06,
|
| 4579 |
+
"loss": 1.8213,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 0.8460543337645536,
|
| 4584 |
+
"grad_norm": 0.4731570780277252,
|
| 4585 |
+
"learning_rate": 5.80789084526836e-06,
|
| 4586 |
+
"loss": 1.6202,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 0.8473479948253557,
|
| 4591 |
+
"grad_norm": 0.48283982276916504,
|
| 4592 |
+
"learning_rate": 5.712584379435482e-06,
|
| 4593 |
+
"loss": 1.5875,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 0.8486416558861578,
|
| 4598 |
+
"grad_norm": 0.5061183571815491,
|
| 4599 |
+
"learning_rate": 5.618018979857309e-06,
|
| 4600 |
+
"loss": 1.5742,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 0.8499353169469599,
|
| 4605 |
+
"grad_norm": 0.48858606815338135,
|
| 4606 |
+
"learning_rate": 5.524196228907203e-06,
|
| 4607 |
+
"loss": 1.5377,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 0.851228978007762,
|
| 4612 |
+
"grad_norm": 0.48720577359199524,
|
| 4613 |
+
"learning_rate": 5.4311176965317065e-06,
|
| 4614 |
+
"loss": 1.5387,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 0.8525226390685641,
|
| 4619 |
+
"grad_norm": 0.49719151854515076,
|
| 4620 |
+
"learning_rate": 5.33878494022424e-06,
|
| 4621 |
+
"loss": 1.6273,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 0.8538163001293662,
|
| 4626 |
+
"grad_norm": 0.4926326274871826,
|
| 4627 |
+
"learning_rate": 5.247199504999106e-06,
|
| 4628 |
+
"loss": 1.7055,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 0.8551099611901681,
|
| 4633 |
+
"grad_norm": 0.5013806819915771,
|
| 4634 |
+
"learning_rate": 5.156362923365588e-06,
|
| 4635 |
+
"loss": 1.6868,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 0.8564036222509702,
|
| 4640 |
+
"grad_norm": 0.5549349784851074,
|
| 4641 |
+
"learning_rate": 5.066276715302304e-06,
|
| 4642 |
+
"loss": 1.571,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 0.8576972833117723,
|
| 4647 |
+
"grad_norm": 0.5115381479263306,
|
| 4648 |
+
"learning_rate": 4.976942388231826e-06,
|
| 4649 |
+
"loss": 1.6374,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 0.8589909443725744,
|
| 4654 |
+
"grad_norm": 0.5344304442405701,
|
| 4655 |
+
"learning_rate": 4.888361436995359e-06,
|
| 4656 |
+
"loss": 1.6862,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 0.8602846054333765,
|
| 4661 |
+
"grad_norm": 0.5527788400650024,
|
| 4662 |
+
"learning_rate": 4.800535343827833e-06,
|
| 4663 |
+
"loss": 1.7234,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 0.8615782664941786,
|
| 4668 |
+
"grad_norm": 0.5573310852050781,
|
| 4669 |
+
"learning_rate": 4.7134655783330425e-06,
|
| 4670 |
+
"loss": 1.6087,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 0.8628719275549805,
|
| 4675 |
+
"grad_norm": 0.5454318523406982,
|
| 4676 |
+
"learning_rate": 4.627153597459072e-06,
|
| 4677 |
+
"loss": 1.6846,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 0.8641655886157826,
|
| 4682 |
+
"grad_norm": 0.5258266925811768,
|
| 4683 |
+
"learning_rate": 4.541600845473881e-06,
|
| 4684 |
+
"loss": 1.7079,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 0.8654592496765847,
|
| 4689 |
+
"grad_norm": 0.5511134266853333,
|
| 4690 |
+
"learning_rate": 4.456808753941205e-06,
|
| 4691 |
+
"loss": 1.6727,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 0.8667529107373868,
|
| 4696 |
+
"grad_norm": 0.5851698517799377,
|
| 4697 |
+
"learning_rate": 4.372778741696559e-06,
|
| 4698 |
+
"loss": 1.8386,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 0.8680465717981889,
|
| 4703 |
+
"grad_norm": 0.5556474328041077,
|
| 4704 |
+
"learning_rate": 4.289512214823466e-06,
|
| 4705 |
+
"loss": 1.6436,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 0.869340232858991,
|
| 4710 |
+
"grad_norm": 0.5633500814437866,
|
| 4711 |
+
"learning_rate": 4.207010566630004e-06,
|
| 4712 |
+
"loss": 1.8678,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 0.8706338939197931,
|
| 4717 |
+
"grad_norm": 0.5625820755958557,
|
| 4718 |
+
"learning_rate": 4.1252751776254375e-06,
|
| 4719 |
+
"loss": 1.6911,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 0.871927554980595,
|
| 4724 |
+
"grad_norm": 0.607752799987793,
|
| 4725 |
+
"learning_rate": 4.044307415497112e-06,
|
| 4726 |
+
"loss": 1.8531,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 0.8732212160413971,
|
| 4731 |
+
"grad_norm": 0.6022449731826782,
|
| 4732 |
+
"learning_rate": 3.964108635087615e-06,
|
| 4733 |
+
"loss": 1.9056,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 0.8745148771021992,
|
| 4738 |
+
"grad_norm": 0.5852234363555908,
|
| 4739 |
+
"learning_rate": 3.884680178372069e-06,
|
| 4740 |
+
"loss": 1.7683,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 0.8758085381630013,
|
| 4745 |
+
"grad_norm": 0.5821849703788757,
|
| 4746 |
+
"learning_rate": 3.8060233744356633e-06,
|
| 4747 |
+
"loss": 1.8066,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 0.8771021992238034,
|
| 4752 |
+
"grad_norm": 0.5757661461830139,
|
| 4753 |
+
"learning_rate": 3.728139539451464e-06,
|
| 4754 |
+
"loss": 1.7934,
|
| 4755 |
+
"step": 678
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 0.8783958602846055,
|
| 4759 |
+
"grad_norm": 0.6299412250518799,
|
| 4760 |
+
"learning_rate": 3.65102997665836e-06,
|
| 4761 |
+
"loss": 1.7985,
|
| 4762 |
+
"step": 679
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 0.8796895213454075,
|
| 4766 |
+
"grad_norm": 0.6304367184638977,
|
| 4767 |
+
"learning_rate": 3.574695976339226e-06,
|
| 4768 |
+
"loss": 1.7298,
|
| 4769 |
+
"step": 680
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 0.8809831824062095,
|
| 4773 |
+
"grad_norm": 0.5838976502418518,
|
| 4774 |
+
"learning_rate": 3.4991388157993966e-06,
|
| 4775 |
+
"loss": 1.7117,
|
| 4776 |
+
"step": 681
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 0.8822768434670116,
|
| 4780 |
+
"grad_norm": 0.6452072262763977,
|
| 4781 |
+
"learning_rate": 3.4243597593452523e-06,
|
| 4782 |
+
"loss": 1.7692,
|
| 4783 |
+
"step": 682
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 0.8835705045278137,
|
| 4787 |
+
"grad_norm": 0.6496413350105286,
|
| 4788 |
+
"learning_rate": 3.350360058263058e-06,
|
| 4789 |
+
"loss": 1.7692,
|
| 4790 |
+
"step": 683
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 0.8848641655886158,
|
| 4794 |
+
"grad_norm": 0.6279682517051697,
|
| 4795 |
+
"learning_rate": 3.277140950798052e-06,
|
| 4796 |
+
"loss": 2.0296,
|
| 4797 |
+
"step": 684
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 0.8861578266494179,
|
| 4801 |
+
"grad_norm": 0.6427062749862671,
|
| 4802 |
+
"learning_rate": 3.2047036621337236e-06,
|
| 4803 |
+
"loss": 1.6774,
|
| 4804 |
+
"step": 685
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 0.88745148771022,
|
| 4808 |
+
"grad_norm": 0.6491368412971497,
|
| 4809 |
+
"learning_rate": 3.133049404371258e-06,
|
| 4810 |
+
"loss": 1.8565,
|
| 4811 |
+
"step": 686
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 0.888745148771022,
|
| 4815 |
+
"grad_norm": 0.6646947860717773,
|
| 4816 |
+
"learning_rate": 3.0621793765093444e-06,
|
| 4817 |
+
"loss": 1.7678,
|
| 4818 |
+
"step": 687
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 0.890038809831824,
|
| 4822 |
+
"grad_norm": 0.7040244936943054,
|
| 4823 |
+
"learning_rate": 2.9920947644240473e-06,
|
| 4824 |
+
"loss": 1.8575,
|
| 4825 |
+
"step": 688
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 0.8913324708926261,
|
| 4829 |
+
"grad_norm": 0.6854454874992371,
|
| 4830 |
+
"learning_rate": 2.9227967408489653e-06,
|
| 4831 |
+
"loss": 1.7821,
|
| 4832 |
+
"step": 689
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 0.8926261319534282,
|
| 4836 |
+
"grad_norm": 0.6900594234466553,
|
| 4837 |
+
"learning_rate": 2.8542864653556546e-06,
|
| 4838 |
+
"loss": 1.8255,
|
| 4839 |
+
"step": 690
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 0.8939197930142303,
|
| 4843 |
+
"grad_norm": 0.6792859435081482,
|
| 4844 |
+
"learning_rate": 2.786565084334175e-06,
|
| 4845 |
+
"loss": 1.9492,
|
| 4846 |
+
"step": 691
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 0.8952134540750324,
|
| 4850 |
+
"grad_norm": 0.6839621663093567,
|
| 4851 |
+
"learning_rate": 2.7196337309739417e-06,
|
| 4852 |
+
"loss": 1.8545,
|
| 4853 |
+
"step": 692
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 0.8965071151358344,
|
| 4857 |
+
"grad_norm": 0.7339411377906799,
|
| 4858 |
+
"learning_rate": 2.653493525244721e-06,
|
| 4859 |
+
"loss": 1.9469,
|
| 4860 |
+
"step": 693
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 0.8978007761966365,
|
| 4864 |
+
"grad_norm": 0.7272126078605652,
|
| 4865 |
+
"learning_rate": 2.5881455738779483e-06,
|
| 4866 |
+
"loss": 1.9897,
|
| 4867 |
+
"step": 694
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 0.8990944372574385,
|
| 4871 |
+
"grad_norm": 0.7714767456054688,
|
| 4872 |
+
"learning_rate": 2.5235909703481665e-06,
|
| 4873 |
+
"loss": 1.9615,
|
| 4874 |
+
"step": 695
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 0.9003880983182406,
|
| 4878 |
+
"grad_norm": 0.7937828302383423,
|
| 4879 |
+
"learning_rate": 2.4598307948547395e-06,
|
| 4880 |
+
"loss": 2.0227,
|
| 4881 |
+
"step": 696
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 0.9016817593790427,
|
| 4885 |
+
"grad_norm": 0.837083637714386,
|
| 4886 |
+
"learning_rate": 2.3968661143037862e-06,
|
| 4887 |
+
"loss": 1.8228,
|
| 4888 |
+
"step": 697
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 0.9029754204398448,
|
| 4892 |
+
"grad_norm": 0.7950493097305298,
|
| 4893 |
+
"learning_rate": 2.334697982290307e-06,
|
| 4894 |
+
"loss": 1.8966,
|
| 4895 |
+
"step": 698
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 0.9042690815006468,
|
| 4899 |
+
"grad_norm": 0.9044060707092285,
|
| 4900 |
+
"learning_rate": 2.273327439080575e-06,
|
| 4901 |
+
"loss": 1.9586,
|
| 4902 |
+
"step": 699
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 0.9055627425614489,
|
| 4906 |
+
"grad_norm": 1.0963821411132812,
|
| 4907 |
+
"learning_rate": 2.2127555115947276e-06,
|
| 4908 |
+
"loss": 2.0201,
|
| 4909 |
+
"step": 700
|
| 4910 |
+
},
|
| 4911 |
+
{
|
| 4912 |
+
"epoch": 0.906856403622251,
|
| 4913 |
+
"grad_norm": 0.42113515734672546,
|
| 4914 |
+
"learning_rate": 2.152983213389559e-06,
|
| 4915 |
+
"loss": 1.4756,
|
| 4916 |
+
"step": 701
|
| 4917 |
+
},
|
| 4918 |
+
{
|
| 4919 |
+
"epoch": 0.908150064683053,
|
| 4920 |
+
"grad_norm": 0.47029221057891846,
|
| 4921 |
+
"learning_rate": 2.0940115446415886e-06,
|
| 4922 |
+
"loss": 1.5055,
|
| 4923 |
+
"step": 702
|
| 4924 |
+
},
|
| 4925 |
+
{
|
| 4926 |
+
"epoch": 0.9094437257438551,
|
| 4927 |
+
"grad_norm": 0.4599394202232361,
|
| 4928 |
+
"learning_rate": 2.035841492130319e-06,
|
| 4929 |
+
"loss": 1.7587,
|
| 4930 |
+
"step": 703
|
| 4931 |
+
},
|
| 4932 |
+
{
|
| 4933 |
+
"epoch": 0.9107373868046572,
|
| 4934 |
+
"grad_norm": 0.4609547257423401,
|
| 4935 |
+
"learning_rate": 1.9784740292217108e-06,
|
| 4936 |
+
"loss": 1.5761,
|
| 4937 |
+
"step": 704
|
| 4938 |
+
},
|
| 4939 |
+
{
|
| 4940 |
+
"epoch": 0.9120310478654593,
|
| 4941 |
+
"grad_norm": 0.4895348846912384,
|
| 4942 |
+
"learning_rate": 1.9219101158518993e-06,
|
| 4943 |
+
"loss": 1.7727,
|
| 4944 |
+
"step": 705
|
| 4945 |
+
},
|
| 4946 |
+
{
|
| 4947 |
+
"epoch": 0.9133247089262613,
|
| 4948 |
+
"grad_norm": 0.48977532982826233,
|
| 4949 |
+
"learning_rate": 1.86615069851116e-06,
|
| 4950 |
+
"loss": 1.6288,
|
| 4951 |
+
"step": 706
|
| 4952 |
+
},
|
| 4953 |
+
{
|
| 4954 |
+
"epoch": 0.9146183699870634,
|
| 4955 |
+
"grad_norm": 0.4814504384994507,
|
| 4956 |
+
"learning_rate": 1.811196710228008e-06,
|
| 4957 |
+
"loss": 1.6022,
|
| 4958 |
+
"step": 707
|
| 4959 |
+
},
|
| 4960 |
+
{
|
| 4961 |
+
"epoch": 0.9159120310478654,
|
| 4962 |
+
"grad_norm": 0.5106490254402161,
|
| 4963 |
+
"learning_rate": 1.757049070553657e-06,
|
| 4964 |
+
"loss": 1.6077,
|
| 4965 |
+
"step": 708
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.9172056921086675,
|
| 4969 |
+
"grad_norm": 0.5112571716308594,
|
| 4970 |
+
"learning_rate": 1.70370868554659e-06,
|
| 4971 |
+
"loss": 1.4259,
|
| 4972 |
+
"step": 709
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.9184993531694696,
|
| 4976 |
+
"grad_norm": 0.49610233306884766,
|
| 4977 |
+
"learning_rate": 1.6511764477573965e-06,
|
| 4978 |
+
"loss": 1.5944,
|
| 4979 |
+
"step": 710
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.9197930142302717,
|
| 4983 |
+
"grad_norm": 0.514818549156189,
|
| 4984 |
+
"learning_rate": 1.599453236213866e-06,
|
| 4985 |
+
"loss": 1.4695,
|
| 4986 |
+
"step": 711
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.9210866752910737,
|
| 4990 |
+
"grad_norm": 0.5288085341453552,
|
| 4991 |
+
"learning_rate": 1.5485399164062552e-06,
|
| 4992 |
+
"loss": 1.6117,
|
| 4993 |
+
"step": 712
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.9223803363518758,
|
| 4997 |
+
"grad_norm": 0.5362600088119507,
|
| 4998 |
+
"learning_rate": 1.4984373402728014e-06,
|
| 4999 |
+
"loss": 1.5812,
|
| 5000 |
+
"step": 713
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.9236739974126779,
|
| 5004 |
+
"grad_norm": 0.5225064754486084,
|
| 5005 |
+
"learning_rate": 1.4491463461854981e-06,
|
| 5006 |
+
"loss": 1.6608,
|
| 5007 |
+
"step": 714
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.92496765847348,
|
| 5011 |
+
"grad_norm": 0.5195352435112,
|
| 5012 |
+
"learning_rate": 1.4006677589360306e-06,
|
| 5013 |
+
"loss": 1.7514,
|
| 5014 |
+
"step": 715
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.926261319534282,
|
| 5018 |
+
"grad_norm": 0.5668680667877197,
|
| 5019 |
+
"learning_rate": 1.3530023897219968e-06,
|
| 5020 |
+
"loss": 1.7045,
|
| 5021 |
+
"step": 716
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.9275549805950841,
|
| 5025 |
+
"grad_norm": 0.5474237203598022,
|
| 5026 |
+
"learning_rate": 1.3061510361333185e-06,
|
| 5027 |
+
"loss": 1.7969,
|
| 5028 |
+
"step": 717
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 0.9288486416558862,
|
| 5032 |
+
"grad_norm": 0.556367814540863,
|
| 5033 |
+
"learning_rate": 1.2601144821389187e-06,
|
| 5034 |
+
"loss": 1.6464,
|
| 5035 |
+
"step": 718
|
| 5036 |
+
},
|
| 5037 |
+
{
|
| 5038 |
+
"epoch": 0.9301423027166882,
|
| 5039 |
+
"grad_norm": 0.5793610215187073,
|
| 5040 |
+
"learning_rate": 1.214893498073577e-06,
|
| 5041 |
+
"loss": 1.7267,
|
| 5042 |
+
"step": 719
|
| 5043 |
+
},
|
| 5044 |
+
{
|
| 5045 |
+
"epoch": 0.9314359637774903,
|
| 5046 |
+
"grad_norm": 0.5605274438858032,
|
| 5047 |
+
"learning_rate": 1.1704888406250447e-06,
|
| 5048 |
+
"loss": 1.6689,
|
| 5049 |
+
"step": 720
|
| 5050 |
+
},
|
| 5051 |
+
{
|
| 5052 |
+
"epoch": 0.9327296248382924,
|
| 5053 |
+
"grad_norm": 0.559226930141449,
|
| 5054 |
+
"learning_rate": 1.1269012528214107e-06,
|
| 5055 |
+
"loss": 1.7192,
|
| 5056 |
+
"step": 721
|
| 5057 |
+
},
|
| 5058 |
+
{
|
| 5059 |
+
"epoch": 0.9340232858990944,
|
| 5060 |
+
"grad_norm": 0.5833441019058228,
|
| 5061 |
+
"learning_rate": 1.0841314640186228e-06,
|
| 5062 |
+
"loss": 1.6467,
|
| 5063 |
+
"step": 722
|
| 5064 |
+
},
|
| 5065 |
+
{
|
| 5066 |
+
"epoch": 0.9353169469598965,
|
| 5067 |
+
"grad_norm": 0.614727795124054,
|
| 5068 |
+
"learning_rate": 1.0421801898883143e-06,
|
| 5069 |
+
"loss": 1.7859,
|
| 5070 |
+
"step": 723
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 0.9366106080206986,
|
| 5074 |
+
"grad_norm": 0.5924329161643982,
|
| 5075 |
+
"learning_rate": 1.0010481324058352e-06,
|
| 5076 |
+
"loss": 1.6152,
|
| 5077 |
+
"step": 724
|
| 5078 |
+
},
|
| 5079 |
+
{
|
| 5080 |
+
"epoch": 0.9379042690815006,
|
| 5081 |
+
"grad_norm": 0.5784980654716492,
|
| 5082 |
+
"learning_rate": 9.607359798384785e-07,
|
| 5083 |
+
"loss": 1.7243,
|
| 5084 |
+
"step": 725
|
| 5085 |
+
},
|
| 5086 |
+
{
|
| 5087 |
+
"epoch": 0.9391979301423027,
|
| 5088 |
+
"grad_norm": 0.5997685790061951,
|
| 5089 |
+
"learning_rate": 9.212444067339787e-07,
|
| 5090 |
+
"loss": 1.7772,
|
| 5091 |
+
"step": 726
|
| 5092 |
+
},
|
| 5093 |
+
{
|
| 5094 |
+
"epoch": 0.9404915912031048,
|
| 5095 |
+
"grad_norm": 0.6061770915985107,
|
| 5096 |
+
"learning_rate": 8.825740739092148e-07,
|
| 5097 |
+
"loss": 1.9133,
|
| 5098 |
+
"step": 727
|
| 5099 |
+
},
|
| 5100 |
+
{
|
| 5101 |
+
"epoch": 0.9417852522639069,
|
| 5102 |
+
"grad_norm": 0.5751582980155945,
|
| 5103 |
+
"learning_rate": 8.447256284391858e-07,
|
| 5104 |
+
"loss": 1.6525,
|
| 5105 |
+
"step": 728
|
| 5106 |
+
},
|
| 5107 |
+
{
|
| 5108 |
+
"epoch": 0.943078913324709,
|
| 5109 |
+
"grad_norm": 0.6227949857711792,
|
| 5110 |
+
"learning_rate": 8.076997036461253e-07,
|
| 5111 |
+
"loss": 1.805,
|
| 5112 |
+
"step": 729
|
| 5113 |
+
},
|
| 5114 |
+
{
|
| 5115 |
+
"epoch": 0.944372574385511,
|
| 5116 |
+
"grad_norm": 0.6226868033409119,
|
| 5117 |
+
"learning_rate": 7.714969190889765e-07,
|
| 5118 |
+
"loss": 1.9385,
|
| 5119 |
+
"step": 730
|
| 5120 |
+
},
|
| 5121 |
+
{
|
| 5122 |
+
"epoch": 0.9456662354463131,
|
| 5123 |
+
"grad_norm": 0.5905411243438721,
|
| 5124 |
+
"learning_rate": 7.36117880552939e-07,
|
| 5125 |
+
"loss": 1.7853,
|
| 5126 |
+
"step": 731
|
| 5127 |
+
},
|
| 5128 |
+
{
|
| 5129 |
+
"epoch": 0.9469598965071151,
|
| 5130 |
+
"grad_norm": 0.6376669406890869,
|
| 5131 |
+
"learning_rate": 7.015631800394107e-07,
|
| 5132 |
+
"loss": 2.0066,
|
| 5133 |
+
"step": 732
|
| 5134 |
+
},
|
| 5135 |
+
{
|
| 5136 |
+
"epoch": 0.9482535575679172,
|
| 5137 |
+
"grad_norm": 0.6215546727180481,
|
| 5138 |
+
"learning_rate": 6.678333957560512e-07,
|
| 5139 |
+
"loss": 1.7759,
|
| 5140 |
+
"step": 733
|
| 5141 |
+
},
|
| 5142 |
+
{
|
| 5143 |
+
"epoch": 0.9495472186287193,
|
| 5144 |
+
"grad_norm": 0.6270310282707214,
|
| 5145 |
+
"learning_rate": 6.349290921070783e-07,
|
| 5146 |
+
"loss": 1.8527,
|
| 5147 |
+
"step": 734
|
| 5148 |
+
},
|
| 5149 |
+
{
|
| 5150 |
+
"epoch": 0.9508408796895214,
|
| 5151 |
+
"grad_norm": 0.6710889935493469,
|
| 5152 |
+
"learning_rate": 6.028508196838811e-07,
|
| 5153 |
+
"loss": 1.863,
|
| 5154 |
+
"step": 735
|
| 5155 |
+
},
|
| 5156 |
+
{
|
| 5157 |
+
"epoch": 0.9521345407503234,
|
| 5158 |
+
"grad_norm": 0.6601763367652893,
|
| 5159 |
+
"learning_rate": 5.715991152557554e-07,
|
| 5160 |
+
"loss": 1.9286,
|
| 5161 |
+
"step": 736
|
| 5162 |
+
},
|
| 5163 |
+
{
|
| 5164 |
+
"epoch": 0.9534282018111255,
|
| 5165 |
+
"grad_norm": 0.6637079119682312,
|
| 5166 |
+
"learning_rate": 5.411745017609493e-07,
|
| 5167 |
+
"loss": 1.9146,
|
| 5168 |
+
"step": 737
|
| 5169 |
+
},
|
| 5170 |
+
{
|
| 5171 |
+
"epoch": 0.9547218628719275,
|
| 5172 |
+
"grad_norm": 0.6605135798454285,
|
| 5173 |
+
"learning_rate": 5.115774882979096e-07,
|
| 5174 |
+
"loss": 1.7605,
|
| 5175 |
+
"step": 738
|
| 5176 |
+
},
|
| 5177 |
+
{
|
| 5178 |
+
"epoch": 0.9560155239327296,
|
| 5179 |
+
"grad_norm": 0.6722599864006042,
|
| 5180 |
+
"learning_rate": 4.828085701167607e-07,
|
| 5181 |
+
"loss": 1.8122,
|
| 5182 |
+
"step": 739
|
| 5183 |
+
},
|
| 5184 |
+
{
|
| 5185 |
+
"epoch": 0.9573091849935317,
|
| 5186 |
+
"grad_norm": 0.6760243773460388,
|
| 5187 |
+
"learning_rate": 4.548682286109995e-07,
|
| 5188 |
+
"loss": 1.6851,
|
| 5189 |
+
"step": 740
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 0.9586028460543338,
|
| 5193 |
+
"grad_norm": 0.6736776232719421,
|
| 5194 |
+
"learning_rate": 4.277569313094809e-07,
|
| 5195 |
+
"loss": 1.786,
|
| 5196 |
+
"step": 741
|
| 5197 |
+
},
|
| 5198 |
+
{
|
| 5199 |
+
"epoch": 0.9598965071151359,
|
| 5200 |
+
"grad_norm": 0.7444686889648438,
|
| 5201 |
+
"learning_rate": 4.0147513186855615e-07,
|
| 5202 |
+
"loss": 1.9045,
|
| 5203 |
+
"step": 742
|
| 5204 |
+
},
|
| 5205 |
+
{
|
| 5206 |
+
"epoch": 0.9611901681759379,
|
| 5207 |
+
"grad_norm": 0.7153714299201965,
|
| 5208 |
+
"learning_rate": 3.7602327006450167e-07,
|
| 5209 |
+
"loss": 2.0627,
|
| 5210 |
+
"step": 743
|
| 5211 |
+
},
|
| 5212 |
+
{
|
| 5213 |
+
"epoch": 0.96248382923674,
|
| 5214 |
+
"grad_norm": 0.7691993117332458,
|
| 5215 |
+
"learning_rate": 3.514017717861529e-07,
|
| 5216 |
+
"loss": 2.0108,
|
| 5217 |
+
"step": 744
|
| 5218 |
+
},
|
| 5219 |
+
{
|
| 5220 |
+
"epoch": 0.963777490297542,
|
| 5221 |
+
"grad_norm": 0.7638096809387207,
|
| 5222 |
+
"learning_rate": 3.2761104902778173e-07,
|
| 5223 |
+
"loss": 1.9658,
|
| 5224 |
+
"step": 745
|
| 5225 |
+
},
|
| 5226 |
+
{
|
| 5227 |
+
"epoch": 0.9650711513583441,
|
| 5228 |
+
"grad_norm": 0.7966986894607544,
|
| 5229 |
+
"learning_rate": 3.04651499882197e-07,
|
| 5230 |
+
"loss": 2.1942,
|
| 5231 |
+
"step": 746
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 0.9663648124191462,
|
| 5235 |
+
"grad_norm": 0.794858455657959,
|
| 5236 |
+
"learning_rate": 2.825235085340938e-07,
|
| 5237 |
+
"loss": 1.936,
|
| 5238 |
+
"step": 747
|
| 5239 |
+
},
|
| 5240 |
+
{
|
| 5241 |
+
"epoch": 0.9676584734799483,
|
| 5242 |
+
"grad_norm": 0.858466386795044,
|
| 5243 |
+
"learning_rate": 2.612274452536201e-07,
|
| 5244 |
+
"loss": 1.9205,
|
| 5245 |
+
"step": 748
|
| 5246 |
+
},
|
| 5247 |
+
{
|
| 5248 |
+
"epoch": 0.9689521345407504,
|
| 5249 |
+
"grad_norm": 1.1012519598007202,
|
| 5250 |
+
"learning_rate": 2.407636663901591e-07,
|
| 5251 |
+
"loss": 2.1395,
|
| 5252 |
+
"step": 749
|
| 5253 |
+
},
|
| 5254 |
+
{
|
| 5255 |
+
"epoch": 0.9702457956015524,
|
| 5256 |
+
"grad_norm": 1.0824167728424072,
|
| 5257 |
+
"learning_rate": 2.211325143664067e-07,
|
| 5258 |
+
"loss": 2.106,
|
| 5259 |
+
"step": 750
|
| 5260 |
+
},
|
| 5261 |
+
{
|
| 5262 |
+
"epoch": 0.9715394566623544,
|
| 5263 |
+
"grad_norm": 0.4427192211151123,
|
| 5264 |
+
"learning_rate": 2.0233431767261447e-07,
|
| 5265 |
+
"loss": 1.7156,
|
| 5266 |
+
"step": 751
|
| 5267 |
+
},
|
| 5268 |
+
{
|
| 5269 |
+
"epoch": 0.9728331177231565,
|
| 5270 |
+
"grad_norm": 0.4759114384651184,
|
| 5271 |
+
"learning_rate": 1.8436939086109995e-07,
|
| 5272 |
+
"loss": 1.5656,
|
| 5273 |
+
"step": 752
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 0.9741267787839586,
|
| 5277 |
+
"grad_norm": 0.48882219195365906,
|
| 5278 |
+
"learning_rate": 1.6723803454098408e-07,
|
| 5279 |
+
"loss": 1.6891,
|
| 5280 |
+
"step": 753
|
| 5281 |
+
},
|
| 5282 |
+
{
|
| 5283 |
+
"epoch": 0.9754204398447607,
|
| 5284 |
+
"grad_norm": 0.48134151101112366,
|
| 5285 |
+
"learning_rate": 1.5094053537316188e-07,
|
| 5286 |
+
"loss": 1.4881,
|
| 5287 |
+
"step": 754
|
| 5288 |
+
},
|
| 5289 |
+
{
|
| 5290 |
+
"epoch": 0.9767141009055628,
|
| 5291 |
+
"grad_norm": 0.528454601764679,
|
| 5292 |
+
"learning_rate": 1.3547716606548966e-07,
|
| 5293 |
+
"loss": 1.4452,
|
| 5294 |
+
"step": 755
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.9780077619663649,
|
| 5298 |
+
"grad_norm": 0.5208367109298706,
|
| 5299 |
+
"learning_rate": 1.2084818536825526e-07,
|
| 5300 |
+
"loss": 1.4119,
|
| 5301 |
+
"step": 756
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.9793014230271668,
|
| 5305 |
+
"grad_norm": 0.5646911859512329,
|
| 5306 |
+
"learning_rate": 1.0705383806982606e-07,
|
| 5307 |
+
"loss": 1.6763,
|
| 5308 |
+
"step": 757
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.9805950840879689,
|
| 5312 |
+
"grad_norm": 0.531313955783844,
|
| 5313 |
+
"learning_rate": 9.409435499254104e-08,
|
| 5314 |
+
"loss": 1.6581,
|
| 5315 |
+
"step": 758
|
| 5316 |
+
},
|
| 5317 |
+
{
|
| 5318 |
+
"epoch": 0.981888745148771,
|
| 5319 |
+
"grad_norm": 0.5821428894996643,
|
| 5320 |
+
"learning_rate": 8.196995298887511e-08,
|
| 5321 |
+
"loss": 1.7181,
|
| 5322 |
+
"step": 759
|
| 5323 |
+
},
|
| 5324 |
+
{
|
| 5325 |
+
"epoch": 0.9831824062095731,
|
| 5326 |
+
"grad_norm": 0.5947988033294678,
|
| 5327 |
+
"learning_rate": 7.068083493779743e-08,
|
| 5328 |
+
"loss": 1.8464,
|
| 5329 |
+
"step": 760
|
| 5330 |
+
},
|
| 5331 |
+
{
|
| 5332 |
+
"epoch": 0.9844760672703752,
|
| 5333 |
+
"grad_norm": 0.5871140956878662,
|
| 5334 |
+
"learning_rate": 6.022718974137975e-08,
|
| 5335 |
+
"loss": 1.7817,
|
| 5336 |
+
"step": 761
|
| 5337 |
+
},
|
| 5338 |
+
{
|
| 5339 |
+
"epoch": 0.9857697283311773,
|
| 5340 |
+
"grad_norm": 0.6064710021018982,
|
| 5341 |
+
"learning_rate": 5.060919232162675e-08,
|
| 5342 |
+
"loss": 1.6367,
|
| 5343 |
+
"step": 762
|
| 5344 |
+
},
|
| 5345 |
+
{
|
| 5346 |
+
"epoch": 0.9870633893919794,
|
| 5347 |
+
"grad_norm": 0.613859236240387,
|
| 5348 |
+
"learning_rate": 4.182700361756164e-08,
|
| 5349 |
+
"loss": 1.8499,
|
| 5350 |
+
"step": 763
|
| 5351 |
+
},
|
| 5352 |
+
{
|
| 5353 |
+
"epoch": 0.9883570504527813,
|
| 5354 |
+
"grad_norm": 0.6105691194534302,
|
| 5355 |
+
"learning_rate": 3.388077058252281e-08,
|
| 5356 |
+
"loss": 1.7682,
|
| 5357 |
+
"step": 764
|
| 5358 |
+
},
|
| 5359 |
+
{
|
| 5360 |
+
"epoch": 0.9896507115135834,
|
| 5361 |
+
"grad_norm": 0.6665717959403992,
|
| 5362 |
+
"learning_rate": 2.6770626181715773e-08,
|
| 5363 |
+
"loss": 1.8694,
|
| 5364 |
+
"step": 765
|
| 5365 |
+
},
|
| 5366 |
+
{
|
| 5367 |
+
"epoch": 0.9909443725743855,
|
| 5368 |
+
"grad_norm": 0.6197919249534607,
|
| 5369 |
+
"learning_rate": 2.0496689389976065e-08,
|
| 5370 |
+
"loss": 1.8789,
|
| 5371 |
+
"step": 766
|
| 5372 |
+
},
|
| 5373 |
+
{
|
| 5374 |
+
"epoch": 0.9922380336351876,
|
| 5375 |
+
"grad_norm": 0.6365950107574463,
|
| 5376 |
+
"learning_rate": 1.5059065189787503e-08,
|
| 5377 |
+
"loss": 1.8171,
|
| 5378 |
+
"step": 767
|
| 5379 |
+
},
|
| 5380 |
+
{
|
| 5381 |
+
"epoch": 0.9935316946959897,
|
| 5382 |
+
"grad_norm": 0.6993787884712219,
|
| 5383 |
+
"learning_rate": 1.045784456952248e-08,
|
| 5384 |
+
"loss": 2.0328,
|
| 5385 |
+
"step": 768
|
| 5386 |
+
},
|
| 5387 |
+
{
|
| 5388 |
+
"epoch": 0.9948253557567918,
|
| 5389 |
+
"grad_norm": 0.7297689318656921,
|
| 5390 |
+
"learning_rate": 6.693104521909854e-09,
|
| 5391 |
+
"loss": 2.0254,
|
| 5392 |
+
"step": 769
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 0.9961190168175937,
|
| 5396 |
+
"grad_norm": 0.7178072333335876,
|
| 5397 |
+
"learning_rate": 3.764908042774851e-09,
|
| 5398 |
+
"loss": 1.8857,
|
| 5399 |
+
"step": 770
|
| 5400 |
+
},
|
| 5401 |
+
{
|
| 5402 |
+
"epoch": 0.9974126778783958,
|
| 5403 |
+
"grad_norm": 0.7650882005691528,
|
| 5404 |
+
"learning_rate": 1.6733041299454855e-09,
|
| 5405 |
+
"loss": 1.8953,
|
| 5406 |
+
"step": 771
|
| 5407 |
+
},
|
| 5408 |
+
{
|
| 5409 |
+
"epoch": 0.9987063389391979,
|
| 5410 |
+
"grad_norm": 0.9221259951591492,
|
| 5411 |
+
"learning_rate": 4.183277824698628e-10,
|
| 5412 |
+
"loss": 1.9443,
|
| 5413 |
+
"step": 772
|
| 5414 |
+
},
|
| 5415 |
+
{
|
| 5416 |
+
"epoch": 1.0,
|
| 5417 |
+
"grad_norm": 1.0081813335418701,
|
| 5418 |
+
"learning_rate": 0.0,
|
| 5419 |
+
"loss": 2.1013,
|
| 5420 |
+
"step": 773
|
| 5421 |
}
|
| 5422 |
],
|
| 5423 |
"logging_steps": 1,
|
|
|
|
| 5432 |
"should_evaluate": false,
|
| 5433 |
"should_log": false,
|
| 5434 |
"should_save": true,
|
| 5435 |
+
"should_training_stop": true
|
| 5436 |
},
|
| 5437 |
"attributes": {}
|
| 5438 |
}
|
| 5439 |
},
|
| 5440 |
+
"total_flos": 5.3869236774764544e+17,
|
| 5441 |
"train_batch_size": 4,
|
| 5442 |
"trial_name": null,
|
| 5443 |
"trial_params": null
|