Training in progress, step 1600, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 891644712
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77e67eae521c3cba24d0ae9eb0d44447bec89794066751cc7d707f2fdc5c29bf
|
| 3 |
size 891644712
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1783444794
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:42e0c34297a9f3d72c5022f0b3dc2e519698ad0189be335ae6d8648524d74bb2
|
| 3 |
size 1783444794
|
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 sha256:8dee4a2b51470c2e565b08aae8a4e5156e5c34e8c236adf6153eeb283fa560ec
|
| 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 sha256:1affb368014b9c4895e09d750783801a20ec7c8f622ab5a02e1bce055904fb15
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -4487,6 +4487,1126 @@
|
|
| 4487 |
"learning_rate": 0.00014958448930255265,
|
| 4488 |
"loss": 0.5353,
|
| 4489 |
"step": 1280
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4490 |
}
|
| 4491 |
],
|
| 4492 |
"logging_steps": 2,
|
|
@@ -4506,7 +5626,7 @@
|
|
| 4506 |
"attributes": {}
|
| 4507 |
}
|
| 4508 |
},
|
| 4509 |
-
"total_flos":
|
| 4510 |
"train_batch_size": 8,
|
| 4511 |
"trial_name": null,
|
| 4512 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8874098724348308,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 4487 |
"learning_rate": 0.00014958448930255265,
|
| 4488 |
"loss": 0.5353,
|
| 4489 |
"step": 1280
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 0.7110371602884082,
|
| 4493 |
+
"grad_norm": 0.29970696568489075,
|
| 4494 |
+
"learning_rate": 0.00014942837881378465,
|
| 4495 |
+
"loss": 0.4261,
|
| 4496 |
+
"step": 1282
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 0.7121464226289518,
|
| 4500 |
+
"grad_norm": 0.36919617652893066,
|
| 4501 |
+
"learning_rate": 0.00014927210875741347,
|
| 4502 |
+
"loss": 0.4935,
|
| 4503 |
+
"step": 1284
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.7132556849694953,
|
| 4507 |
+
"grad_norm": 0.3045351505279541,
|
| 4508 |
+
"learning_rate": 0.00014911567963791928,
|
| 4509 |
+
"loss": 0.4191,
|
| 4510 |
+
"step": 1286
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 0.7143649473100389,
|
| 4514 |
+
"grad_norm": 0.30813243985176086,
|
| 4515 |
+
"learning_rate": 0.00014895909196029585,
|
| 4516 |
+
"loss": 0.3992,
|
| 4517 |
+
"step": 1288
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 0.7154742096505824,
|
| 4521 |
+
"grad_norm": 0.3444893956184387,
|
| 4522 |
+
"learning_rate": 0.00014880234623004866,
|
| 4523 |
+
"loss": 0.4351,
|
| 4524 |
+
"step": 1290
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 0.7165834719911259,
|
| 4528 |
+
"grad_norm": 0.29215720295906067,
|
| 4529 |
+
"learning_rate": 0.00014864544295319356,
|
| 4530 |
+
"loss": 0.4323,
|
| 4531 |
+
"step": 1292
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 0.7176927343316695,
|
| 4535 |
+
"grad_norm": 0.37148571014404297,
|
| 4536 |
+
"learning_rate": 0.00014848838263625496,
|
| 4537 |
+
"loss": 0.4463,
|
| 4538 |
+
"step": 1294
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 0.718801996672213,
|
| 4542 |
+
"grad_norm": 0.38779592514038086,
|
| 4543 |
+
"learning_rate": 0.00014833116578626417,
|
| 4544 |
+
"loss": 0.5293,
|
| 4545 |
+
"step": 1296
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 0.7199112590127565,
|
| 4549 |
+
"grad_norm": 0.31231850385665894,
|
| 4550 |
+
"learning_rate": 0.00014817379291075792,
|
| 4551 |
+
"loss": 0.4575,
|
| 4552 |
+
"step": 1298
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 0.7210205213533001,
|
| 4556 |
+
"grad_norm": 0.2966194450855255,
|
| 4557 |
+
"learning_rate": 0.00014801626451777658,
|
| 4558 |
+
"loss": 0.4337,
|
| 4559 |
+
"step": 1300
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 0.7221297836938436,
|
| 4563 |
+
"grad_norm": 0.2993515133857727,
|
| 4564 |
+
"learning_rate": 0.00014785858111586258,
|
| 4565 |
+
"loss": 0.3994,
|
| 4566 |
+
"step": 1302
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 0.7232390460343872,
|
| 4570 |
+
"grad_norm": 0.27827030420303345,
|
| 4571 |
+
"learning_rate": 0.00014770074321405878,
|
| 4572 |
+
"loss": 0.4079,
|
| 4573 |
+
"step": 1304
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 0.7243483083749307,
|
| 4577 |
+
"grad_norm": 0.2576323449611664,
|
| 4578 |
+
"learning_rate": 0.00014754275132190678,
|
| 4579 |
+
"loss": 0.3798,
|
| 4580 |
+
"step": 1306
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 0.7254575707154742,
|
| 4584 |
+
"grad_norm": 0.24843810498714447,
|
| 4585 |
+
"learning_rate": 0.0001473846059494453,
|
| 4586 |
+
"loss": 0.4564,
|
| 4587 |
+
"step": 1308
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 0.7265668330560178,
|
| 4591 |
+
"grad_norm": 0.2799087166786194,
|
| 4592 |
+
"learning_rate": 0.00014722630760720856,
|
| 4593 |
+
"loss": 0.3906,
|
| 4594 |
+
"step": 1310
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 0.7276760953965613,
|
| 4598 |
+
"grad_norm": 0.3635210394859314,
|
| 4599 |
+
"learning_rate": 0.00014706785680622462,
|
| 4600 |
+
"loss": 0.4409,
|
| 4601 |
+
"step": 1312
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 0.7287853577371048,
|
| 4605 |
+
"grad_norm": 0.30838248133659363,
|
| 4606 |
+
"learning_rate": 0.0001469092540580136,
|
| 4607 |
+
"loss": 0.4268,
|
| 4608 |
+
"step": 1314
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 0.7298946200776484,
|
| 4612 |
+
"grad_norm": 0.2162405252456665,
|
| 4613 |
+
"learning_rate": 0.0001467504998745863,
|
| 4614 |
+
"loss": 0.4085,
|
| 4615 |
+
"step": 1316
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 0.7310038824181919,
|
| 4619 |
+
"grad_norm": 0.30635640025138855,
|
| 4620 |
+
"learning_rate": 0.00014659159476844232,
|
| 4621 |
+
"loss": 0.4089,
|
| 4622 |
+
"step": 1318
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 0.7321131447587355,
|
| 4626 |
+
"grad_norm": 0.4048236012458801,
|
| 4627 |
+
"learning_rate": 0.00014643253925256846,
|
| 4628 |
+
"loss": 0.5283,
|
| 4629 |
+
"step": 1320
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 0.733222407099279,
|
| 4633 |
+
"grad_norm": 0.4575350284576416,
|
| 4634 |
+
"learning_rate": 0.00014627333384043713,
|
| 4635 |
+
"loss": 0.4675,
|
| 4636 |
+
"step": 1322
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 0.7343316694398225,
|
| 4640 |
+
"grad_norm": 0.3289128243923187,
|
| 4641 |
+
"learning_rate": 0.00014611397904600458,
|
| 4642 |
+
"loss": 0.4777,
|
| 4643 |
+
"step": 1324
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 0.7354409317803661,
|
| 4647 |
+
"grad_norm": 0.29231977462768555,
|
| 4648 |
+
"learning_rate": 0.00014595447538370935,
|
| 4649 |
+
"loss": 0.3851,
|
| 4650 |
+
"step": 1326
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 0.7365501941209096,
|
| 4654 |
+
"grad_norm": 0.33795663714408875,
|
| 4655 |
+
"learning_rate": 0.00014579482336847058,
|
| 4656 |
+
"loss": 0.475,
|
| 4657 |
+
"step": 1328
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 0.7376594564614531,
|
| 4661 |
+
"grad_norm": 0.20978425443172455,
|
| 4662 |
+
"learning_rate": 0.00014563502351568625,
|
| 4663 |
+
"loss": 0.3455,
|
| 4664 |
+
"step": 1330
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 0.7387687188019967,
|
| 4668 |
+
"grad_norm": 0.4113425016403198,
|
| 4669 |
+
"learning_rate": 0.00014547507634123176,
|
| 4670 |
+
"loss": 0.3741,
|
| 4671 |
+
"step": 1332
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 0.7398779811425402,
|
| 4675 |
+
"grad_norm": 0.38009563088417053,
|
| 4676 |
+
"learning_rate": 0.0001453149823614579,
|
| 4677 |
+
"loss": 0.53,
|
| 4678 |
+
"step": 1334
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 0.7409872434830838,
|
| 4682 |
+
"grad_norm": 0.2306888997554779,
|
| 4683 |
+
"learning_rate": 0.00014515474209318948,
|
| 4684 |
+
"loss": 0.4229,
|
| 4685 |
+
"step": 1336
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 0.7420965058236273,
|
| 4689 |
+
"grad_norm": 0.326107382774353,
|
| 4690 |
+
"learning_rate": 0.00014499435605372366,
|
| 4691 |
+
"loss": 0.3731,
|
| 4692 |
+
"step": 1338
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 0.7432057681641708,
|
| 4696 |
+
"grad_norm": 0.26597580313682556,
|
| 4697 |
+
"learning_rate": 0.00014483382476082802,
|
| 4698 |
+
"loss": 0.3841,
|
| 4699 |
+
"step": 1340
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 0.7443150305047144,
|
| 4703 |
+
"grad_norm": 0.3690161406993866,
|
| 4704 |
+
"learning_rate": 0.00014467314873273918,
|
| 4705 |
+
"loss": 0.4814,
|
| 4706 |
+
"step": 1342
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 0.7454242928452579,
|
| 4710 |
+
"grad_norm": 0.30208808183670044,
|
| 4711 |
+
"learning_rate": 0.0001445123284881609,
|
| 4712 |
+
"loss": 0.4522,
|
| 4713 |
+
"step": 1344
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 0.7465335551858014,
|
| 4717 |
+
"grad_norm": 0.28505003452301025,
|
| 4718 |
+
"learning_rate": 0.00014435136454626264,
|
| 4719 |
+
"loss": 0.4283,
|
| 4720 |
+
"step": 1346
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 0.747642817526345,
|
| 4724 |
+
"grad_norm": 0.2772189974784851,
|
| 4725 |
+
"learning_rate": 0.0001441902574266776,
|
| 4726 |
+
"loss": 0.3964,
|
| 4727 |
+
"step": 1348
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 0.7487520798668885,
|
| 4731 |
+
"grad_norm": 0.4741387665271759,
|
| 4732 |
+
"learning_rate": 0.0001440290076495013,
|
| 4733 |
+
"loss": 0.4577,
|
| 4734 |
+
"step": 1350
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 0.7498613422074321,
|
| 4738 |
+
"grad_norm": 0.3234037160873413,
|
| 4739 |
+
"learning_rate": 0.00014386761573528976,
|
| 4740 |
+
"loss": 0.4542,
|
| 4741 |
+
"step": 1352
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 0.7509706045479756,
|
| 4745 |
+
"grad_norm": 0.3096826374530792,
|
| 4746 |
+
"learning_rate": 0.0001437060822050579,
|
| 4747 |
+
"loss": 0.3892,
|
| 4748 |
+
"step": 1354
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 0.7520798668885191,
|
| 4752 |
+
"grad_norm": 0.2626829743385315,
|
| 4753 |
+
"learning_rate": 0.00014354440758027772,
|
| 4754 |
+
"loss": 0.4441,
|
| 4755 |
+
"step": 1356
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 0.7531891292290627,
|
| 4759 |
+
"grad_norm": 0.2835614085197449,
|
| 4760 |
+
"learning_rate": 0.00014338259238287678,
|
| 4761 |
+
"loss": 0.4905,
|
| 4762 |
+
"step": 1358
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 0.7542983915696062,
|
| 4766 |
+
"grad_norm": 0.23285934329032898,
|
| 4767 |
+
"learning_rate": 0.00014322063713523647,
|
| 4768 |
+
"loss": 0.3911,
|
| 4769 |
+
"step": 1360
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 0.7554076539101497,
|
| 4773 |
+
"grad_norm": 0.2746225595474243,
|
| 4774 |
+
"learning_rate": 0.00014305854236019018,
|
| 4775 |
+
"loss": 0.4681,
|
| 4776 |
+
"step": 1362
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 0.7565169162506933,
|
| 4780 |
+
"grad_norm": 0.30551669001579285,
|
| 4781 |
+
"learning_rate": 0.0001428963085810219,
|
| 4782 |
+
"loss": 0.3883,
|
| 4783 |
+
"step": 1364
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 0.7576261785912368,
|
| 4787 |
+
"grad_norm": 0.3372795283794403,
|
| 4788 |
+
"learning_rate": 0.0001427339363214642,
|
| 4789 |
+
"loss": 0.4452,
|
| 4790 |
+
"step": 1366
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 0.7587354409317804,
|
| 4794 |
+
"grad_norm": 0.2697629928588867,
|
| 4795 |
+
"learning_rate": 0.00014257142610569682,
|
| 4796 |
+
"loss": 0.3853,
|
| 4797 |
+
"step": 1368
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 0.7598447032723239,
|
| 4801 |
+
"grad_norm": 0.29290249943733215,
|
| 4802 |
+
"learning_rate": 0.00014240877845834472,
|
| 4803 |
+
"loss": 0.4694,
|
| 4804 |
+
"step": 1370
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 0.7609539656128674,
|
| 4808 |
+
"grad_norm": 0.298874169588089,
|
| 4809 |
+
"learning_rate": 0.00014224599390447672,
|
| 4810 |
+
"loss": 0.4644,
|
| 4811 |
+
"step": 1372
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 0.762063227953411,
|
| 4815 |
+
"grad_norm": 0.23659248650074005,
|
| 4816 |
+
"learning_rate": 0.00014208307296960344,
|
| 4817 |
+
"loss": 0.395,
|
| 4818 |
+
"step": 1374
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 0.7631724902939545,
|
| 4822 |
+
"grad_norm": 0.2965550124645233,
|
| 4823 |
+
"learning_rate": 0.00014192001617967587,
|
| 4824 |
+
"loss": 0.5332,
|
| 4825 |
+
"step": 1376
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 0.764281752634498,
|
| 4829 |
+
"grad_norm": 0.3338853716850281,
|
| 4830 |
+
"learning_rate": 0.00014175682406108352,
|
| 4831 |
+
"loss": 0.5176,
|
| 4832 |
+
"step": 1378
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 0.7653910149750416,
|
| 4836 |
+
"grad_norm": 0.24134789407253265,
|
| 4837 |
+
"learning_rate": 0.0001415934971406528,
|
| 4838 |
+
"loss": 0.4224,
|
| 4839 |
+
"step": 1380
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 0.7665002773155851,
|
| 4843 |
+
"grad_norm": 0.3920575678348541,
|
| 4844 |
+
"learning_rate": 0.00014143003594564528,
|
| 4845 |
+
"loss": 0.4627,
|
| 4846 |
+
"step": 1382
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 0.7676095396561287,
|
| 4850 |
+
"grad_norm": 0.3521714508533478,
|
| 4851 |
+
"learning_rate": 0.00014126644100375603,
|
| 4852 |
+
"loss": 0.446,
|
| 4853 |
+
"step": 1384
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 0.7687188019966722,
|
| 4857 |
+
"grad_norm": 0.2819899022579193,
|
| 4858 |
+
"learning_rate": 0.0001411027128431119,
|
| 4859 |
+
"loss": 0.3637,
|
| 4860 |
+
"step": 1386
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 0.7698280643372157,
|
| 4864 |
+
"grad_norm": 0.1896730363368988,
|
| 4865 |
+
"learning_rate": 0.00014093885199226972,
|
| 4866 |
+
"loss": 0.3206,
|
| 4867 |
+
"step": 1388
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 0.7709373266777593,
|
| 4871 |
+
"grad_norm": 0.21066512167453766,
|
| 4872 |
+
"learning_rate": 0.0001407748589802148,
|
| 4873 |
+
"loss": 0.3081,
|
| 4874 |
+
"step": 1390
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 0.7720465890183028,
|
| 4878 |
+
"grad_norm": 0.2417469471693039,
|
| 4879 |
+
"learning_rate": 0.000140610734336359,
|
| 4880 |
+
"loss": 0.3954,
|
| 4881 |
+
"step": 1392
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 0.7731558513588463,
|
| 4885 |
+
"grad_norm": 0.41810843348503113,
|
| 4886 |
+
"learning_rate": 0.00014044647859053915,
|
| 4887 |
+
"loss": 0.521,
|
| 4888 |
+
"step": 1394
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 0.7742651136993899,
|
| 4892 |
+
"grad_norm": 0.21894732117652893,
|
| 4893 |
+
"learning_rate": 0.00014028209227301533,
|
| 4894 |
+
"loss": 0.342,
|
| 4895 |
+
"step": 1396
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 0.7753743760399334,
|
| 4899 |
+
"grad_norm": 0.31191563606262207,
|
| 4900 |
+
"learning_rate": 0.00014011757591446918,
|
| 4901 |
+
"loss": 0.4173,
|
| 4902 |
+
"step": 1398
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 0.776483638380477,
|
| 4906 |
+
"grad_norm": 0.34966176748275757,
|
| 4907 |
+
"learning_rate": 0.00013995293004600206,
|
| 4908 |
+
"loss": 0.4993,
|
| 4909 |
+
"step": 1400
|
| 4910 |
+
},
|
| 4911 |
+
{
|
| 4912 |
+
"epoch": 0.7775929007210205,
|
| 4913 |
+
"grad_norm": 0.3630419671535492,
|
| 4914 |
+
"learning_rate": 0.00013978815519913345,
|
| 4915 |
+
"loss": 0.4331,
|
| 4916 |
+
"step": 1402
|
| 4917 |
+
},
|
| 4918 |
+
{
|
| 4919 |
+
"epoch": 0.778702163061564,
|
| 4920 |
+
"grad_norm": 0.2934836447238922,
|
| 4921 |
+
"learning_rate": 0.00013962325190579919,
|
| 4922 |
+
"loss": 0.4618,
|
| 4923 |
+
"step": 1404
|
| 4924 |
+
},
|
| 4925 |
+
{
|
| 4926 |
+
"epoch": 0.7798114254021076,
|
| 4927 |
+
"grad_norm": 0.330842524766922,
|
| 4928 |
+
"learning_rate": 0.00013945822069834983,
|
| 4929 |
+
"loss": 0.4437,
|
| 4930 |
+
"step": 1406
|
| 4931 |
+
},
|
| 4932 |
+
{
|
| 4933 |
+
"epoch": 0.7809206877426511,
|
| 4934 |
+
"grad_norm": 0.24101948738098145,
|
| 4935 |
+
"learning_rate": 0.0001392930621095489,
|
| 4936 |
+
"loss": 0.4065,
|
| 4937 |
+
"step": 1408
|
| 4938 |
+
},
|
| 4939 |
+
{
|
| 4940 |
+
"epoch": 0.7820299500831946,
|
| 4941 |
+
"grad_norm": 0.5105953216552734,
|
| 4942 |
+
"learning_rate": 0.00013912777667257094,
|
| 4943 |
+
"loss": 0.5135,
|
| 4944 |
+
"step": 1410
|
| 4945 |
+
},
|
| 4946 |
+
{
|
| 4947 |
+
"epoch": 0.7831392124237382,
|
| 4948 |
+
"grad_norm": 0.2571849226951599,
|
| 4949 |
+
"learning_rate": 0.00013896236492100025,
|
| 4950 |
+
"loss": 0.4153,
|
| 4951 |
+
"step": 1412
|
| 4952 |
+
},
|
| 4953 |
+
{
|
| 4954 |
+
"epoch": 0.7842484747642817,
|
| 4955 |
+
"grad_norm": 0.27226725220680237,
|
| 4956 |
+
"learning_rate": 0.00013879682738882873,
|
| 4957 |
+
"loss": 0.3522,
|
| 4958 |
+
"step": 1414
|
| 4959 |
+
},
|
| 4960 |
+
{
|
| 4961 |
+
"epoch": 0.7853577371048253,
|
| 4962 |
+
"grad_norm": 0.2498832792043686,
|
| 4963 |
+
"learning_rate": 0.0001386311646104544,
|
| 4964 |
+
"loss": 0.4186,
|
| 4965 |
+
"step": 1416
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.7864669994453688,
|
| 4969 |
+
"grad_norm": 0.18799692392349243,
|
| 4970 |
+
"learning_rate": 0.00013846537712067962,
|
| 4971 |
+
"loss": 0.3278,
|
| 4972 |
+
"step": 1418
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.7875762617859123,
|
| 4976 |
+
"grad_norm": 0.3741225600242615,
|
| 4977 |
+
"learning_rate": 0.0001382994654547093,
|
| 4978 |
+
"loss": 0.4816,
|
| 4979 |
+
"step": 1420
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.7886855241264559,
|
| 4983 |
+
"grad_norm": 0.34847941994667053,
|
| 4984 |
+
"learning_rate": 0.00013813343014814925,
|
| 4985 |
+
"loss": 0.5673,
|
| 4986 |
+
"step": 1422
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.7897947864669994,
|
| 4990 |
+
"grad_norm": 0.38350898027420044,
|
| 4991 |
+
"learning_rate": 0.00013796727173700444,
|
| 4992 |
+
"loss": 0.5009,
|
| 4993 |
+
"step": 1424
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.790904048807543,
|
| 4997 |
+
"grad_norm": 0.2870427668094635,
|
| 4998 |
+
"learning_rate": 0.0001378009907576772,
|
| 4999 |
+
"loss": 0.4333,
|
| 5000 |
+
"step": 1426
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.7920133111480865,
|
| 5004 |
+
"grad_norm": 0.2194262146949768,
|
| 5005 |
+
"learning_rate": 0.00013763458774696563,
|
| 5006 |
+
"loss": 0.3433,
|
| 5007 |
+
"step": 1428
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.79312257348863,
|
| 5011 |
+
"grad_norm": 0.26900714635849,
|
| 5012 |
+
"learning_rate": 0.00013746806324206173,
|
| 5013 |
+
"loss": 0.5099,
|
| 5014 |
+
"step": 1430
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.7942318358291736,
|
| 5018 |
+
"grad_norm": 0.2423945814371109,
|
| 5019 |
+
"learning_rate": 0.00013730141778054962,
|
| 5020 |
+
"loss": 0.333,
|
| 5021 |
+
"step": 1432
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.7953410981697171,
|
| 5025 |
+
"grad_norm": 0.31677019596099854,
|
| 5026 |
+
"learning_rate": 0.00013713465190040415,
|
| 5027 |
+
"loss": 0.4285,
|
| 5028 |
+
"step": 1434
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 0.7964503605102606,
|
| 5032 |
+
"grad_norm": 0.33169832825660706,
|
| 5033 |
+
"learning_rate": 0.0001369677661399886,
|
| 5034 |
+
"loss": 0.4058,
|
| 5035 |
+
"step": 1436
|
| 5036 |
+
},
|
| 5037 |
+
{
|
| 5038 |
+
"epoch": 0.7975596228508042,
|
| 5039 |
+
"grad_norm": 0.3489621579647064,
|
| 5040 |
+
"learning_rate": 0.0001368007610380535,
|
| 5041 |
+
"loss": 0.4153,
|
| 5042 |
+
"step": 1438
|
| 5043 |
+
},
|
| 5044 |
+
{
|
| 5045 |
+
"epoch": 0.7986688851913477,
|
| 5046 |
+
"grad_norm": 0.2639998495578766,
|
| 5047 |
+
"learning_rate": 0.00013663363713373454,
|
| 5048 |
+
"loss": 0.3959,
|
| 5049 |
+
"step": 1440
|
| 5050 |
+
},
|
| 5051 |
+
{
|
| 5052 |
+
"epoch": 0.7997781475318912,
|
| 5053 |
+
"grad_norm": 0.27134037017822266,
|
| 5054 |
+
"learning_rate": 0.0001364663949665509,
|
| 5055 |
+
"loss": 0.4331,
|
| 5056 |
+
"step": 1442
|
| 5057 |
+
},
|
| 5058 |
+
{
|
| 5059 |
+
"epoch": 0.8008874098724348,
|
| 5060 |
+
"grad_norm": 0.28077998757362366,
|
| 5061 |
+
"learning_rate": 0.00013629903507640369,
|
| 5062 |
+
"loss": 0.5282,
|
| 5063 |
+
"step": 1444
|
| 5064 |
+
},
|
| 5065 |
+
{
|
| 5066 |
+
"epoch": 0.8019966722129783,
|
| 5067 |
+
"grad_norm": 0.2948680818080902,
|
| 5068 |
+
"learning_rate": 0.00013613155800357385,
|
| 5069 |
+
"loss": 0.3951,
|
| 5070 |
+
"step": 1446
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 0.8031059345535219,
|
| 5074 |
+
"grad_norm": 0.3196534812450409,
|
| 5075 |
+
"learning_rate": 0.0001359639642887208,
|
| 5076 |
+
"loss": 0.4191,
|
| 5077 |
+
"step": 1448
|
| 5078 |
+
},
|
| 5079 |
+
{
|
| 5080 |
+
"epoch": 0.8042151968940654,
|
| 5081 |
+
"grad_norm": 0.26202476024627686,
|
| 5082 |
+
"learning_rate": 0.00013579625447288044,
|
| 5083 |
+
"loss": 0.352,
|
| 5084 |
+
"step": 1450
|
| 5085 |
+
},
|
| 5086 |
+
{
|
| 5087 |
+
"epoch": 0.8053244592346089,
|
| 5088 |
+
"grad_norm": 0.24866290390491486,
|
| 5089 |
+
"learning_rate": 0.00013562842909746342,
|
| 5090 |
+
"loss": 0.3203,
|
| 5091 |
+
"step": 1452
|
| 5092 |
+
},
|
| 5093 |
+
{
|
| 5094 |
+
"epoch": 0.8064337215751525,
|
| 5095 |
+
"grad_norm": 0.41027557849884033,
|
| 5096 |
+
"learning_rate": 0.00013546048870425356,
|
| 5097 |
+
"loss": 0.4187,
|
| 5098 |
+
"step": 1454
|
| 5099 |
+
},
|
| 5100 |
+
{
|
| 5101 |
+
"epoch": 0.807542983915696,
|
| 5102 |
+
"grad_norm": 0.34640318155288696,
|
| 5103 |
+
"learning_rate": 0.0001352924338354059,
|
| 5104 |
+
"loss": 0.3204,
|
| 5105 |
+
"step": 1456
|
| 5106 |
+
},
|
| 5107 |
+
{
|
| 5108 |
+
"epoch": 0.8086522462562395,
|
| 5109 |
+
"grad_norm": 0.3971330523490906,
|
| 5110 |
+
"learning_rate": 0.0001351242650334451,
|
| 5111 |
+
"loss": 0.4598,
|
| 5112 |
+
"step": 1458
|
| 5113 |
+
},
|
| 5114 |
+
{
|
| 5115 |
+
"epoch": 0.8097615085967831,
|
| 5116 |
+
"grad_norm": 0.3868078887462616,
|
| 5117 |
+
"learning_rate": 0.0001349559828412635,
|
| 5118 |
+
"loss": 0.3641,
|
| 5119 |
+
"step": 1460
|
| 5120 |
+
},
|
| 5121 |
+
{
|
| 5122 |
+
"epoch": 0.8108707709373266,
|
| 5123 |
+
"grad_norm": 0.26136353611946106,
|
| 5124 |
+
"learning_rate": 0.00013478758780211965,
|
| 5125 |
+
"loss": 0.4286,
|
| 5126 |
+
"step": 1462
|
| 5127 |
+
},
|
| 5128 |
+
{
|
| 5129 |
+
"epoch": 0.8119800332778702,
|
| 5130 |
+
"grad_norm": 0.2987557351589203,
|
| 5131 |
+
"learning_rate": 0.00013461908045963634,
|
| 5132 |
+
"loss": 0.4286,
|
| 5133 |
+
"step": 1464
|
| 5134 |
+
},
|
| 5135 |
+
{
|
| 5136 |
+
"epoch": 0.8130892956184138,
|
| 5137 |
+
"grad_norm": 0.33275967836380005,
|
| 5138 |
+
"learning_rate": 0.00013445046135779885,
|
| 5139 |
+
"loss": 0.3616,
|
| 5140 |
+
"step": 1466
|
| 5141 |
+
},
|
| 5142 |
+
{
|
| 5143 |
+
"epoch": 0.8141985579589573,
|
| 5144 |
+
"grad_norm": 0.2795950770378113,
|
| 5145 |
+
"learning_rate": 0.00013428173104095331,
|
| 5146 |
+
"loss": 0.5246,
|
| 5147 |
+
"step": 1468
|
| 5148 |
+
},
|
| 5149 |
+
{
|
| 5150 |
+
"epoch": 0.8153078202995009,
|
| 5151 |
+
"grad_norm": 0.3052369952201843,
|
| 5152 |
+
"learning_rate": 0.00013411289005380494,
|
| 5153 |
+
"loss": 0.3672,
|
| 5154 |
+
"step": 1470
|
| 5155 |
+
},
|
| 5156 |
+
{
|
| 5157 |
+
"epoch": 0.8164170826400444,
|
| 5158 |
+
"grad_norm": 0.2728108763694763,
|
| 5159 |
+
"learning_rate": 0.00013394393894141605,
|
| 5160 |
+
"loss": 0.4897,
|
| 5161 |
+
"step": 1472
|
| 5162 |
+
},
|
| 5163 |
+
{
|
| 5164 |
+
"epoch": 0.817526344980588,
|
| 5165 |
+
"grad_norm": 0.3310782015323639,
|
| 5166 |
+
"learning_rate": 0.00013377487824920459,
|
| 5167 |
+
"loss": 0.5144,
|
| 5168 |
+
"step": 1474
|
| 5169 |
+
},
|
| 5170 |
+
{
|
| 5171 |
+
"epoch": 0.8186356073211315,
|
| 5172 |
+
"grad_norm": 0.4153352677822113,
|
| 5173 |
+
"learning_rate": 0.00013360570852294227,
|
| 5174 |
+
"loss": 0.5313,
|
| 5175 |
+
"step": 1476
|
| 5176 |
+
},
|
| 5177 |
+
{
|
| 5178 |
+
"epoch": 0.819744869661675,
|
| 5179 |
+
"grad_norm": 0.4107600450515747,
|
| 5180 |
+
"learning_rate": 0.00013343643030875276,
|
| 5181 |
+
"loss": 0.4873,
|
| 5182 |
+
"step": 1478
|
| 5183 |
+
},
|
| 5184 |
+
{
|
| 5185 |
+
"epoch": 0.8208541320022186,
|
| 5186 |
+
"grad_norm": 0.2750249207019806,
|
| 5187 |
+
"learning_rate": 0.00013326704415311,
|
| 5188 |
+
"loss": 0.4373,
|
| 5189 |
+
"step": 1480
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 0.8219633943427621,
|
| 5193 |
+
"grad_norm": 0.2961103618144989,
|
| 5194 |
+
"learning_rate": 0.00013309755060283626,
|
| 5195 |
+
"loss": 0.4252,
|
| 5196 |
+
"step": 1482
|
| 5197 |
+
},
|
| 5198 |
+
{
|
| 5199 |
+
"epoch": 0.8230726566833056,
|
| 5200 |
+
"grad_norm": 0.25156864523887634,
|
| 5201 |
+
"learning_rate": 0.00013292795020510066,
|
| 5202 |
+
"loss": 0.307,
|
| 5203 |
+
"step": 1484
|
| 5204 |
+
},
|
| 5205 |
+
{
|
| 5206 |
+
"epoch": 0.8241819190238492,
|
| 5207 |
+
"grad_norm": 0.2984169125556946,
|
| 5208 |
+
"learning_rate": 0.00013275824350741716,
|
| 5209 |
+
"loss": 0.4416,
|
| 5210 |
+
"step": 1486
|
| 5211 |
+
},
|
| 5212 |
+
{
|
| 5213 |
+
"epoch": 0.8252911813643927,
|
| 5214 |
+
"grad_norm": 0.27711641788482666,
|
| 5215 |
+
"learning_rate": 0.00013258843105764297,
|
| 5216 |
+
"loss": 0.3382,
|
| 5217 |
+
"step": 1488
|
| 5218 |
+
},
|
| 5219 |
+
{
|
| 5220 |
+
"epoch": 0.8264004437049363,
|
| 5221 |
+
"grad_norm": 0.3105831742286682,
|
| 5222 |
+
"learning_rate": 0.00013241851340397656,
|
| 5223 |
+
"loss": 0.4276,
|
| 5224 |
+
"step": 1490
|
| 5225 |
+
},
|
| 5226 |
+
{
|
| 5227 |
+
"epoch": 0.8275097060454798,
|
| 5228 |
+
"grad_norm": 0.3303448557853699,
|
| 5229 |
+
"learning_rate": 0.00013224849109495622,
|
| 5230 |
+
"loss": 0.4679,
|
| 5231 |
+
"step": 1492
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 0.8286189683860233,
|
| 5235 |
+
"grad_norm": 0.3527246415615082,
|
| 5236 |
+
"learning_rate": 0.00013207836467945785,
|
| 5237 |
+
"loss": 0.5059,
|
| 5238 |
+
"step": 1494
|
| 5239 |
+
},
|
| 5240 |
+
{
|
| 5241 |
+
"epoch": 0.8297282307265669,
|
| 5242 |
+
"grad_norm": 0.41127604246139526,
|
| 5243 |
+
"learning_rate": 0.00013190813470669363,
|
| 5244 |
+
"loss": 0.5412,
|
| 5245 |
+
"step": 1496
|
| 5246 |
+
},
|
| 5247 |
+
{
|
| 5248 |
+
"epoch": 0.8308374930671104,
|
| 5249 |
+
"grad_norm": 0.24502909183502197,
|
| 5250 |
+
"learning_rate": 0.00013173780172620999,
|
| 5251 |
+
"loss": 0.3072,
|
| 5252 |
+
"step": 1498
|
| 5253 |
+
},
|
| 5254 |
+
{
|
| 5255 |
+
"epoch": 0.831946755407654,
|
| 5256 |
+
"grad_norm": 0.21016305685043335,
|
| 5257 |
+
"learning_rate": 0.00013156736628788584,
|
| 5258 |
+
"loss": 0.3889,
|
| 5259 |
+
"step": 1500
|
| 5260 |
+
},
|
| 5261 |
+
{
|
| 5262 |
+
"epoch": 0.8330560177481975,
|
| 5263 |
+
"grad_norm": 0.5300863981246948,
|
| 5264 |
+
"learning_rate": 0.000131396828941931,
|
| 5265 |
+
"loss": 0.5858,
|
| 5266 |
+
"step": 1502
|
| 5267 |
+
},
|
| 5268 |
+
{
|
| 5269 |
+
"epoch": 0.834165280088741,
|
| 5270 |
+
"grad_norm": 0.2194632738828659,
|
| 5271 |
+
"learning_rate": 0.00013122619023888402,
|
| 5272 |
+
"loss": 0.3433,
|
| 5273 |
+
"step": 1504
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 0.8352745424292846,
|
| 5277 |
+
"grad_norm": 0.7686997056007385,
|
| 5278 |
+
"learning_rate": 0.00013105545072961093,
|
| 5279 |
+
"loss": 0.5005,
|
| 5280 |
+
"step": 1506
|
| 5281 |
+
},
|
| 5282 |
+
{
|
| 5283 |
+
"epoch": 0.8363838047698281,
|
| 5284 |
+
"grad_norm": 0.2888992428779602,
|
| 5285 |
+
"learning_rate": 0.00013088461096530304,
|
| 5286 |
+
"loss": 0.3987,
|
| 5287 |
+
"step": 1508
|
| 5288 |
+
},
|
| 5289 |
+
{
|
| 5290 |
+
"epoch": 0.8374930671103716,
|
| 5291 |
+
"grad_norm": 0.2540237009525299,
|
| 5292 |
+
"learning_rate": 0.00013071367149747535,
|
| 5293 |
+
"loss": 0.4531,
|
| 5294 |
+
"step": 1510
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.8386023294509152,
|
| 5298 |
+
"grad_norm": 0.3219870924949646,
|
| 5299 |
+
"learning_rate": 0.00013054263287796465,
|
| 5300 |
+
"loss": 0.561,
|
| 5301 |
+
"step": 1512
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.8397115917914587,
|
| 5305 |
+
"grad_norm": 0.24539148807525635,
|
| 5306 |
+
"learning_rate": 0.00013037149565892794,
|
| 5307 |
+
"loss": 0.3225,
|
| 5308 |
+
"step": 1514
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.8408208541320022,
|
| 5312 |
+
"grad_norm": 0.23382632434368134,
|
| 5313 |
+
"learning_rate": 0.00013020026039284045,
|
| 5314 |
+
"loss": 0.3557,
|
| 5315 |
+
"step": 1516
|
| 5316 |
+
},
|
| 5317 |
+
{
|
| 5318 |
+
"epoch": 0.8419301164725458,
|
| 5319 |
+
"grad_norm": 0.2739519476890564,
|
| 5320 |
+
"learning_rate": 0.00013002892763249398,
|
| 5321 |
+
"loss": 0.3635,
|
| 5322 |
+
"step": 1518
|
| 5323 |
+
},
|
| 5324 |
+
{
|
| 5325 |
+
"epoch": 0.8430393788130893,
|
| 5326 |
+
"grad_norm": 0.23864848911762238,
|
| 5327 |
+
"learning_rate": 0.000129857497930995,
|
| 5328 |
+
"loss": 0.393,
|
| 5329 |
+
"step": 1520
|
| 5330 |
+
},
|
| 5331 |
+
{
|
| 5332 |
+
"epoch": 0.8441486411536329,
|
| 5333 |
+
"grad_norm": 0.3150426149368286,
|
| 5334 |
+
"learning_rate": 0.00012968597184176298,
|
| 5335 |
+
"loss": 0.4337,
|
| 5336 |
+
"step": 1522
|
| 5337 |
+
},
|
| 5338 |
+
{
|
| 5339 |
+
"epoch": 0.8452579034941764,
|
| 5340 |
+
"grad_norm": 0.25435131788253784,
|
| 5341 |
+
"learning_rate": 0.00012951434991852857,
|
| 5342 |
+
"loss": 0.3681,
|
| 5343 |
+
"step": 1524
|
| 5344 |
+
},
|
| 5345 |
+
{
|
| 5346 |
+
"epoch": 0.8463671658347199,
|
| 5347 |
+
"grad_norm": 0.3123234808444977,
|
| 5348 |
+
"learning_rate": 0.0001293426327153317,
|
| 5349 |
+
"loss": 0.4089,
|
| 5350 |
+
"step": 1526
|
| 5351 |
+
},
|
| 5352 |
+
{
|
| 5353 |
+
"epoch": 0.8474764281752635,
|
| 5354 |
+
"grad_norm": 0.52870112657547,
|
| 5355 |
+
"learning_rate": 0.00012917082078652,
|
| 5356 |
+
"loss": 0.4779,
|
| 5357 |
+
"step": 1528
|
| 5358 |
+
},
|
| 5359 |
+
{
|
| 5360 |
+
"epoch": 0.848585690515807,
|
| 5361 |
+
"grad_norm": 0.2784457206726074,
|
| 5362 |
+
"learning_rate": 0.00012899891468674688,
|
| 5363 |
+
"loss": 0.4163,
|
| 5364 |
+
"step": 1530
|
| 5365 |
+
},
|
| 5366 |
+
{
|
| 5367 |
+
"epoch": 0.8496949528563505,
|
| 5368 |
+
"grad_norm": 0.26158082485198975,
|
| 5369 |
+
"learning_rate": 0.0001288269149709697,
|
| 5370 |
+
"loss": 0.464,
|
| 5371 |
+
"step": 1532
|
| 5372 |
+
},
|
| 5373 |
+
{
|
| 5374 |
+
"epoch": 0.8508042151968941,
|
| 5375 |
+
"grad_norm": 0.2595478594303131,
|
| 5376 |
+
"learning_rate": 0.00012865482219444804,
|
| 5377 |
+
"loss": 0.3473,
|
| 5378 |
+
"step": 1534
|
| 5379 |
+
},
|
| 5380 |
+
{
|
| 5381 |
+
"epoch": 0.8519134775374376,
|
| 5382 |
+
"grad_norm": 0.28629809617996216,
|
| 5383 |
+
"learning_rate": 0.000128482636912742,
|
| 5384 |
+
"loss": 0.4626,
|
| 5385 |
+
"step": 1536
|
| 5386 |
+
},
|
| 5387 |
+
{
|
| 5388 |
+
"epoch": 0.8530227398779812,
|
| 5389 |
+
"grad_norm": 0.2769101560115814,
|
| 5390 |
+
"learning_rate": 0.00012831035968171025,
|
| 5391 |
+
"loss": 0.5318,
|
| 5392 |
+
"step": 1538
|
| 5393 |
+
},
|
| 5394 |
+
{
|
| 5395 |
+
"epoch": 0.8541320022185247,
|
| 5396 |
+
"grad_norm": 0.32585474848747253,
|
| 5397 |
+
"learning_rate": 0.00012813799105750823,
|
| 5398 |
+
"loss": 0.4428,
|
| 5399 |
+
"step": 1540
|
| 5400 |
+
},
|
| 5401 |
+
{
|
| 5402 |
+
"epoch": 0.8552412645590682,
|
| 5403 |
+
"grad_norm": 0.2759438157081604,
|
| 5404 |
+
"learning_rate": 0.00012796553159658653,
|
| 5405 |
+
"loss": 0.3853,
|
| 5406 |
+
"step": 1542
|
| 5407 |
+
},
|
| 5408 |
+
{
|
| 5409 |
+
"epoch": 0.8563505268996118,
|
| 5410 |
+
"grad_norm": 0.34193405508995056,
|
| 5411 |
+
"learning_rate": 0.0001277929818556889,
|
| 5412 |
+
"loss": 0.5144,
|
| 5413 |
+
"step": 1544
|
| 5414 |
+
},
|
| 5415 |
+
{
|
| 5416 |
+
"epoch": 0.8574597892401553,
|
| 5417 |
+
"grad_norm": 0.3385421335697174,
|
| 5418 |
+
"learning_rate": 0.00012762034239185063,
|
| 5419 |
+
"loss": 0.4466,
|
| 5420 |
+
"step": 1546
|
| 5421 |
+
},
|
| 5422 |
+
{
|
| 5423 |
+
"epoch": 0.8585690515806988,
|
| 5424 |
+
"grad_norm": 0.3747256398200989,
|
| 5425 |
+
"learning_rate": 0.00012744761376239655,
|
| 5426 |
+
"loss": 0.517,
|
| 5427 |
+
"step": 1548
|
| 5428 |
+
},
|
| 5429 |
+
{
|
| 5430 |
+
"epoch": 0.8596783139212424,
|
| 5431 |
+
"grad_norm": 0.31231391429901123,
|
| 5432 |
+
"learning_rate": 0.00012727479652493943,
|
| 5433 |
+
"loss": 0.3797,
|
| 5434 |
+
"step": 1550
|
| 5435 |
+
},
|
| 5436 |
+
{
|
| 5437 |
+
"epoch": 0.8607875762617859,
|
| 5438 |
+
"grad_norm": 0.2660077214241028,
|
| 5439 |
+
"learning_rate": 0.00012710189123737802,
|
| 5440 |
+
"loss": 0.3481,
|
| 5441 |
+
"step": 1552
|
| 5442 |
+
},
|
| 5443 |
+
{
|
| 5444 |
+
"epoch": 0.8618968386023295,
|
| 5445 |
+
"grad_norm": 0.34666526317596436,
|
| 5446 |
+
"learning_rate": 0.00012692889845789538,
|
| 5447 |
+
"loss": 0.4369,
|
| 5448 |
+
"step": 1554
|
| 5449 |
+
},
|
| 5450 |
+
{
|
| 5451 |
+
"epoch": 0.863006100942873,
|
| 5452 |
+
"grad_norm": 0.26898959279060364,
|
| 5453 |
+
"learning_rate": 0.00012675581874495697,
|
| 5454 |
+
"loss": 0.4428,
|
| 5455 |
+
"step": 1556
|
| 5456 |
+
},
|
| 5457 |
+
{
|
| 5458 |
+
"epoch": 0.8641153632834165,
|
| 5459 |
+
"grad_norm": 0.378214567899704,
|
| 5460 |
+
"learning_rate": 0.0001265826526573089,
|
| 5461 |
+
"loss": 0.5425,
|
| 5462 |
+
"step": 1558
|
| 5463 |
+
},
|
| 5464 |
+
{
|
| 5465 |
+
"epoch": 0.8652246256239601,
|
| 5466 |
+
"grad_norm": 0.3740502595901489,
|
| 5467 |
+
"learning_rate": 0.0001264094007539762,
|
| 5468 |
+
"loss": 0.5318,
|
| 5469 |
+
"step": 1560
|
| 5470 |
+
},
|
| 5471 |
+
{
|
| 5472 |
+
"epoch": 0.8663338879645036,
|
| 5473 |
+
"grad_norm": 0.25904256105422974,
|
| 5474 |
+
"learning_rate": 0.00012623606359426077,
|
| 5475 |
+
"loss": 0.4032,
|
| 5476 |
+
"step": 1562
|
| 5477 |
+
},
|
| 5478 |
+
{
|
| 5479 |
+
"epoch": 0.8674431503050472,
|
| 5480 |
+
"grad_norm": 0.29247143864631653,
|
| 5481 |
+
"learning_rate": 0.00012606264173773988,
|
| 5482 |
+
"loss": 0.4446,
|
| 5483 |
+
"step": 1564
|
| 5484 |
+
},
|
| 5485 |
+
{
|
| 5486 |
+
"epoch": 0.8685524126455907,
|
| 5487 |
+
"grad_norm": 0.2748325765132904,
|
| 5488 |
+
"learning_rate": 0.0001258891357442642,
|
| 5489 |
+
"loss": 0.4435,
|
| 5490 |
+
"step": 1566
|
| 5491 |
+
},
|
| 5492 |
+
{
|
| 5493 |
+
"epoch": 0.8696616749861342,
|
| 5494 |
+
"grad_norm": 0.4215485155582428,
|
| 5495 |
+
"learning_rate": 0.00012571554617395598,
|
| 5496 |
+
"loss": 0.4285,
|
| 5497 |
+
"step": 1568
|
| 5498 |
+
},
|
| 5499 |
+
{
|
| 5500 |
+
"epoch": 0.8707709373266778,
|
| 5501 |
+
"grad_norm": 0.2875049412250519,
|
| 5502 |
+
"learning_rate": 0.00012554187358720725,
|
| 5503 |
+
"loss": 0.4743,
|
| 5504 |
+
"step": 1570
|
| 5505 |
+
},
|
| 5506 |
+
{
|
| 5507 |
+
"epoch": 0.8718801996672213,
|
| 5508 |
+
"grad_norm": 0.3484318256378174,
|
| 5509 |
+
"learning_rate": 0.00012536811854467817,
|
| 5510 |
+
"loss": 0.4997,
|
| 5511 |
+
"step": 1572
|
| 5512 |
+
},
|
| 5513 |
+
{
|
| 5514 |
+
"epoch": 0.8729894620077648,
|
| 5515 |
+
"grad_norm": 0.33628854155540466,
|
| 5516 |
+
"learning_rate": 0.0001251942816072949,
|
| 5517 |
+
"loss": 0.5044,
|
| 5518 |
+
"step": 1574
|
| 5519 |
+
},
|
| 5520 |
+
{
|
| 5521 |
+
"epoch": 0.8740987243483084,
|
| 5522 |
+
"grad_norm": 0.28208592534065247,
|
| 5523 |
+
"learning_rate": 0.00012502036333624815,
|
| 5524 |
+
"loss": 0.3405,
|
| 5525 |
+
"step": 1576
|
| 5526 |
+
},
|
| 5527 |
+
{
|
| 5528 |
+
"epoch": 0.8752079866888519,
|
| 5529 |
+
"grad_norm": 0.2113669067621231,
|
| 5530 |
+
"learning_rate": 0.00012484636429299114,
|
| 5531 |
+
"loss": 0.3996,
|
| 5532 |
+
"step": 1578
|
| 5533 |
+
},
|
| 5534 |
+
{
|
| 5535 |
+
"epoch": 0.8763172490293955,
|
| 5536 |
+
"grad_norm": 0.2464189976453781,
|
| 5537 |
+
"learning_rate": 0.00012467228503923773,
|
| 5538 |
+
"loss": 0.3572,
|
| 5539 |
+
"step": 1580
|
| 5540 |
+
},
|
| 5541 |
+
{
|
| 5542 |
+
"epoch": 0.877426511369939,
|
| 5543 |
+
"grad_norm": 0.391923189163208,
|
| 5544 |
+
"learning_rate": 0.00012449812613696094,
|
| 5545 |
+
"loss": 0.6324,
|
| 5546 |
+
"step": 1582
|
| 5547 |
+
},
|
| 5548 |
+
{
|
| 5549 |
+
"epoch": 0.8785357737104825,
|
| 5550 |
+
"grad_norm": 0.2782968282699585,
|
| 5551 |
+
"learning_rate": 0.0001243238881483907,
|
| 5552 |
+
"loss": 0.352,
|
| 5553 |
+
"step": 1584
|
| 5554 |
+
},
|
| 5555 |
+
{
|
| 5556 |
+
"epoch": 0.8796450360510261,
|
| 5557 |
+
"grad_norm": 0.30209866166114807,
|
| 5558 |
+
"learning_rate": 0.00012414957163601236,
|
| 5559 |
+
"loss": 0.4611,
|
| 5560 |
+
"step": 1586
|
| 5561 |
+
},
|
| 5562 |
+
{
|
| 5563 |
+
"epoch": 0.8807542983915696,
|
| 5564 |
+
"grad_norm": 0.2990707457065582,
|
| 5565 |
+
"learning_rate": 0.0001239751771625648,
|
| 5566 |
+
"loss": 0.5419,
|
| 5567 |
+
"step": 1588
|
| 5568 |
+
},
|
| 5569 |
+
{
|
| 5570 |
+
"epoch": 0.8818635607321131,
|
| 5571 |
+
"grad_norm": 0.25346022844314575,
|
| 5572 |
+
"learning_rate": 0.00012380070529103852,
|
| 5573 |
+
"loss": 0.4205,
|
| 5574 |
+
"step": 1590
|
| 5575 |
+
},
|
| 5576 |
+
{
|
| 5577 |
+
"epoch": 0.8829728230726567,
|
| 5578 |
+
"grad_norm": 0.33787423372268677,
|
| 5579 |
+
"learning_rate": 0.00012362615658467377,
|
| 5580 |
+
"loss": 0.4186,
|
| 5581 |
+
"step": 1592
|
| 5582 |
+
},
|
| 5583 |
+
{
|
| 5584 |
+
"epoch": 0.8840820854132002,
|
| 5585 |
+
"grad_norm": 0.23708541691303253,
|
| 5586 |
+
"learning_rate": 0.00012345153160695917,
|
| 5587 |
+
"loss": 0.3261,
|
| 5588 |
+
"step": 1594
|
| 5589 |
+
},
|
| 5590 |
+
{
|
| 5591 |
+
"epoch": 0.8851913477537438,
|
| 5592 |
+
"grad_norm": 0.22642351686954498,
|
| 5593 |
+
"learning_rate": 0.00012327683092162918,
|
| 5594 |
+
"loss": 0.3415,
|
| 5595 |
+
"step": 1596
|
| 5596 |
+
},
|
| 5597 |
+
{
|
| 5598 |
+
"epoch": 0.8863006100942873,
|
| 5599 |
+
"grad_norm": 0.27935388684272766,
|
| 5600 |
+
"learning_rate": 0.00012310205509266292,
|
| 5601 |
+
"loss": 0.3471,
|
| 5602 |
+
"step": 1598
|
| 5603 |
+
},
|
| 5604 |
+
{
|
| 5605 |
+
"epoch": 0.8874098724348308,
|
| 5606 |
+
"grad_norm": 0.23772071301937103,
|
| 5607 |
+
"learning_rate": 0.000122927204684282,
|
| 5608 |
+
"loss": 0.3386,
|
| 5609 |
+
"step": 1600
|
| 5610 |
}
|
| 5611 |
],
|
| 5612 |
"logging_steps": 2,
|
|
|
|
| 5626 |
"attributes": {}
|
| 5627 |
}
|
| 5628 |
},
|
| 5629 |
+
"total_flos": 3897330499584000.0,
|
| 5630 |
"train_batch_size": 8,
|
| 5631 |
"trial_name": null,
|
| 5632 |
"trial_params": null
|