Fanucci
commited on
Training in progress, step 600, 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 671149168
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf23c078e9609fce3d6406e8cf3318bd1c803b0e4a1390242292fdcc82f0a604
|
| 3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1342555602
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f445125568e8c2ec052c8043e5d4738490c13a3f54d07729fc25a1f8f8f74204
|
| 3 |
size 1342555602
|
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 sha256:72fbfa530194f7b2cf0e13b6a439f72a3da375141106db14ca42e185603cf635
|
| 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:fe80b57817b0d8cc8f62cfc1dc5ac6112abe389906e014f58f93bc482e7e6c0e
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 3.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2831,6 +2831,1414 @@
|
|
| 2831 |
"eval_samples_per_second": 5.634,
|
| 2832 |
"eval_steps_per_second": 1.409,
|
| 2833 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2834 |
}
|
| 2835 |
],
|
| 2836 |
"logging_steps": 1,
|
|
@@ -2859,7 +4267,7 @@
|
|
| 2859 |
"attributes": {}
|
| 2860 |
}
|
| 2861 |
},
|
| 2862 |
-
"total_flos":
|
| 2863 |
"train_batch_size": 4,
|
| 2864 |
"trial_name": null,
|
| 2865 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 3.1351470947265625,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
+
"epoch": 0.1919692849144137,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2831 |
"eval_samples_per_second": 5.634,
|
| 2832 |
"eval_steps_per_second": 1.409,
|
| 2833 |
"step": 400
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 0.1282994720844665,
|
| 2837 |
+
"grad_norm": 15.050670623779297,
|
| 2838 |
+
"learning_rate": 0.0003847980611605696,
|
| 2839 |
+
"loss": 6.5357,
|
| 2840 |
+
"step": 401
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 0.12861942089265718,
|
| 2844 |
+
"grad_norm": 10.210688591003418,
|
| 2845 |
+
"learning_rate": 0.00038472126307432245,
|
| 2846 |
+
"loss": 6.5739,
|
| 2847 |
+
"step": 402
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.12893936970084788,
|
| 2851 |
+
"grad_norm": 9.43270492553711,
|
| 2852 |
+
"learning_rate": 0.00038464427961536955,
|
| 2853 |
+
"loss": 6.8275,
|
| 2854 |
+
"step": 403
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.12925931850903855,
|
| 2858 |
+
"grad_norm": 8.865079879760742,
|
| 2859 |
+
"learning_rate": 0.0003845671108653876,
|
| 2860 |
+
"loss": 6.6035,
|
| 2861 |
+
"step": 404
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.12957926731722924,
|
| 2865 |
+
"grad_norm": 8.50759506225586,
|
| 2866 |
+
"learning_rate": 0.0003844897569062497,
|
| 2867 |
+
"loss": 6.4124,
|
| 2868 |
+
"step": 405
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.12989921612541994,
|
| 2872 |
+
"grad_norm": 10.129754066467285,
|
| 2873 |
+
"learning_rate": 0.00038441221782002546,
|
| 2874 |
+
"loss": 6.5082,
|
| 2875 |
+
"step": 406
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.13021916493361063,
|
| 2879 |
+
"grad_norm": 8.176865577697754,
|
| 2880 |
+
"learning_rate": 0.000384334493688981,
|
| 2881 |
+
"loss": 6.3945,
|
| 2882 |
+
"step": 407
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.1305391137418013,
|
| 2886 |
+
"grad_norm": 7.027590751647949,
|
| 2887 |
+
"learning_rate": 0.0003842565845955786,
|
| 2888 |
+
"loss": 6.3583,
|
| 2889 |
+
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.130859062549992,
|
| 2893 |
+
"grad_norm": 9.852948188781738,
|
| 2894 |
+
"learning_rate": 0.000384178490622477,
|
| 2895 |
+
"loss": 6.1044,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.1311790113581827,
|
| 2900 |
+
"grad_norm": 8.926599502563477,
|
| 2901 |
+
"learning_rate": 0.0003841002118525309,
|
| 2902 |
+
"loss": 6.2928,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.13149896016637339,
|
| 2907 |
+
"grad_norm": 8.307205200195312,
|
| 2908 |
+
"learning_rate": 0.0003840217483687912,
|
| 2909 |
+
"loss": 6.5742,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.13181890897456408,
|
| 2914 |
+
"grad_norm": 7.831887722015381,
|
| 2915 |
+
"learning_rate": 0.0003839431002545047,
|
| 2916 |
+
"loss": 6.2626,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.13213885778275475,
|
| 2921 |
+
"grad_norm": 7.879708766937256,
|
| 2922 |
+
"learning_rate": 0.0003838642675931141,
|
| 2923 |
+
"loss": 6.5028,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.13245880659094544,
|
| 2928 |
+
"grad_norm": 5.989245891571045,
|
| 2929 |
+
"learning_rate": 0.00038378525046825783,
|
| 2930 |
+
"loss": 6.3706,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.13277875539913614,
|
| 2935 |
+
"grad_norm": 7.094750881195068,
|
| 2936 |
+
"learning_rate": 0.0003837060489637702,
|
| 2937 |
+
"loss": 6.2248,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.13309870420732683,
|
| 2942 |
+
"grad_norm": 7.013360977172852,
|
| 2943 |
+
"learning_rate": 0.00038362666316368095,
|
| 2944 |
+
"loss": 6.6161,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.13341865301551753,
|
| 2949 |
+
"grad_norm": 6.406157970428467,
|
| 2950 |
+
"learning_rate": 0.00038354709315221547,
|
| 2951 |
+
"loss": 6.4046,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.1337386018237082,
|
| 2956 |
+
"grad_norm": 7.2502055168151855,
|
| 2957 |
+
"learning_rate": 0.00038346733901379454,
|
| 2958 |
+
"loss": 6.2165,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.1340585506318989,
|
| 2963 |
+
"grad_norm": 7.486465930938721,
|
| 2964 |
+
"learning_rate": 0.00038338740083303436,
|
| 2965 |
+
"loss": 6.3809,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.1343784994400896,
|
| 2970 |
+
"grad_norm": 4.760215759277344,
|
| 2971 |
+
"learning_rate": 0.0003833072786947463,
|
| 2972 |
+
"loss": 6.2624,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.13469844824828028,
|
| 2977 |
+
"grad_norm": 7.41804838180542,
|
| 2978 |
+
"learning_rate": 0.0003832269726839369,
|
| 2979 |
+
"loss": 6.219,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.13501839705647095,
|
| 2984 |
+
"grad_norm": 5.859868049621582,
|
| 2985 |
+
"learning_rate": 0.00038314648288580795,
|
| 2986 |
+
"loss": 6.2853,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.13533834586466165,
|
| 2991 |
+
"grad_norm": 7.259220123291016,
|
| 2992 |
+
"learning_rate": 0.000383065809385756,
|
| 2993 |
+
"loss": 6.4797,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.13565829467285234,
|
| 2998 |
+
"grad_norm": 5.617652893066406,
|
| 2999 |
+
"learning_rate": 0.0003829849522693727,
|
| 3000 |
+
"loss": 6.1839,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.13597824348104304,
|
| 3005 |
+
"grad_norm": 7.392098426818848,
|
| 3006 |
+
"learning_rate": 0.00038290391162244423,
|
| 3007 |
+
"loss": 6.3839,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.13629819228923373,
|
| 3012 |
+
"grad_norm": 5.7139997482299805,
|
| 3013 |
+
"learning_rate": 0.0003828226875309518,
|
| 3014 |
+
"loss": 6.2123,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.1366181410974244,
|
| 3019 |
+
"grad_norm": 5.242659568786621,
|
| 3020 |
+
"learning_rate": 0.0003827412800810712,
|
| 3021 |
+
"loss": 6.3428,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.1369380899056151,
|
| 3026 |
+
"grad_norm": 5.612354278564453,
|
| 3027 |
+
"learning_rate": 0.00038265968935917254,
|
| 3028 |
+
"loss": 6.3251,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.1372580387138058,
|
| 3033 |
+
"grad_norm": 5.880339622497559,
|
| 3034 |
+
"learning_rate": 0.0003825779154518207,
|
| 3035 |
+
"loss": 6.3662,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.1375779875219965,
|
| 3040 |
+
"grad_norm": 5.417006492614746,
|
| 3041 |
+
"learning_rate": 0.00038249595844577446,
|
| 3042 |
+
"loss": 6.4292,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.13789793633018718,
|
| 3047 |
+
"grad_norm": 8.451333999633789,
|
| 3048 |
+
"learning_rate": 0.0003824138184279874,
|
| 3049 |
+
"loss": 6.3248,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.13821788513837785,
|
| 3054 |
+
"grad_norm": 7.415943622589111,
|
| 3055 |
+
"learning_rate": 0.00038233149548560683,
|
| 3056 |
+
"loss": 6.3606,
|
| 3057 |
+
"step": 432
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.13853783394656854,
|
| 3061 |
+
"grad_norm": 4.678287029266357,
|
| 3062 |
+
"learning_rate": 0.00038224898970597426,
|
| 3063 |
+
"loss": 6.2415,
|
| 3064 |
+
"step": 433
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.13885778275475924,
|
| 3068 |
+
"grad_norm": 7.330494403839111,
|
| 3069 |
+
"learning_rate": 0.0003821663011766254,
|
| 3070 |
+
"loss": 6.4611,
|
| 3071 |
+
"step": 434
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.13917773156294994,
|
| 3075 |
+
"grad_norm": 5.664525985717773,
|
| 3076 |
+
"learning_rate": 0.0003820834299852896,
|
| 3077 |
+
"loss": 6.0102,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.13949768037114063,
|
| 3082 |
+
"grad_norm": 4.248838901519775,
|
| 3083 |
+
"learning_rate": 0.00038200037621989005,
|
| 3084 |
+
"loss": 6.2752,
|
| 3085 |
+
"step": 436
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.1398176291793313,
|
| 3089 |
+
"grad_norm": 7.762814998626709,
|
| 3090 |
+
"learning_rate": 0.0003819171399685437,
|
| 3091 |
+
"loss": 6.0838,
|
| 3092 |
+
"step": 437
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.140137577987522,
|
| 3096 |
+
"grad_norm": 6.642805576324463,
|
| 3097 |
+
"learning_rate": 0.00038183372131956114,
|
| 3098 |
+
"loss": 6.055,
|
| 3099 |
+
"step": 438
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.1404575267957127,
|
| 3103 |
+
"grad_norm": 12.796443939208984,
|
| 3104 |
+
"learning_rate": 0.0003817501203614464,
|
| 3105 |
+
"loss": 6.1023,
|
| 3106 |
+
"step": 439
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.14077747560390338,
|
| 3110 |
+
"grad_norm": 13.292308807373047,
|
| 3111 |
+
"learning_rate": 0.0003816663371828969,
|
| 3112 |
+
"loss": 6.6153,
|
| 3113 |
+
"step": 440
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.14109742441209405,
|
| 3117 |
+
"grad_norm": 5.322185516357422,
|
| 3118 |
+
"learning_rate": 0.00038158237187280343,
|
| 3119 |
+
"loss": 6.0856,
|
| 3120 |
+
"step": 441
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.14141737322028475,
|
| 3124 |
+
"grad_norm": 9.385347366333008,
|
| 3125 |
+
"learning_rate": 0.00038149822452025024,
|
| 3126 |
+
"loss": 6.1498,
|
| 3127 |
+
"step": 442
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.14173732202847544,
|
| 3131 |
+
"grad_norm": 8.704651832580566,
|
| 3132 |
+
"learning_rate": 0.00038141389521451427,
|
| 3133 |
+
"loss": 6.1346,
|
| 3134 |
+
"step": 443
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.14205727083666614,
|
| 3138 |
+
"grad_norm": 6.753723621368408,
|
| 3139 |
+
"learning_rate": 0.0003813293840450659,
|
| 3140 |
+
"loss": 6.3897,
|
| 3141 |
+
"step": 444
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 0.14237721964485683,
|
| 3145 |
+
"grad_norm": 9.21335220336914,
|
| 3146 |
+
"learning_rate": 0.0003812446911015682,
|
| 3147 |
+
"loss": 6.3132,
|
| 3148 |
+
"step": 445
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.1426971684530475,
|
| 3152 |
+
"grad_norm": 6.543299674987793,
|
| 3153 |
+
"learning_rate": 0.00038115981647387733,
|
| 3154 |
+
"loss": 6.2129,
|
| 3155 |
+
"step": 446
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 0.1430171172612382,
|
| 3159 |
+
"grad_norm": 21.838804244995117,
|
| 3160 |
+
"learning_rate": 0.000381074760252042,
|
| 3161 |
+
"loss": 6.3327,
|
| 3162 |
+
"step": 447
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 0.1433370660694289,
|
| 3166 |
+
"grad_norm": 28.30170440673828,
|
| 3167 |
+
"learning_rate": 0.00038098952252630373,
|
| 3168 |
+
"loss": 6.2693,
|
| 3169 |
+
"step": 448
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 0.1436570148776196,
|
| 3173 |
+
"grad_norm": 20.272245407104492,
|
| 3174 |
+
"learning_rate": 0.0003809041033870965,
|
| 3175 |
+
"loss": 6.173,
|
| 3176 |
+
"step": 449
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 0.14397696368581028,
|
| 3180 |
+
"grad_norm": 11.198784828186035,
|
| 3181 |
+
"learning_rate": 0.00038081850292504677,
|
| 3182 |
+
"loss": 6.2571,
|
| 3183 |
+
"step": 450
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 0.14429691249400095,
|
| 3187 |
+
"grad_norm": 6.138251304626465,
|
| 3188 |
+
"learning_rate": 0.00038073272123097354,
|
| 3189 |
+
"loss": 6.2325,
|
| 3190 |
+
"step": 451
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"epoch": 0.14461686130219165,
|
| 3194 |
+
"grad_norm": 8.177600860595703,
|
| 3195 |
+
"learning_rate": 0.0003806467583958879,
|
| 3196 |
+
"loss": 6.3511,
|
| 3197 |
+
"step": 452
|
| 3198 |
+
},
|
| 3199 |
+
{
|
| 3200 |
+
"epoch": 0.14493681011038234,
|
| 3201 |
+
"grad_norm": 6.5708184242248535,
|
| 3202 |
+
"learning_rate": 0.0003805606145109932,
|
| 3203 |
+
"loss": 6.073,
|
| 3204 |
+
"step": 453
|
| 3205 |
+
},
|
| 3206 |
+
{
|
| 3207 |
+
"epoch": 0.14525675891857304,
|
| 3208 |
+
"grad_norm": 7.593567848205566,
|
| 3209 |
+
"learning_rate": 0.00038047428966768486,
|
| 3210 |
+
"loss": 6.5693,
|
| 3211 |
+
"step": 454
|
| 3212 |
+
},
|
| 3213 |
+
{
|
| 3214 |
+
"epoch": 0.1455767077267637,
|
| 3215 |
+
"grad_norm": 10.439419746398926,
|
| 3216 |
+
"learning_rate": 0.00038038778395755034,
|
| 3217 |
+
"loss": 6.4299,
|
| 3218 |
+
"step": 455
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 0.1458966565349544,
|
| 3222 |
+
"grad_norm": 11.682433128356934,
|
| 3223 |
+
"learning_rate": 0.00038030109747236887,
|
| 3224 |
+
"loss": 6.3646,
|
| 3225 |
+
"step": 456
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 0.1462166053431451,
|
| 3229 |
+
"grad_norm": 6.629786491394043,
|
| 3230 |
+
"learning_rate": 0.00038021423030411164,
|
| 3231 |
+
"loss": 6.2166,
|
| 3232 |
+
"step": 457
|
| 3233 |
+
},
|
| 3234 |
+
{
|
| 3235 |
+
"epoch": 0.1465365541513358,
|
| 3236 |
+
"grad_norm": 9.529027938842773,
|
| 3237 |
+
"learning_rate": 0.0003801271825449414,
|
| 3238 |
+
"loss": 6.7073,
|
| 3239 |
+
"step": 458
|
| 3240 |
+
},
|
| 3241 |
+
{
|
| 3242 |
+
"epoch": 0.14685650295952649,
|
| 3243 |
+
"grad_norm": 13.9060697555542,
|
| 3244 |
+
"learning_rate": 0.0003800399542872126,
|
| 3245 |
+
"loss": 6.2038,
|
| 3246 |
+
"step": 459
|
| 3247 |
+
},
|
| 3248 |
+
{
|
| 3249 |
+
"epoch": 0.14717645176771715,
|
| 3250 |
+
"grad_norm": 7.617699146270752,
|
| 3251 |
+
"learning_rate": 0.0003799525456234712,
|
| 3252 |
+
"loss": 6.247,
|
| 3253 |
+
"step": 460
|
| 3254 |
+
},
|
| 3255 |
+
{
|
| 3256 |
+
"epoch": 0.14749640057590785,
|
| 3257 |
+
"grad_norm": 6.537087440490723,
|
| 3258 |
+
"learning_rate": 0.0003798649566464544,
|
| 3259 |
+
"loss": 6.3626,
|
| 3260 |
+
"step": 461
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 0.14781634938409854,
|
| 3264 |
+
"grad_norm": 5.183995723724365,
|
| 3265 |
+
"learning_rate": 0.00037977718744909084,
|
| 3266 |
+
"loss": 6.332,
|
| 3267 |
+
"step": 462
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.14813629819228924,
|
| 3271 |
+
"grad_norm": 6.627659797668457,
|
| 3272 |
+
"learning_rate": 0.00037968923812450037,
|
| 3273 |
+
"loss": 6.3186,
|
| 3274 |
+
"step": 463
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 0.14845624700047994,
|
| 3278 |
+
"grad_norm": 5.604552745819092,
|
| 3279 |
+
"learning_rate": 0.00037960110876599396,
|
| 3280 |
+
"loss": 6.3636,
|
| 3281 |
+
"step": 464
|
| 3282 |
+
},
|
| 3283 |
+
{
|
| 3284 |
+
"epoch": 0.1487761958086706,
|
| 3285 |
+
"grad_norm": 5.800177097320557,
|
| 3286 |
+
"learning_rate": 0.0003795127994670735,
|
| 3287 |
+
"loss": 6.4762,
|
| 3288 |
+
"step": 465
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 0.1490961446168613,
|
| 3292 |
+
"grad_norm": 7.276813983917236,
|
| 3293 |
+
"learning_rate": 0.0003794243103214319,
|
| 3294 |
+
"loss": 6.2994,
|
| 3295 |
+
"step": 466
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.149416093425052,
|
| 3299 |
+
"grad_norm": 7.041287422180176,
|
| 3300 |
+
"learning_rate": 0.00037933564142295274,
|
| 3301 |
+
"loss": 6.3795,
|
| 3302 |
+
"step": 467
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.1497360422332427,
|
| 3306 |
+
"grad_norm": 7.782250881195068,
|
| 3307 |
+
"learning_rate": 0.0003792467928657105,
|
| 3308 |
+
"loss": 6.506,
|
| 3309 |
+
"step": 468
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.15005599104143338,
|
| 3313 |
+
"grad_norm": 8.970823287963867,
|
| 3314 |
+
"learning_rate": 0.0003791577647439701,
|
| 3315 |
+
"loss": 6.4383,
|
| 3316 |
+
"step": 469
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.15037593984962405,
|
| 3320 |
+
"grad_norm": 7.6800360679626465,
|
| 3321 |
+
"learning_rate": 0.00037906855715218695,
|
| 3322 |
+
"loss": 6.307,
|
| 3323 |
+
"step": 470
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.15069588865781475,
|
| 3327 |
+
"grad_norm": 10.980112075805664,
|
| 3328 |
+
"learning_rate": 0.0003789791701850072,
|
| 3329 |
+
"loss": 6.4015,
|
| 3330 |
+
"step": 471
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.15101583746600544,
|
| 3334 |
+
"grad_norm": 6.547383785247803,
|
| 3335 |
+
"learning_rate": 0.00037888960393726677,
|
| 3336 |
+
"loss": 6.229,
|
| 3337 |
+
"step": 472
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.15133578627419614,
|
| 3341 |
+
"grad_norm": 8.423219680786133,
|
| 3342 |
+
"learning_rate": 0.0003787998585039923,
|
| 3343 |
+
"loss": 6.2121,
|
| 3344 |
+
"step": 473
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.1516557350823868,
|
| 3348 |
+
"grad_norm": 5.707985877990723,
|
| 3349 |
+
"learning_rate": 0.0003787099339804003,
|
| 3350 |
+
"loss": 6.3282,
|
| 3351 |
+
"step": 474
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.1519756838905775,
|
| 3355 |
+
"grad_norm": 6.869352340698242,
|
| 3356 |
+
"learning_rate": 0.0003786198304618973,
|
| 3357 |
+
"loss": 6.1333,
|
| 3358 |
+
"step": 475
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.1522956326987682,
|
| 3362 |
+
"grad_norm": 6.416158676147461,
|
| 3363 |
+
"learning_rate": 0.00037852954804407974,
|
| 3364 |
+
"loss": 6.37,
|
| 3365 |
+
"step": 476
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.1526155815069589,
|
| 3369 |
+
"grad_norm": 8.767664909362793,
|
| 3370 |
+
"learning_rate": 0.0003784390868227339,
|
| 3371 |
+
"loss": 6.1501,
|
| 3372 |
+
"step": 477
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.1529355303151496,
|
| 3376 |
+
"grad_norm": 10.395624160766602,
|
| 3377 |
+
"learning_rate": 0.0003783484468938356,
|
| 3378 |
+
"loss": 6.5037,
|
| 3379 |
+
"step": 478
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.15325547912334025,
|
| 3383 |
+
"grad_norm": 6.741879940032959,
|
| 3384 |
+
"learning_rate": 0.0003782576283535506,
|
| 3385 |
+
"loss": 6.4512,
|
| 3386 |
+
"step": 479
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.15357542793153095,
|
| 3390 |
+
"grad_norm": 9.098990440368652,
|
| 3391 |
+
"learning_rate": 0.00037816663129823394,
|
| 3392 |
+
"loss": 6.2566,
|
| 3393 |
+
"step": 480
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.15389537673972165,
|
| 3397 |
+
"grad_norm": 10.70953369140625,
|
| 3398 |
+
"learning_rate": 0.0003780754558244299,
|
| 3399 |
+
"loss": 6.1686,
|
| 3400 |
+
"step": 481
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.15421532554791234,
|
| 3404 |
+
"grad_norm": 6.690832138061523,
|
| 3405 |
+
"learning_rate": 0.0003779841020288725,
|
| 3406 |
+
"loss": 6.1794,
|
| 3407 |
+
"step": 482
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.15453527435610304,
|
| 3411 |
+
"grad_norm": 6.798882484436035,
|
| 3412 |
+
"learning_rate": 0.0003778925700084845,
|
| 3413 |
+
"loss": 6.0711,
|
| 3414 |
+
"step": 483
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.1548552231642937,
|
| 3418 |
+
"grad_norm": 11.702363014221191,
|
| 3419 |
+
"learning_rate": 0.00037780085986037796,
|
| 3420 |
+
"loss": 6.6549,
|
| 3421 |
+
"step": 484
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.1551751719724844,
|
| 3425 |
+
"grad_norm": 8.392105102539062,
|
| 3426 |
+
"learning_rate": 0.000377708971681854,
|
| 3427 |
+
"loss": 6.1966,
|
| 3428 |
+
"step": 485
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.1554951207806751,
|
| 3432 |
+
"grad_norm": 9.752997398376465,
|
| 3433 |
+
"learning_rate": 0.0003776169055704025,
|
| 3434 |
+
"loss": 6.8049,
|
| 3435 |
+
"step": 486
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.1558150695888658,
|
| 3439 |
+
"grad_norm": 55.3693733215332,
|
| 3440 |
+
"learning_rate": 0.00037752466162370207,
|
| 3441 |
+
"loss": 6.1044,
|
| 3442 |
+
"step": 487
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.15613501839705646,
|
| 3446 |
+
"grad_norm": 65.37632751464844,
|
| 3447 |
+
"learning_rate": 0.0003774322399396202,
|
| 3448 |
+
"loss": 6.4519,
|
| 3449 |
+
"step": 488
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.15645496720524715,
|
| 3453 |
+
"grad_norm": 7.421685218811035,
|
| 3454 |
+
"learning_rate": 0.0003773396406162127,
|
| 3455 |
+
"loss": 6.3113,
|
| 3456 |
+
"step": 489
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.15677491601343785,
|
| 3460 |
+
"grad_norm": 9.326229095458984,
|
| 3461 |
+
"learning_rate": 0.00037724686375172403,
|
| 3462 |
+
"loss": 6.4404,
|
| 3463 |
+
"step": 490
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.15709486482162854,
|
| 3467 |
+
"grad_norm": 8.919943809509277,
|
| 3468 |
+
"learning_rate": 0.000377153909444587,
|
| 3469 |
+
"loss": 6.3796,
|
| 3470 |
+
"step": 491
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.15741481362981924,
|
| 3474 |
+
"grad_norm": 12.594194412231445,
|
| 3475 |
+
"learning_rate": 0.00037706077779342263,
|
| 3476 |
+
"loss": 6.1395,
|
| 3477 |
+
"step": 492
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.1577347624380099,
|
| 3481 |
+
"grad_norm": 13.710543632507324,
|
| 3482 |
+
"learning_rate": 0.0003769674688970401,
|
| 3483 |
+
"loss": 6.1843,
|
| 3484 |
+
"step": 493
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.1580547112462006,
|
| 3488 |
+
"grad_norm": 7.772939682006836,
|
| 3489 |
+
"learning_rate": 0.0003768739828544366,
|
| 3490 |
+
"loss": 6.2831,
|
| 3491 |
+
"step": 494
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.1583746600543913,
|
| 3495 |
+
"grad_norm": 89.74982452392578,
|
| 3496 |
+
"learning_rate": 0.00037678031976479733,
|
| 3497 |
+
"loss": 6.3358,
|
| 3498 |
+
"step": 495
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.158694608862582,
|
| 3502 |
+
"grad_norm": 9.788384437561035,
|
| 3503 |
+
"learning_rate": 0.0003766864797274954,
|
| 3504 |
+
"loss": 6.4116,
|
| 3505 |
+
"step": 496
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.1590145576707727,
|
| 3509 |
+
"grad_norm": 13.687607765197754,
|
| 3510 |
+
"learning_rate": 0.0003765924628420916,
|
| 3511 |
+
"loss": 6.3342,
|
| 3512 |
+
"step": 497
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.15933450647896336,
|
| 3516 |
+
"grad_norm": 10.730764389038086,
|
| 3517 |
+
"learning_rate": 0.0003764982692083341,
|
| 3518 |
+
"loss": 6.3028,
|
| 3519 |
+
"step": 498
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.15965445528715405,
|
| 3523 |
+
"grad_norm": 7.144099712371826,
|
| 3524 |
+
"learning_rate": 0.00037640389892615897,
|
| 3525 |
+
"loss": 6.1418,
|
| 3526 |
+
"step": 499
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.15997440409534475,
|
| 3530 |
+
"grad_norm": 6.012965202331543,
|
| 3531 |
+
"learning_rate": 0.00037630935209568954,
|
| 3532 |
+
"loss": 6.3804,
|
| 3533 |
+
"step": 500
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.16029435290353544,
|
| 3537 |
+
"grad_norm": 19.48784065246582,
|
| 3538 |
+
"learning_rate": 0.0003762146288172364,
|
| 3539 |
+
"loss": 6.3017,
|
| 3540 |
+
"step": 501
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.16061430171172614,
|
| 3544 |
+
"grad_norm": 10.153864860534668,
|
| 3545 |
+
"learning_rate": 0.0003761197291912974,
|
| 3546 |
+
"loss": 6.2839,
|
| 3547 |
+
"step": 502
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.1609342505199168,
|
| 3551 |
+
"grad_norm": 10.028372764587402,
|
| 3552 |
+
"learning_rate": 0.00037602465331855754,
|
| 3553 |
+
"loss": 6.1487,
|
| 3554 |
+
"step": 503
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.1612541993281075,
|
| 3558 |
+
"grad_norm": 7.426041126251221,
|
| 3559 |
+
"learning_rate": 0.00037592940129988866,
|
| 3560 |
+
"loss": 6.3812,
|
| 3561 |
+
"step": 504
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.1615741481362982,
|
| 3565 |
+
"grad_norm": 5.519888877868652,
|
| 3566 |
+
"learning_rate": 0.00037583397323634964,
|
| 3567 |
+
"loss": 6.1925,
|
| 3568 |
+
"step": 505
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.1618940969444889,
|
| 3572 |
+
"grad_norm": 7.253342628479004,
|
| 3573 |
+
"learning_rate": 0.00037573836922918615,
|
| 3574 |
+
"loss": 6.2379,
|
| 3575 |
+
"step": 506
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.16221404575267956,
|
| 3579 |
+
"grad_norm": 5.459297180175781,
|
| 3580 |
+
"learning_rate": 0.00037564258937983035,
|
| 3581 |
+
"loss": 6.3595,
|
| 3582 |
+
"step": 507
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.16253399456087025,
|
| 3586 |
+
"grad_norm": 4.96316385269165,
|
| 3587 |
+
"learning_rate": 0.000375546633789901,
|
| 3588 |
+
"loss": 6.0939,
|
| 3589 |
+
"step": 508
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.16285394336906095,
|
| 3593 |
+
"grad_norm": 4.441860675811768,
|
| 3594 |
+
"learning_rate": 0.00037545050256120353,
|
| 3595 |
+
"loss": 6.1986,
|
| 3596 |
+
"step": 509
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 0.16317389217725164,
|
| 3600 |
+
"grad_norm": 9.1715087890625,
|
| 3601 |
+
"learning_rate": 0.0003753541957957295,
|
| 3602 |
+
"loss": 6.2663,
|
| 3603 |
+
"step": 510
|
| 3604 |
+
},
|
| 3605 |
+
{
|
| 3606 |
+
"epoch": 0.16349384098544234,
|
| 3607 |
+
"grad_norm": 5.704549789428711,
|
| 3608 |
+
"learning_rate": 0.0003752577135956568,
|
| 3609 |
+
"loss": 6.0975,
|
| 3610 |
+
"step": 511
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 0.163813789793633,
|
| 3614 |
+
"grad_norm": 6.922268390655518,
|
| 3615 |
+
"learning_rate": 0.00037516105606334943,
|
| 3616 |
+
"loss": 6.4677,
|
| 3617 |
+
"step": 512
|
| 3618 |
+
},
|
| 3619 |
+
{
|
| 3620 |
+
"epoch": 0.1641337386018237,
|
| 3621 |
+
"grad_norm": 5.864443778991699,
|
| 3622 |
+
"learning_rate": 0.0003750642233013573,
|
| 3623 |
+
"loss": 6.0617,
|
| 3624 |
+
"step": 513
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"epoch": 0.1644536874100144,
|
| 3628 |
+
"grad_norm": 6.080179691314697,
|
| 3629 |
+
"learning_rate": 0.0003749672154124165,
|
| 3630 |
+
"loss": 6.2692,
|
| 3631 |
+
"step": 514
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 0.1647736362182051,
|
| 3635 |
+
"grad_norm": 7.958470344543457,
|
| 3636 |
+
"learning_rate": 0.00037487003249944863,
|
| 3637 |
+
"loss": 6.3754,
|
| 3638 |
+
"step": 515
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 0.1650935850263958,
|
| 3642 |
+
"grad_norm": 7.878583908081055,
|
| 3643 |
+
"learning_rate": 0.0003747726746655612,
|
| 3644 |
+
"loss": 6.3376,
|
| 3645 |
+
"step": 516
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 0.16541353383458646,
|
| 3649 |
+
"grad_norm": 6.210682392120361,
|
| 3650 |
+
"learning_rate": 0.00037467514201404726,
|
| 3651 |
+
"loss": 6.4979,
|
| 3652 |
+
"step": 517
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 0.16573348264277715,
|
| 3656 |
+
"grad_norm": 13.755587577819824,
|
| 3657 |
+
"learning_rate": 0.0003745774346483851,
|
| 3658 |
+
"loss": 6.4135,
|
| 3659 |
+
"step": 518
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 0.16605343145096785,
|
| 3663 |
+
"grad_norm": 8.76400375366211,
|
| 3664 |
+
"learning_rate": 0.00037447955267223875,
|
| 3665 |
+
"loss": 6.1651,
|
| 3666 |
+
"step": 519
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 0.16637338025915854,
|
| 3670 |
+
"grad_norm": 6.616308212280273,
|
| 3671 |
+
"learning_rate": 0.00037438149618945727,
|
| 3672 |
+
"loss": 6.2581,
|
| 3673 |
+
"step": 520
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.1666933290673492,
|
| 3677 |
+
"grad_norm": 8.678804397583008,
|
| 3678 |
+
"learning_rate": 0.0003742832653040748,
|
| 3679 |
+
"loss": 6.3311,
|
| 3680 |
+
"step": 521
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 0.1670132778755399,
|
| 3684 |
+
"grad_norm": 6.813952445983887,
|
| 3685 |
+
"learning_rate": 0.00037418486012031087,
|
| 3686 |
+
"loss": 6.2027,
|
| 3687 |
+
"step": 522
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 0.1673332266837306,
|
| 3691 |
+
"grad_norm": 8.717082023620605,
|
| 3692 |
+
"learning_rate": 0.0003740862807425694,
|
| 3693 |
+
"loss": 6.3091,
|
| 3694 |
+
"step": 523
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 0.1676531754919213,
|
| 3698 |
+
"grad_norm": 7.110479354858398,
|
| 3699 |
+
"learning_rate": 0.00037398752727543967,
|
| 3700 |
+
"loss": 6.3897,
|
| 3701 |
+
"step": 524
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 0.167973124300112,
|
| 3705 |
+
"grad_norm": 5.483648777008057,
|
| 3706 |
+
"learning_rate": 0.0003738885998236952,
|
| 3707 |
+
"loss": 6.2346,
|
| 3708 |
+
"step": 525
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 0.16829307310830266,
|
| 3712 |
+
"grad_norm": 6.923319339752197,
|
| 3713 |
+
"learning_rate": 0.00037378949849229445,
|
| 3714 |
+
"loss": 6.2363,
|
| 3715 |
+
"step": 526
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.16861302191649336,
|
| 3719 |
+
"grad_norm": 5.014647960662842,
|
| 3720 |
+
"learning_rate": 0.0003736902233863801,
|
| 3721 |
+
"loss": 6.249,
|
| 3722 |
+
"step": 527
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 0.16893297072468405,
|
| 3726 |
+
"grad_norm": 8.451704978942871,
|
| 3727 |
+
"learning_rate": 0.00037359077461127945,
|
| 3728 |
+
"loss": 6.3163,
|
| 3729 |
+
"step": 528
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 0.16925291953287475,
|
| 3733 |
+
"grad_norm": 5.6020073890686035,
|
| 3734 |
+
"learning_rate": 0.00037349115227250373,
|
| 3735 |
+
"loss": 6.3071,
|
| 3736 |
+
"step": 529
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 0.16957286834106544,
|
| 3740 |
+
"grad_norm": 38.06607437133789,
|
| 3741 |
+
"learning_rate": 0.0003733913564757487,
|
| 3742 |
+
"loss": 6.1327,
|
| 3743 |
+
"step": 530
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 0.1698928171492561,
|
| 3747 |
+
"grad_norm": 8.072343826293945,
|
| 3748 |
+
"learning_rate": 0.00037329138732689385,
|
| 3749 |
+
"loss": 6.2813,
|
| 3750 |
+
"step": 531
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 0.1702127659574468,
|
| 3754 |
+
"grad_norm": 14.569067001342773,
|
| 3755 |
+
"learning_rate": 0.0003731912449320027,
|
| 3756 |
+
"loss": 6.14,
|
| 3757 |
+
"step": 532
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.1705327147656375,
|
| 3761 |
+
"grad_norm": 10.168042182922363,
|
| 3762 |
+
"learning_rate": 0.0003730909293973226,
|
| 3763 |
+
"loss": 6.3262,
|
| 3764 |
+
"step": 533
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 0.1708526635738282,
|
| 3768 |
+
"grad_norm": 10.32205581665039,
|
| 3769 |
+
"learning_rate": 0.00037299044082928455,
|
| 3770 |
+
"loss": 6.1903,
|
| 3771 |
+
"step": 534
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 0.1711726123820189,
|
| 3775 |
+
"grad_norm": 6.994221210479736,
|
| 3776 |
+
"learning_rate": 0.0003728897793345032,
|
| 3777 |
+
"loss": 6.2946,
|
| 3778 |
+
"step": 535
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 0.17149256119020956,
|
| 3782 |
+
"grad_norm": 6.205694675445557,
|
| 3783 |
+
"learning_rate": 0.0003727889450197765,
|
| 3784 |
+
"loss": 6.4397,
|
| 3785 |
+
"step": 536
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 0.17181250999840025,
|
| 3789 |
+
"grad_norm": 6.186336517333984,
|
| 3790 |
+
"learning_rate": 0.000372687937992086,
|
| 3791 |
+
"loss": 6.3172,
|
| 3792 |
+
"step": 537
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 0.17213245880659095,
|
| 3796 |
+
"grad_norm": 12.758987426757812,
|
| 3797 |
+
"learning_rate": 0.00037258675835859635,
|
| 3798 |
+
"loss": 6.3897,
|
| 3799 |
+
"step": 538
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 0.17245240761478164,
|
| 3803 |
+
"grad_norm": 11.200096130371094,
|
| 3804 |
+
"learning_rate": 0.0003724854062266553,
|
| 3805 |
+
"loss": 6.0816,
|
| 3806 |
+
"step": 539
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 0.1727723564229723,
|
| 3810 |
+
"grad_norm": 13.330513954162598,
|
| 3811 |
+
"learning_rate": 0.00037238388170379364,
|
| 3812 |
+
"loss": 6.5373,
|
| 3813 |
+
"step": 540
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 0.173092305231163,
|
| 3817 |
+
"grad_norm": 6.715458869934082,
|
| 3818 |
+
"learning_rate": 0.00037228218489772515,
|
| 3819 |
+
"loss": 6.1138,
|
| 3820 |
+
"step": 541
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 0.1734122540393537,
|
| 3824 |
+
"grad_norm": 8.349139213562012,
|
| 3825 |
+
"learning_rate": 0.0003721803159163463,
|
| 3826 |
+
"loss": 6.1373,
|
| 3827 |
+
"step": 542
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 0.1737322028475444,
|
| 3831 |
+
"grad_norm": 9.18417739868164,
|
| 3832 |
+
"learning_rate": 0.00037207827486773624,
|
| 3833 |
+
"loss": 6.2775,
|
| 3834 |
+
"step": 543
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.1740521516557351,
|
| 3838 |
+
"grad_norm": 8.832735061645508,
|
| 3839 |
+
"learning_rate": 0.0003719760618601567,
|
| 3840 |
+
"loss": 6.2763,
|
| 3841 |
+
"step": 544
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 0.17437210046392576,
|
| 3845 |
+
"grad_norm": 14.054030418395996,
|
| 3846 |
+
"learning_rate": 0.00037187367700205185,
|
| 3847 |
+
"loss": 6.6554,
|
| 3848 |
+
"step": 545
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 0.17469204927211646,
|
| 3852 |
+
"grad_norm": 8.51115894317627,
|
| 3853 |
+
"learning_rate": 0.0003717711204020482,
|
| 3854 |
+
"loss": 6.3151,
|
| 3855 |
+
"step": 546
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 0.17501199808030715,
|
| 3859 |
+
"grad_norm": 10.541391372680664,
|
| 3860 |
+
"learning_rate": 0.00037166839216895445,
|
| 3861 |
+
"loss": 6.1285,
|
| 3862 |
+
"step": 547
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 0.17533194688849785,
|
| 3866 |
+
"grad_norm": 21.057283401489258,
|
| 3867 |
+
"learning_rate": 0.0003715654924117614,
|
| 3868 |
+
"loss": 6.4187,
|
| 3869 |
+
"step": 548
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 0.17565189569668854,
|
| 3873 |
+
"grad_norm": 17.178653717041016,
|
| 3874 |
+
"learning_rate": 0.00037146242123964183,
|
| 3875 |
+
"loss": 6.2921,
|
| 3876 |
+
"step": 549
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.1759718445048792,
|
| 3880 |
+
"grad_norm": 10.637142181396484,
|
| 3881 |
+
"learning_rate": 0.00037135917876195037,
|
| 3882 |
+
"loss": 6.168,
|
| 3883 |
+
"step": 550
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 0.1762917933130699,
|
| 3887 |
+
"grad_norm": 7.821049213409424,
|
| 3888 |
+
"learning_rate": 0.0003712557650882234,
|
| 3889 |
+
"loss": 6.3907,
|
| 3890 |
+
"step": 551
|
| 3891 |
+
},
|
| 3892 |
+
{
|
| 3893 |
+
"epoch": 0.1766117421212606,
|
| 3894 |
+
"grad_norm": 9.258932113647461,
|
| 3895 |
+
"learning_rate": 0.000371152180328179,
|
| 3896 |
+
"loss": 6.2813,
|
| 3897 |
+
"step": 552
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 0.1769316909294513,
|
| 3901 |
+
"grad_norm": 8.969828605651855,
|
| 3902 |
+
"learning_rate": 0.0003710484245917167,
|
| 3903 |
+
"loss": 6.2085,
|
| 3904 |
+
"step": 553
|
| 3905 |
+
},
|
| 3906 |
+
{
|
| 3907 |
+
"epoch": 0.17725163973764196,
|
| 3908 |
+
"grad_norm": 10.377073287963867,
|
| 3909 |
+
"learning_rate": 0.0003709444979889174,
|
| 3910 |
+
"loss": 6.3274,
|
| 3911 |
+
"step": 554
|
| 3912 |
+
},
|
| 3913 |
+
{
|
| 3914 |
+
"epoch": 0.17757158854583266,
|
| 3915 |
+
"grad_norm": 13.76203727722168,
|
| 3916 |
+
"learning_rate": 0.0003708404006300434,
|
| 3917 |
+
"loss": 6.7071,
|
| 3918 |
+
"step": 555
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 0.17789153735402335,
|
| 3922 |
+
"grad_norm": 9.469388961791992,
|
| 3923 |
+
"learning_rate": 0.00037073613262553803,
|
| 3924 |
+
"loss": 6.1904,
|
| 3925 |
+
"step": 556
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 0.17821148616221405,
|
| 3929 |
+
"grad_norm": 7.818362712860107,
|
| 3930 |
+
"learning_rate": 0.00037063169408602586,
|
| 3931 |
+
"loss": 6.3993,
|
| 3932 |
+
"step": 557
|
| 3933 |
+
},
|
| 3934 |
+
{
|
| 3935 |
+
"epoch": 0.17853143497040475,
|
| 3936 |
+
"grad_norm": 10.85080337524414,
|
| 3937 |
+
"learning_rate": 0.000370527085122312,
|
| 3938 |
+
"loss": 6.2407,
|
| 3939 |
+
"step": 558
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 0.1788513837785954,
|
| 3943 |
+
"grad_norm": 7.716219902038574,
|
| 3944 |
+
"learning_rate": 0.0003704223058453829,
|
| 3945 |
+
"loss": 6.3401,
|
| 3946 |
+
"step": 559
|
| 3947 |
+
},
|
| 3948 |
+
{
|
| 3949 |
+
"epoch": 0.1791713325867861,
|
| 3950 |
+
"grad_norm": 4.769658088684082,
|
| 3951 |
+
"learning_rate": 0.00037031735636640524,
|
| 3952 |
+
"loss": 6.3418,
|
| 3953 |
+
"step": 560
|
| 3954 |
+
},
|
| 3955 |
+
{
|
| 3956 |
+
"epoch": 0.1794912813949768,
|
| 3957 |
+
"grad_norm": 7.051829814910889,
|
| 3958 |
+
"learning_rate": 0.00037021223679672673,
|
| 3959 |
+
"loss": 6.2469,
|
| 3960 |
+
"step": 561
|
| 3961 |
+
},
|
| 3962 |
+
{
|
| 3963 |
+
"epoch": 0.1798112302031675,
|
| 3964 |
+
"grad_norm": 9.345251083374023,
|
| 3965 |
+
"learning_rate": 0.000370106947247875,
|
| 3966 |
+
"loss": 6.5334,
|
| 3967 |
+
"step": 562
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 0.1801311790113582,
|
| 3971 |
+
"grad_norm": 8.516511917114258,
|
| 3972 |
+
"learning_rate": 0.0003700014878315585,
|
| 3973 |
+
"loss": 6.2604,
|
| 3974 |
+
"step": 563
|
| 3975 |
+
},
|
| 3976 |
+
{
|
| 3977 |
+
"epoch": 0.18045112781954886,
|
| 3978 |
+
"grad_norm": 8.440566062927246,
|
| 3979 |
+
"learning_rate": 0.0003698958586596657,
|
| 3980 |
+
"loss": 6.2302,
|
| 3981 |
+
"step": 564
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 0.18077107662773956,
|
| 3985 |
+
"grad_norm": 5.112747669219971,
|
| 3986 |
+
"learning_rate": 0.0003697900598442651,
|
| 3987 |
+
"loss": 6.3096,
|
| 3988 |
+
"step": 565
|
| 3989 |
+
},
|
| 3990 |
+
{
|
| 3991 |
+
"epoch": 0.18109102543593025,
|
| 3992 |
+
"grad_norm": 5.404264450073242,
|
| 3993 |
+
"learning_rate": 0.00036968409149760534,
|
| 3994 |
+
"loss": 6.3788,
|
| 3995 |
+
"step": 566
|
| 3996 |
+
},
|
| 3997 |
+
{
|
| 3998 |
+
"epoch": 0.18141097424412095,
|
| 3999 |
+
"grad_norm": 8.593727111816406,
|
| 4000 |
+
"learning_rate": 0.0003695779537321149,
|
| 4001 |
+
"loss": 6.0832,
|
| 4002 |
+
"step": 567
|
| 4003 |
+
},
|
| 4004 |
+
{
|
| 4005 |
+
"epoch": 0.18173092305231164,
|
| 4006 |
+
"grad_norm": 5.876023769378662,
|
| 4007 |
+
"learning_rate": 0.00036947164666040184,
|
| 4008 |
+
"loss": 5.996,
|
| 4009 |
+
"step": 568
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 0.1820508718605023,
|
| 4013 |
+
"grad_norm": 5.251745700836182,
|
| 4014 |
+
"learning_rate": 0.0003693651703952541,
|
| 4015 |
+
"loss": 6.1838,
|
| 4016 |
+
"step": 569
|
| 4017 |
+
},
|
| 4018 |
+
{
|
| 4019 |
+
"epoch": 0.182370820668693,
|
| 4020 |
+
"grad_norm": 6.7820539474487305,
|
| 4021 |
+
"learning_rate": 0.00036925852504963893,
|
| 4022 |
+
"loss": 6.2634,
|
| 4023 |
+
"step": 570
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 0.1826907694768837,
|
| 4027 |
+
"grad_norm": 6.269637584686279,
|
| 4028 |
+
"learning_rate": 0.00036915171073670307,
|
| 4029 |
+
"loss": 6.4802,
|
| 4030 |
+
"step": 571
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 0.1830107182850744,
|
| 4034 |
+
"grad_norm": 23.37257957458496,
|
| 4035 |
+
"learning_rate": 0.0003690447275697725,
|
| 4036 |
+
"loss": 6.1899,
|
| 4037 |
+
"step": 572
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 0.18333066709326507,
|
| 4041 |
+
"grad_norm": 5.05233097076416,
|
| 4042 |
+
"learning_rate": 0.00036893757566235227,
|
| 4043 |
+
"loss": 6.1816,
|
| 4044 |
+
"step": 573
|
| 4045 |
+
},
|
| 4046 |
+
{
|
| 4047 |
+
"epoch": 0.18365061590145576,
|
| 4048 |
+
"grad_norm": 8.209317207336426,
|
| 4049 |
+
"learning_rate": 0.00036883025512812675,
|
| 4050 |
+
"loss": 6.4029,
|
| 4051 |
+
"step": 574
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 0.18397056470964646,
|
| 4055 |
+
"grad_norm": 10.015338897705078,
|
| 4056 |
+
"learning_rate": 0.00036872276608095874,
|
| 4057 |
+
"loss": 6.1534,
|
| 4058 |
+
"step": 575
|
| 4059 |
+
},
|
| 4060 |
+
{
|
| 4061 |
+
"epoch": 0.18429051351783715,
|
| 4062 |
+
"grad_norm": 6.191135883331299,
|
| 4063 |
+
"learning_rate": 0.00036861510863489023,
|
| 4064 |
+
"loss": 6.3621,
|
| 4065 |
+
"step": 576
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 0.18461046232602785,
|
| 4069 |
+
"grad_norm": 9.056325912475586,
|
| 4070 |
+
"learning_rate": 0.0003685072829041417,
|
| 4071 |
+
"loss": 6.3678,
|
| 4072 |
+
"step": 577
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 0.18493041113421851,
|
| 4076 |
+
"grad_norm": 11.877366065979004,
|
| 4077 |
+
"learning_rate": 0.00036839928900311223,
|
| 4078 |
+
"loss": 6.2934,
|
| 4079 |
+
"step": 578
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 0.1852503599424092,
|
| 4083 |
+
"grad_norm": 11.008338928222656,
|
| 4084 |
+
"learning_rate": 0.0003682911270463793,
|
| 4085 |
+
"loss": 6.2898,
|
| 4086 |
+
"step": 579
|
| 4087 |
+
},
|
| 4088 |
+
{
|
| 4089 |
+
"epoch": 0.1855703087505999,
|
| 4090 |
+
"grad_norm": 8.387709617614746,
|
| 4091 |
+
"learning_rate": 0.0003681827971486986,
|
| 4092 |
+
"loss": 6.3564,
|
| 4093 |
+
"step": 580
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 0.1858902575587906,
|
| 4097 |
+
"grad_norm": 5.683431148529053,
|
| 4098 |
+
"learning_rate": 0.0003680742994250042,
|
| 4099 |
+
"loss": 6.358,
|
| 4100 |
+
"step": 581
|
| 4101 |
+
},
|
| 4102 |
+
{
|
| 4103 |
+
"epoch": 0.1862102063669813,
|
| 4104 |
+
"grad_norm": 9.834399223327637,
|
| 4105 |
+
"learning_rate": 0.000367965633990408,
|
| 4106 |
+
"loss": 6.4083,
|
| 4107 |
+
"step": 582
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 0.18653015517517196,
|
| 4111 |
+
"grad_norm": 6.106937408447266,
|
| 4112 |
+
"learning_rate": 0.0003678568009602001,
|
| 4113 |
+
"loss": 6.1561,
|
| 4114 |
+
"step": 583
|
| 4115 |
+
},
|
| 4116 |
+
{
|
| 4117 |
+
"epoch": 0.18685010398336266,
|
| 4118 |
+
"grad_norm": 10.490397453308105,
|
| 4119 |
+
"learning_rate": 0.0003677478004498481,
|
| 4120 |
+
"loss": 6.326,
|
| 4121 |
+
"step": 584
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 0.18717005279155335,
|
| 4125 |
+
"grad_norm": 7.496698379516602,
|
| 4126 |
+
"learning_rate": 0.0003676386325749976,
|
| 4127 |
+
"loss": 6.3687,
|
| 4128 |
+
"step": 585
|
| 4129 |
+
},
|
| 4130 |
+
{
|
| 4131 |
+
"epoch": 0.18749000159974405,
|
| 4132 |
+
"grad_norm": 7.097468376159668,
|
| 4133 |
+
"learning_rate": 0.0003675292974514715,
|
| 4134 |
+
"loss": 6.2169,
|
| 4135 |
+
"step": 586
|
| 4136 |
+
},
|
| 4137 |
+
{
|
| 4138 |
+
"epoch": 0.18780995040793472,
|
| 4139 |
+
"grad_norm": 14.11148738861084,
|
| 4140 |
+
"learning_rate": 0.0003674197951952704,
|
| 4141 |
+
"loss": 6.1241,
|
| 4142 |
+
"step": 587
|
| 4143 |
+
},
|
| 4144 |
+
{
|
| 4145 |
+
"epoch": 0.1881298992161254,
|
| 4146 |
+
"grad_norm": 13.260645866394043,
|
| 4147 |
+
"learning_rate": 0.00036731012592257194,
|
| 4148 |
+
"loss": 6.1671,
|
| 4149 |
+
"step": 588
|
| 4150 |
+
},
|
| 4151 |
+
{
|
| 4152 |
+
"epoch": 0.1884498480243161,
|
| 4153 |
+
"grad_norm": 8.077284812927246,
|
| 4154 |
+
"learning_rate": 0.0003672002897497312,
|
| 4155 |
+
"loss": 6.2584,
|
| 4156 |
+
"step": 589
|
| 4157 |
+
},
|
| 4158 |
+
{
|
| 4159 |
+
"epoch": 0.1887697968325068,
|
| 4160 |
+
"grad_norm": 5.156368255615234,
|
| 4161 |
+
"learning_rate": 0.00036709028679328013,
|
| 4162 |
+
"loss": 6.2548,
|
| 4163 |
+
"step": 590
|
| 4164 |
+
},
|
| 4165 |
+
{
|
| 4166 |
+
"epoch": 0.1890897456406975,
|
| 4167 |
+
"grad_norm": 7.575767517089844,
|
| 4168 |
+
"learning_rate": 0.0003669801171699279,
|
| 4169 |
+
"loss": 6.426,
|
| 4170 |
+
"step": 591
|
| 4171 |
+
},
|
| 4172 |
+
{
|
| 4173 |
+
"epoch": 0.18940969444888817,
|
| 4174 |
+
"grad_norm": 8.419235229492188,
|
| 4175 |
+
"learning_rate": 0.0003668697809965602,
|
| 4176 |
+
"loss": 6.2028,
|
| 4177 |
+
"step": 592
|
| 4178 |
+
},
|
| 4179 |
+
{
|
| 4180 |
+
"epoch": 0.18972964325707886,
|
| 4181 |
+
"grad_norm": 6.528421401977539,
|
| 4182 |
+
"learning_rate": 0.0003667592783902397,
|
| 4183 |
+
"loss": 6.2471,
|
| 4184 |
+
"step": 593
|
| 4185 |
+
},
|
| 4186 |
+
{
|
| 4187 |
+
"epoch": 0.19004959206526956,
|
| 4188 |
+
"grad_norm": 7.039434909820557,
|
| 4189 |
+
"learning_rate": 0.00036664860946820536,
|
| 4190 |
+
"loss": 6.4782,
|
| 4191 |
+
"step": 594
|
| 4192 |
+
},
|
| 4193 |
+
{
|
| 4194 |
+
"epoch": 0.19036954087346025,
|
| 4195 |
+
"grad_norm": 5.257431507110596,
|
| 4196 |
+
"learning_rate": 0.00036653777434787294,
|
| 4197 |
+
"loss": 6.1964,
|
| 4198 |
+
"step": 595
|
| 4199 |
+
},
|
| 4200 |
+
{
|
| 4201 |
+
"epoch": 0.19068948968165095,
|
| 4202 |
+
"grad_norm": 7.877770900726318,
|
| 4203 |
+
"learning_rate": 0.0003664267731468342,
|
| 4204 |
+
"loss": 6.501,
|
| 4205 |
+
"step": 596
|
| 4206 |
+
},
|
| 4207 |
+
{
|
| 4208 |
+
"epoch": 0.19100943848984162,
|
| 4209 |
+
"grad_norm": 6.9174933433532715,
|
| 4210 |
+
"learning_rate": 0.00036631560598285735,
|
| 4211 |
+
"loss": 6.1698,
|
| 4212 |
+
"step": 597
|
| 4213 |
+
},
|
| 4214 |
+
{
|
| 4215 |
+
"epoch": 0.1913293872980323,
|
| 4216 |
+
"grad_norm": 5.704330921173096,
|
| 4217 |
+
"learning_rate": 0.00036620427297388646,
|
| 4218 |
+
"loss": 6.1756,
|
| 4219 |
+
"step": 598
|
| 4220 |
+
},
|
| 4221 |
+
{
|
| 4222 |
+
"epoch": 0.191649336106223,
|
| 4223 |
+
"grad_norm": 7.497727870941162,
|
| 4224 |
+
"learning_rate": 0.00036609277423804183,
|
| 4225 |
+
"loss": 6.1859,
|
| 4226 |
+
"step": 599
|
| 4227 |
+
},
|
| 4228 |
+
{
|
| 4229 |
+
"epoch": 0.1919692849144137,
|
| 4230 |
+
"grad_norm": 10.520272254943848,
|
| 4231 |
+
"learning_rate": 0.0003659811098936193,
|
| 4232 |
+
"loss": 6.2496,
|
| 4233 |
+
"step": 600
|
| 4234 |
+
},
|
| 4235 |
+
{
|
| 4236 |
+
"epoch": 0.1919692849144137,
|
| 4237 |
+
"eval_loss": 3.1351470947265625,
|
| 4238 |
+
"eval_runtime": 233.6712,
|
| 4239 |
+
"eval_samples_per_second": 5.632,
|
| 4240 |
+
"eval_steps_per_second": 1.408,
|
| 4241 |
+
"step": 600
|
| 4242 |
}
|
| 4243 |
],
|
| 4244 |
"logging_steps": 1,
|
|
|
|
| 4267 |
"attributes": {}
|
| 4268 |
}
|
| 4269 |
},
|
| 4270 |
+
"total_flos": 5.573737245006889e+17,
|
| 4271 |
"train_batch_size": 4,
|
| 4272 |
"trial_name": null,
|
| 4273 |
"trial_params": null
|