loss 2.2
Browse files- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +1113 -3
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 352532601
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ac120fe0be67d7968589a3cff544029928b6f77e268ee2cedbf3b1859418daa
|
| 3 |
size 352532601
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 688496379
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67c94fb4ffcee06e9c4e7cb3024c86be9cee95ca73b0f2660b397877828b8d8f
|
| 3 |
size 688496379
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d112fe17318190ebeb1922d4d2da7834c36312edc1f5bdc0c651043716930cb
|
| 3 |
size 14503
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb787f0d035f8eb29c22b3bc6a15e850bc7e6a3e174e71d31e92c39f38834e49
|
| 3 |
size 623
|
trainer_state.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
@@ -2646,11 +2646,1121 @@
|
|
| 2646 |
"learning_rate": 0.0005221027479091995,
|
| 2647 |
"loss": 2.2734,
|
| 2648 |
"step": 44000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2649 |
}
|
| 2650 |
],
|
| 2651 |
"max_steps": 92070,
|
| 2652 |
"num_train_epochs": 10,
|
| 2653 |
-
"total_flos":
|
| 2654 |
"trial_name": null,
|
| 2655 |
"trial_params": null
|
| 2656 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.788291835238535,
|
| 5 |
+
"global_step": 62500,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
|
|
|
| 2646 |
"learning_rate": 0.0005221027479091995,
|
| 2647 |
"loss": 2.2734,
|
| 2648 |
"step": 44000
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 4.79,
|
| 2652 |
+
"learning_rate": 0.0005210166177908114,
|
| 2653 |
+
"loss": 2.2511,
|
| 2654 |
+
"step": 44100
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 4.8,
|
| 2658 |
+
"learning_rate": 0.0005199304876724231,
|
| 2659 |
+
"loss": 2.3169,
|
| 2660 |
+
"step": 44200
|
| 2661 |
+
},
|
| 2662 |
+
{
|
| 2663 |
+
"epoch": 4.81,
|
| 2664 |
+
"learning_rate": 0.0005188443575540351,
|
| 2665 |
+
"loss": 2.2753,
|
| 2666 |
+
"step": 44300
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 4.82,
|
| 2670 |
+
"learning_rate": 0.0005177582274356468,
|
| 2671 |
+
"loss": 2.2067,
|
| 2672 |
+
"step": 44400
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 4.83,
|
| 2676 |
+
"learning_rate": 0.0005166720973172587,
|
| 2677 |
+
"loss": 2.1947,
|
| 2678 |
+
"step": 44500
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 4.84,
|
| 2682 |
+
"learning_rate": 0.0005155859671988704,
|
| 2683 |
+
"loss": 2.2377,
|
| 2684 |
+
"step": 44600
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 4.85,
|
| 2688 |
+
"learning_rate": 0.0005144998370804822,
|
| 2689 |
+
"loss": 2.2191,
|
| 2690 |
+
"step": 44700
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 4.87,
|
| 2694 |
+
"learning_rate": 0.0005134137069620941,
|
| 2695 |
+
"loss": 2.2433,
|
| 2696 |
+
"step": 44800
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 4.88,
|
| 2700 |
+
"learning_rate": 0.0005123275768437059,
|
| 2701 |
+
"loss": 2.2032,
|
| 2702 |
+
"step": 44900
|
| 2703 |
+
},
|
| 2704 |
+
{
|
| 2705 |
+
"epoch": 4.89,
|
| 2706 |
+
"learning_rate": 0.0005112414467253177,
|
| 2707 |
+
"loss": 2.2909,
|
| 2708 |
+
"step": 45000
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 4.9,
|
| 2712 |
+
"learning_rate": 0.0005101553166069295,
|
| 2713 |
+
"loss": 2.2409,
|
| 2714 |
+
"step": 45100
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 4.91,
|
| 2718 |
+
"learning_rate": 0.0005090691864885414,
|
| 2719 |
+
"loss": 2.222,
|
| 2720 |
+
"step": 45200
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 4.92,
|
| 2724 |
+
"learning_rate": 0.0005079830563701532,
|
| 2725 |
+
"loss": 2.1896,
|
| 2726 |
+
"step": 45300
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 4.93,
|
| 2730 |
+
"learning_rate": 0.000506896926251765,
|
| 2731 |
+
"loss": 2.2844,
|
| 2732 |
+
"step": 45400
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 4.94,
|
| 2736 |
+
"learning_rate": 0.0005058107961333768,
|
| 2737 |
+
"loss": 2.2575,
|
| 2738 |
+
"step": 45500
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 4.95,
|
| 2742 |
+
"learning_rate": 0.0005047246660149886,
|
| 2743 |
+
"loss": 2.2653,
|
| 2744 |
+
"step": 45600
|
| 2745 |
+
},
|
| 2746 |
+
{
|
| 2747 |
+
"epoch": 4.96,
|
| 2748 |
+
"learning_rate": 0.0005036385358966005,
|
| 2749 |
+
"loss": 2.3424,
|
| 2750 |
+
"step": 45700
|
| 2751 |
+
},
|
| 2752 |
+
{
|
| 2753 |
+
"epoch": 4.97,
|
| 2754 |
+
"learning_rate": 0.0005025524057782122,
|
| 2755 |
+
"loss": 2.226,
|
| 2756 |
+
"step": 45800
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 4.99,
|
| 2760 |
+
"learning_rate": 0.0005014662756598241,
|
| 2761 |
+
"loss": 2.2514,
|
| 2762 |
+
"step": 45900
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 5.0,
|
| 2766 |
+
"learning_rate": 0.0005003801455414358,
|
| 2767 |
+
"loss": 2.221,
|
| 2768 |
+
"step": 46000
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 5.01,
|
| 2772 |
+
"learning_rate": 0.0004992940154230477,
|
| 2773 |
+
"loss": 2.2391,
|
| 2774 |
+
"step": 46100
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 5.02,
|
| 2778 |
+
"learning_rate": 0.0004982078853046595,
|
| 2779 |
+
"loss": 2.1091,
|
| 2780 |
+
"step": 46200
|
| 2781 |
+
},
|
| 2782 |
+
{
|
| 2783 |
+
"epoch": 5.03,
|
| 2784 |
+
"learning_rate": 0.0004971217551862714,
|
| 2785 |
+
"loss": 2.1705,
|
| 2786 |
+
"step": 46300
|
| 2787 |
+
},
|
| 2788 |
+
{
|
| 2789 |
+
"epoch": 5.04,
|
| 2790 |
+
"learning_rate": 0.0004960356250678831,
|
| 2791 |
+
"loss": 2.2045,
|
| 2792 |
+
"step": 46400
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 5.05,
|
| 2796 |
+
"learning_rate": 0.000494949494949495,
|
| 2797 |
+
"loss": 2.0964,
|
| 2798 |
+
"step": 46500
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 5.06,
|
| 2802 |
+
"learning_rate": 0.0004938633648311068,
|
| 2803 |
+
"loss": 2.1545,
|
| 2804 |
+
"step": 46600
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 5.07,
|
| 2808 |
+
"learning_rate": 0.0004927772347127186,
|
| 2809 |
+
"loss": 2.1766,
|
| 2810 |
+
"step": 46700
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 5.08,
|
| 2814 |
+
"learning_rate": 0.0004916911045943304,
|
| 2815 |
+
"loss": 2.2168,
|
| 2816 |
+
"step": 46800
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 5.09,
|
| 2820 |
+
"learning_rate": 0.0004906049744759422,
|
| 2821 |
+
"loss": 2.1429,
|
| 2822 |
+
"step": 46900
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"epoch": 5.1,
|
| 2826 |
+
"learning_rate": 0.000489518844357554,
|
| 2827 |
+
"loss": 2.1623,
|
| 2828 |
+
"step": 47000
|
| 2829 |
+
},
|
| 2830 |
+
{
|
| 2831 |
+
"epoch": 5.12,
|
| 2832 |
+
"learning_rate": 0.0004884327142391659,
|
| 2833 |
+
"loss": 2.1718,
|
| 2834 |
+
"step": 47100
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 5.13,
|
| 2838 |
+
"learning_rate": 0.0004873465841207777,
|
| 2839 |
+
"loss": 2.2335,
|
| 2840 |
+
"step": 47200
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 5.14,
|
| 2844 |
+
"learning_rate": 0.00048626045400238947,
|
| 2845 |
+
"loss": 2.1698,
|
| 2846 |
+
"step": 47300
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 5.15,
|
| 2850 |
+
"learning_rate": 0.00048517432388400127,
|
| 2851 |
+
"loss": 2.0601,
|
| 2852 |
+
"step": 47400
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 5.16,
|
| 2856 |
+
"learning_rate": 0.0004840881937656131,
|
| 2857 |
+
"loss": 2.187,
|
| 2858 |
+
"step": 47500
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 5.17,
|
| 2862 |
+
"learning_rate": 0.0004830020636472249,
|
| 2863 |
+
"loss": 2.0521,
|
| 2864 |
+
"step": 47600
|
| 2865 |
+
},
|
| 2866 |
+
{
|
| 2867 |
+
"epoch": 5.18,
|
| 2868 |
+
"learning_rate": 0.0004819159335288368,
|
| 2869 |
+
"loss": 2.1387,
|
| 2870 |
+
"step": 47700
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"epoch": 5.19,
|
| 2874 |
+
"learning_rate": 0.0004808298034104486,
|
| 2875 |
+
"loss": 2.1577,
|
| 2876 |
+
"step": 47800
|
| 2877 |
+
},
|
| 2878 |
+
{
|
| 2879 |
+
"epoch": 5.2,
|
| 2880 |
+
"learning_rate": 0.00047974367329206043,
|
| 2881 |
+
"loss": 2.0751,
|
| 2882 |
+
"step": 47900
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 5.21,
|
| 2886 |
+
"learning_rate": 0.00047865754317367223,
|
| 2887 |
+
"loss": 2.1627,
|
| 2888 |
+
"step": 48000
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 5.22,
|
| 2892 |
+
"learning_rate": 0.000477571413055284,
|
| 2893 |
+
"loss": 2.1614,
|
| 2894 |
+
"step": 48100
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 5.24,
|
| 2898 |
+
"learning_rate": 0.0004764852829368959,
|
| 2899 |
+
"loss": 2.092,
|
| 2900 |
+
"step": 48200
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 5.25,
|
| 2904 |
+
"learning_rate": 0.0004753991528185077,
|
| 2905 |
+
"loss": 2.1432,
|
| 2906 |
+
"step": 48300
|
| 2907 |
+
},
|
| 2908 |
+
{
|
| 2909 |
+
"epoch": 5.26,
|
| 2910 |
+
"learning_rate": 0.0004743130227001195,
|
| 2911 |
+
"loss": 2.1394,
|
| 2912 |
+
"step": 48400
|
| 2913 |
+
},
|
| 2914 |
+
{
|
| 2915 |
+
"epoch": 5.27,
|
| 2916 |
+
"learning_rate": 0.0004732268925817313,
|
| 2917 |
+
"loss": 2.1308,
|
| 2918 |
+
"step": 48500
|
| 2919 |
+
},
|
| 2920 |
+
{
|
| 2921 |
+
"epoch": 5.28,
|
| 2922 |
+
"learning_rate": 0.00047214076246334313,
|
| 2923 |
+
"loss": 2.1903,
|
| 2924 |
+
"step": 48600
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 5.29,
|
| 2928 |
+
"learning_rate": 0.00047105463234495493,
|
| 2929 |
+
"loss": 2.1931,
|
| 2930 |
+
"step": 48700
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 5.3,
|
| 2934 |
+
"learning_rate": 0.00046996850222656673,
|
| 2935 |
+
"loss": 2.1703,
|
| 2936 |
+
"step": 48800
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 5.31,
|
| 2940 |
+
"learning_rate": 0.0004688823721081786,
|
| 2941 |
+
"loss": 2.2013,
|
| 2942 |
+
"step": 48900
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 5.32,
|
| 2946 |
+
"learning_rate": 0.0004677962419897904,
|
| 2947 |
+
"loss": 2.1552,
|
| 2948 |
+
"step": 49000
|
| 2949 |
+
},
|
| 2950 |
+
{
|
| 2951 |
+
"epoch": 5.33,
|
| 2952 |
+
"learning_rate": 0.00046671011187140223,
|
| 2953 |
+
"loss": 2.1587,
|
| 2954 |
+
"step": 49100
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 5.34,
|
| 2958 |
+
"learning_rate": 0.00046562398175301403,
|
| 2959 |
+
"loss": 2.1813,
|
| 2960 |
+
"step": 49200
|
| 2961 |
+
},
|
| 2962 |
+
{
|
| 2963 |
+
"epoch": 5.35,
|
| 2964 |
+
"learning_rate": 0.0004645378516346259,
|
| 2965 |
+
"loss": 2.1841,
|
| 2966 |
+
"step": 49300
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 5.37,
|
| 2970 |
+
"learning_rate": 0.00046345172151623763,
|
| 2971 |
+
"loss": 2.1264,
|
| 2972 |
+
"step": 49400
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 5.38,
|
| 2976 |
+
"learning_rate": 0.00046236559139784943,
|
| 2977 |
+
"loss": 2.1126,
|
| 2978 |
+
"step": 49500
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 5.39,
|
| 2982 |
+
"learning_rate": 0.0004612794612794613,
|
| 2983 |
+
"loss": 2.071,
|
| 2984 |
+
"step": 49600
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 5.4,
|
| 2988 |
+
"learning_rate": 0.0004601933311610731,
|
| 2989 |
+
"loss": 2.173,
|
| 2990 |
+
"step": 49700
|
| 2991 |
+
},
|
| 2992 |
+
{
|
| 2993 |
+
"epoch": 5.41,
|
| 2994 |
+
"learning_rate": 0.00045910720104268493,
|
| 2995 |
+
"loss": 2.0715,
|
| 2996 |
+
"step": 49800
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 5.42,
|
| 3000 |
+
"learning_rate": 0.00045802107092429673,
|
| 3001 |
+
"loss": 2.2097,
|
| 3002 |
+
"step": 49900
|
| 3003 |
+
},
|
| 3004 |
+
{
|
| 3005 |
+
"epoch": 5.43,
|
| 3006 |
+
"learning_rate": 0.0004569349408059086,
|
| 3007 |
+
"loss": 2.1484,
|
| 3008 |
+
"step": 50000
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 5.44,
|
| 3012 |
+
"learning_rate": 0.0004558488106875204,
|
| 3013 |
+
"loss": 2.1509,
|
| 3014 |
+
"step": 50100
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 5.45,
|
| 3018 |
+
"learning_rate": 0.0004547626805691322,
|
| 3019 |
+
"loss": 2.2253,
|
| 3020 |
+
"step": 50200
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 5.46,
|
| 3024 |
+
"learning_rate": 0.00045367655045074404,
|
| 3025 |
+
"loss": 2.1305,
|
| 3026 |
+
"step": 50300
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 5.47,
|
| 3030 |
+
"learning_rate": 0.00045259042033235584,
|
| 3031 |
+
"loss": 2.1049,
|
| 3032 |
+
"step": 50400
|
| 3033 |
+
},
|
| 3034 |
+
{
|
| 3035 |
+
"epoch": 5.48,
|
| 3036 |
+
"learning_rate": 0.00045150429021396764,
|
| 3037 |
+
"loss": 2.1722,
|
| 3038 |
+
"step": 50500
|
| 3039 |
+
},
|
| 3040 |
+
{
|
| 3041 |
+
"epoch": 5.5,
|
| 3042 |
+
"learning_rate": 0.00045041816009557944,
|
| 3043 |
+
"loss": 2.1202,
|
| 3044 |
+
"step": 50600
|
| 3045 |
+
},
|
| 3046 |
+
{
|
| 3047 |
+
"epoch": 5.51,
|
| 3048 |
+
"learning_rate": 0.0004493320299771913,
|
| 3049 |
+
"loss": 2.1804,
|
| 3050 |
+
"step": 50700
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 5.52,
|
| 3054 |
+
"learning_rate": 0.0004482458998588031,
|
| 3055 |
+
"loss": 2.2228,
|
| 3056 |
+
"step": 50800
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 5.53,
|
| 3060 |
+
"learning_rate": 0.0004471597697404149,
|
| 3061 |
+
"loss": 2.2324,
|
| 3062 |
+
"step": 50900
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 5.54,
|
| 3066 |
+
"learning_rate": 0.00044607363962202674,
|
| 3067 |
+
"loss": 2.1689,
|
| 3068 |
+
"step": 51000
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 5.55,
|
| 3072 |
+
"learning_rate": 0.00044498750950363854,
|
| 3073 |
+
"loss": 2.077,
|
| 3074 |
+
"step": 51100
|
| 3075 |
+
},
|
| 3076 |
+
{
|
| 3077 |
+
"epoch": 5.56,
|
| 3078 |
+
"learning_rate": 0.0004439013793852504,
|
| 3079 |
+
"loss": 2.0564,
|
| 3080 |
+
"step": 51200
|
| 3081 |
+
},
|
| 3082 |
+
{
|
| 3083 |
+
"epoch": 5.57,
|
| 3084 |
+
"learning_rate": 0.0004428152492668622,
|
| 3085 |
+
"loss": 2.2077,
|
| 3086 |
+
"step": 51300
|
| 3087 |
+
},
|
| 3088 |
+
{
|
| 3089 |
+
"epoch": 5.58,
|
| 3090 |
+
"learning_rate": 0.00044172911914847404,
|
| 3091 |
+
"loss": 2.1496,
|
| 3092 |
+
"step": 51400
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 5.59,
|
| 3096 |
+
"learning_rate": 0.00044064298903008584,
|
| 3097 |
+
"loss": 2.0937,
|
| 3098 |
+
"step": 51500
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 5.6,
|
| 3102 |
+
"learning_rate": 0.0004395568589116976,
|
| 3103 |
+
"loss": 2.1854,
|
| 3104 |
+
"step": 51600
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 5.62,
|
| 3108 |
+
"learning_rate": 0.00043847072879330944,
|
| 3109 |
+
"loss": 2.165,
|
| 3110 |
+
"step": 51700
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 5.63,
|
| 3114 |
+
"learning_rate": 0.00043738459867492124,
|
| 3115 |
+
"loss": 2.1304,
|
| 3116 |
+
"step": 51800
|
| 3117 |
+
},
|
| 3118 |
+
{
|
| 3119 |
+
"epoch": 5.64,
|
| 3120 |
+
"learning_rate": 0.0004362984685565331,
|
| 3121 |
+
"loss": 2.166,
|
| 3122 |
+
"step": 51900
|
| 3123 |
+
},
|
| 3124 |
+
{
|
| 3125 |
+
"epoch": 5.65,
|
| 3126 |
+
"learning_rate": 0.0004352123384381449,
|
| 3127 |
+
"loss": 2.1038,
|
| 3128 |
+
"step": 52000
|
| 3129 |
+
},
|
| 3130 |
+
{
|
| 3131 |
+
"epoch": 5.66,
|
| 3132 |
+
"learning_rate": 0.00043412620831975675,
|
| 3133 |
+
"loss": 2.1373,
|
| 3134 |
+
"step": 52100
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 5.67,
|
| 3138 |
+
"learning_rate": 0.00043304007820136855,
|
| 3139 |
+
"loss": 2.1287,
|
| 3140 |
+
"step": 52200
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 5.68,
|
| 3144 |
+
"learning_rate": 0.00043195394808298034,
|
| 3145 |
+
"loss": 2.2025,
|
| 3146 |
+
"step": 52300
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 5.69,
|
| 3150 |
+
"learning_rate": 0.0004308678179645922,
|
| 3151 |
+
"loss": 2.1951,
|
| 3152 |
+
"step": 52400
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 5.7,
|
| 3156 |
+
"learning_rate": 0.000429781687846204,
|
| 3157 |
+
"loss": 2.1729,
|
| 3158 |
+
"step": 52500
|
| 3159 |
+
},
|
| 3160 |
+
{
|
| 3161 |
+
"epoch": 5.71,
|
| 3162 |
+
"learning_rate": 0.0004286955577278158,
|
| 3163 |
+
"loss": 2.148,
|
| 3164 |
+
"step": 52600
|
| 3165 |
+
},
|
| 3166 |
+
{
|
| 3167 |
+
"epoch": 5.72,
|
| 3168 |
+
"learning_rate": 0.0004276094276094276,
|
| 3169 |
+
"loss": 2.1946,
|
| 3170 |
+
"step": 52700
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 5.73,
|
| 3174 |
+
"learning_rate": 0.0004265232974910394,
|
| 3175 |
+
"loss": 2.1026,
|
| 3176 |
+
"step": 52800
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 5.75,
|
| 3180 |
+
"learning_rate": 0.00042543716737265125,
|
| 3181 |
+
"loss": 2.1142,
|
| 3182 |
+
"step": 52900
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 5.76,
|
| 3186 |
+
"learning_rate": 0.00042435103725426305,
|
| 3187 |
+
"loss": 2.2227,
|
| 3188 |
+
"step": 53000
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 5.77,
|
| 3192 |
+
"learning_rate": 0.0004232649071358749,
|
| 3193 |
+
"loss": 2.1227,
|
| 3194 |
+
"step": 53100
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 5.78,
|
| 3198 |
+
"learning_rate": 0.0004221787770174867,
|
| 3199 |
+
"loss": 2.1521,
|
| 3200 |
+
"step": 53200
|
| 3201 |
+
},
|
| 3202 |
+
{
|
| 3203 |
+
"epoch": 5.79,
|
| 3204 |
+
"learning_rate": 0.00042109264689909855,
|
| 3205 |
+
"loss": 2.1634,
|
| 3206 |
+
"step": 53300
|
| 3207 |
+
},
|
| 3208 |
+
{
|
| 3209 |
+
"epoch": 5.8,
|
| 3210 |
+
"learning_rate": 0.00042000651678071035,
|
| 3211 |
+
"loss": 2.1214,
|
| 3212 |
+
"step": 53400
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 5.81,
|
| 3216 |
+
"learning_rate": 0.00041892038666232215,
|
| 3217 |
+
"loss": 2.1831,
|
| 3218 |
+
"step": 53500
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 5.82,
|
| 3222 |
+
"learning_rate": 0.000417834256543934,
|
| 3223 |
+
"loss": 2.1929,
|
| 3224 |
+
"step": 53600
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 5.83,
|
| 3228 |
+
"learning_rate": 0.00041674812642554575,
|
| 3229 |
+
"loss": 2.1744,
|
| 3230 |
+
"step": 53700
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 5.84,
|
| 3234 |
+
"learning_rate": 0.0004156619963071576,
|
| 3235 |
+
"loss": 2.0353,
|
| 3236 |
+
"step": 53800
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 5.85,
|
| 3240 |
+
"learning_rate": 0.0004145758661887694,
|
| 3241 |
+
"loss": 2.1524,
|
| 3242 |
+
"step": 53900
|
| 3243 |
+
},
|
| 3244 |
+
{
|
| 3245 |
+
"epoch": 5.87,
|
| 3246 |
+
"learning_rate": 0.00041348973607038125,
|
| 3247 |
+
"loss": 2.1994,
|
| 3248 |
+
"step": 54000
|
| 3249 |
+
},
|
| 3250 |
+
{
|
| 3251 |
+
"epoch": 5.88,
|
| 3252 |
+
"learning_rate": 0.00041240360595199305,
|
| 3253 |
+
"loss": 2.2203,
|
| 3254 |
+
"step": 54100
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 5.89,
|
| 3258 |
+
"learning_rate": 0.00041131747583360485,
|
| 3259 |
+
"loss": 2.1567,
|
| 3260 |
+
"step": 54200
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 5.9,
|
| 3264 |
+
"learning_rate": 0.0004102313457152167,
|
| 3265 |
+
"loss": 2.1588,
|
| 3266 |
+
"step": 54300
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 5.91,
|
| 3270 |
+
"learning_rate": 0.0004091452155968285,
|
| 3271 |
+
"loss": 2.1359,
|
| 3272 |
+
"step": 54400
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 5.92,
|
| 3276 |
+
"learning_rate": 0.00040805908547844036,
|
| 3277 |
+
"loss": 2.1171,
|
| 3278 |
+
"step": 54500
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 5.93,
|
| 3282 |
+
"learning_rate": 0.00040697295536005216,
|
| 3283 |
+
"loss": 2.1239,
|
| 3284 |
+
"step": 54600
|
| 3285 |
+
},
|
| 3286 |
+
{
|
| 3287 |
+
"epoch": 5.94,
|
| 3288 |
+
"learning_rate": 0.00040588682524166395,
|
| 3289 |
+
"loss": 2.0922,
|
| 3290 |
+
"step": 54700
|
| 3291 |
+
},
|
| 3292 |
+
{
|
| 3293 |
+
"epoch": 5.95,
|
| 3294 |
+
"learning_rate": 0.00040480069512327575,
|
| 3295 |
+
"loss": 2.1319,
|
| 3296 |
+
"step": 54800
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 5.96,
|
| 3300 |
+
"learning_rate": 0.00040371456500488755,
|
| 3301 |
+
"loss": 2.1331,
|
| 3302 |
+
"step": 54900
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 5.97,
|
| 3306 |
+
"learning_rate": 0.0004026284348864994,
|
| 3307 |
+
"loss": 2.1449,
|
| 3308 |
+
"step": 55000
|
| 3309 |
+
},
|
| 3310 |
+
{
|
| 3311 |
+
"epoch": 5.98,
|
| 3312 |
+
"learning_rate": 0.0004015423047681112,
|
| 3313 |
+
"loss": 2.1566,
|
| 3314 |
+
"step": 55100
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 6.0,
|
| 3318 |
+
"learning_rate": 0.00040045617464972306,
|
| 3319 |
+
"loss": 2.1217,
|
| 3320 |
+
"step": 55200
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 6.01,
|
| 3324 |
+
"learning_rate": 0.00039937004453133486,
|
| 3325 |
+
"loss": 2.0642,
|
| 3326 |
+
"step": 55300
|
| 3327 |
+
},
|
| 3328 |
+
{
|
| 3329 |
+
"epoch": 6.02,
|
| 3330 |
+
"learning_rate": 0.0003982839144129467,
|
| 3331 |
+
"loss": 2.043,
|
| 3332 |
+
"step": 55400
|
| 3333 |
+
},
|
| 3334 |
+
{
|
| 3335 |
+
"epoch": 6.03,
|
| 3336 |
+
"learning_rate": 0.0003971977842945585,
|
| 3337 |
+
"loss": 2.0665,
|
| 3338 |
+
"step": 55500
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 6.04,
|
| 3342 |
+
"learning_rate": 0.0003961116541761703,
|
| 3343 |
+
"loss": 2.0171,
|
| 3344 |
+
"step": 55600
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 6.05,
|
| 3348 |
+
"learning_rate": 0.00039502552405778216,
|
| 3349 |
+
"loss": 2.0151,
|
| 3350 |
+
"step": 55700
|
| 3351 |
+
},
|
| 3352 |
+
{
|
| 3353 |
+
"epoch": 6.06,
|
| 3354 |
+
"learning_rate": 0.0003939393939393939,
|
| 3355 |
+
"loss": 2.107,
|
| 3356 |
+
"step": 55800
|
| 3357 |
+
},
|
| 3358 |
+
{
|
| 3359 |
+
"epoch": 6.07,
|
| 3360 |
+
"learning_rate": 0.00039285326382100576,
|
| 3361 |
+
"loss": 1.9811,
|
| 3362 |
+
"step": 55900
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 6.08,
|
| 3366 |
+
"learning_rate": 0.00039176713370261756,
|
| 3367 |
+
"loss": 2.0737,
|
| 3368 |
+
"step": 56000
|
| 3369 |
+
},
|
| 3370 |
+
{
|
| 3371 |
+
"epoch": 6.09,
|
| 3372 |
+
"learning_rate": 0.0003906810035842294,
|
| 3373 |
+
"loss": 2.1152,
|
| 3374 |
+
"step": 56100
|
| 3375 |
+
},
|
| 3376 |
+
{
|
| 3377 |
+
"epoch": 6.1,
|
| 3378 |
+
"learning_rate": 0.0003895948734658412,
|
| 3379 |
+
"loss": 2.0868,
|
| 3380 |
+
"step": 56200
|
| 3381 |
+
},
|
| 3382 |
+
{
|
| 3383 |
+
"epoch": 6.11,
|
| 3384 |
+
"learning_rate": 0.000388508743347453,
|
| 3385 |
+
"loss": 2.0529,
|
| 3386 |
+
"step": 56300
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 6.13,
|
| 3390 |
+
"learning_rate": 0.00038742261322906486,
|
| 3391 |
+
"loss": 2.0739,
|
| 3392 |
+
"step": 56400
|
| 3393 |
+
},
|
| 3394 |
+
{
|
| 3395 |
+
"epoch": 6.14,
|
| 3396 |
+
"learning_rate": 0.00038633648311067666,
|
| 3397 |
+
"loss": 2.0837,
|
| 3398 |
+
"step": 56500
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 6.15,
|
| 3402 |
+
"learning_rate": 0.0003852503529922885,
|
| 3403 |
+
"loss": 2.0741,
|
| 3404 |
+
"step": 56600
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 6.16,
|
| 3408 |
+
"learning_rate": 0.0003841642228739003,
|
| 3409 |
+
"loss": 2.1264,
|
| 3410 |
+
"step": 56700
|
| 3411 |
+
},
|
| 3412 |
+
{
|
| 3413 |
+
"epoch": 6.17,
|
| 3414 |
+
"learning_rate": 0.00038307809275551217,
|
| 3415 |
+
"loss": 2.0742,
|
| 3416 |
+
"step": 56800
|
| 3417 |
+
},
|
| 3418 |
+
{
|
| 3419 |
+
"epoch": 6.18,
|
| 3420 |
+
"learning_rate": 0.0003819919626371239,
|
| 3421 |
+
"loss": 2.0261,
|
| 3422 |
+
"step": 56900
|
| 3423 |
+
},
|
| 3424 |
+
{
|
| 3425 |
+
"epoch": 6.19,
|
| 3426 |
+
"learning_rate": 0.0003809058325187357,
|
| 3427 |
+
"loss": 2.0186,
|
| 3428 |
+
"step": 57000
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 6.2,
|
| 3432 |
+
"learning_rate": 0.00037981970240034756,
|
| 3433 |
+
"loss": 2.0157,
|
| 3434 |
+
"step": 57100
|
| 3435 |
+
},
|
| 3436 |
+
{
|
| 3437 |
+
"epoch": 6.21,
|
| 3438 |
+
"learning_rate": 0.00037873357228195936,
|
| 3439 |
+
"loss": 2.0094,
|
| 3440 |
+
"step": 57200
|
| 3441 |
+
},
|
| 3442 |
+
{
|
| 3443 |
+
"epoch": 6.22,
|
| 3444 |
+
"learning_rate": 0.0003776474421635712,
|
| 3445 |
+
"loss": 2.1602,
|
| 3446 |
+
"step": 57300
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 6.23,
|
| 3450 |
+
"learning_rate": 0.000376561312045183,
|
| 3451 |
+
"loss": 2.0572,
|
| 3452 |
+
"step": 57400
|
| 3453 |
+
},
|
| 3454 |
+
{
|
| 3455 |
+
"epoch": 6.25,
|
| 3456 |
+
"learning_rate": 0.00037547518192679487,
|
| 3457 |
+
"loss": 2.1229,
|
| 3458 |
+
"step": 57500
|
| 3459 |
+
},
|
| 3460 |
+
{
|
| 3461 |
+
"epoch": 6.26,
|
| 3462 |
+
"learning_rate": 0.00037438905180840667,
|
| 3463 |
+
"loss": 1.9764,
|
| 3464 |
+
"step": 57600
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 6.27,
|
| 3468 |
+
"learning_rate": 0.00037330292169001847,
|
| 3469 |
+
"loss": 2.0217,
|
| 3470 |
+
"step": 57700
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 6.28,
|
| 3474 |
+
"learning_rate": 0.0003722167915716303,
|
| 3475 |
+
"loss": 2.0748,
|
| 3476 |
+
"step": 57800
|
| 3477 |
+
},
|
| 3478 |
+
{
|
| 3479 |
+
"epoch": 6.29,
|
| 3480 |
+
"learning_rate": 0.00037113066145324207,
|
| 3481 |
+
"loss": 2.0651,
|
| 3482 |
+
"step": 57900
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 6.3,
|
| 3486 |
+
"learning_rate": 0.0003700445313348539,
|
| 3487 |
+
"loss": 2.1312,
|
| 3488 |
+
"step": 58000
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 6.31,
|
| 3492 |
+
"learning_rate": 0.0003689584012164657,
|
| 3493 |
+
"loss": 2.0224,
|
| 3494 |
+
"step": 58100
|
| 3495 |
+
},
|
| 3496 |
+
{
|
| 3497 |
+
"epoch": 6.32,
|
| 3498 |
+
"learning_rate": 0.00036787227109807757,
|
| 3499 |
+
"loss": 2.0589,
|
| 3500 |
+
"step": 58200
|
| 3501 |
+
},
|
| 3502 |
+
{
|
| 3503 |
+
"epoch": 6.33,
|
| 3504 |
+
"learning_rate": 0.00036678614097968937,
|
| 3505 |
+
"loss": 1.9721,
|
| 3506 |
+
"step": 58300
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 6.34,
|
| 3510 |
+
"learning_rate": 0.00036570001086130117,
|
| 3511 |
+
"loss": 2.0588,
|
| 3512 |
+
"step": 58400
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 6.35,
|
| 3516 |
+
"learning_rate": 0.000364613880742913,
|
| 3517 |
+
"loss": 2.1054,
|
| 3518 |
+
"step": 58500
|
| 3519 |
+
},
|
| 3520 |
+
{
|
| 3521 |
+
"epoch": 6.36,
|
| 3522 |
+
"learning_rate": 0.0003635277506245248,
|
| 3523 |
+
"loss": 2.1048,
|
| 3524 |
+
"step": 58600
|
| 3525 |
+
},
|
| 3526 |
+
{
|
| 3527 |
+
"epoch": 6.38,
|
| 3528 |
+
"learning_rate": 0.0003624416205061367,
|
| 3529 |
+
"loss": 2.0284,
|
| 3530 |
+
"step": 58700
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 6.39,
|
| 3534 |
+
"learning_rate": 0.0003613554903877485,
|
| 3535 |
+
"loss": 2.0858,
|
| 3536 |
+
"step": 58800
|
| 3537 |
+
},
|
| 3538 |
+
{
|
| 3539 |
+
"epoch": 6.4,
|
| 3540 |
+
"learning_rate": 0.0003602693602693603,
|
| 3541 |
+
"loss": 2.0396,
|
| 3542 |
+
"step": 58900
|
| 3543 |
+
},
|
| 3544 |
+
{
|
| 3545 |
+
"epoch": 6.41,
|
| 3546 |
+
"learning_rate": 0.00035918323015097207,
|
| 3547 |
+
"loss": 2.1102,
|
| 3548 |
+
"step": 59000
|
| 3549 |
+
},
|
| 3550 |
+
{
|
| 3551 |
+
"epoch": 6.42,
|
| 3552 |
+
"learning_rate": 0.00035809710003258387,
|
| 3553 |
+
"loss": 2.0202,
|
| 3554 |
+
"step": 59100
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 6.43,
|
| 3558 |
+
"learning_rate": 0.0003570109699141957,
|
| 3559 |
+
"loss": 2.1099,
|
| 3560 |
+
"step": 59200
|
| 3561 |
+
},
|
| 3562 |
+
{
|
| 3563 |
+
"epoch": 6.44,
|
| 3564 |
+
"learning_rate": 0.0003559248397958075,
|
| 3565 |
+
"loss": 2.1964,
|
| 3566 |
+
"step": 59300
|
| 3567 |
+
},
|
| 3568 |
+
{
|
| 3569 |
+
"epoch": 6.45,
|
| 3570 |
+
"learning_rate": 0.0003548387096774194,
|
| 3571 |
+
"loss": 2.0122,
|
| 3572 |
+
"step": 59400
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 6.46,
|
| 3576 |
+
"learning_rate": 0.0003537525795590312,
|
| 3577 |
+
"loss": 2.0572,
|
| 3578 |
+
"step": 59500
|
| 3579 |
+
},
|
| 3580 |
+
{
|
| 3581 |
+
"epoch": 6.47,
|
| 3582 |
+
"learning_rate": 0.00035266644944064303,
|
| 3583 |
+
"loss": 2.0727,
|
| 3584 |
+
"step": 59600
|
| 3585 |
+
},
|
| 3586 |
+
{
|
| 3587 |
+
"epoch": 6.48,
|
| 3588 |
+
"learning_rate": 0.00035158031932225483,
|
| 3589 |
+
"loss": 2.0338,
|
| 3590 |
+
"step": 59700
|
| 3591 |
+
},
|
| 3592 |
+
{
|
| 3593 |
+
"epoch": 6.5,
|
| 3594 |
+
"learning_rate": 0.0003504941892038666,
|
| 3595 |
+
"loss": 2.1002,
|
| 3596 |
+
"step": 59800
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 6.51,
|
| 3600 |
+
"learning_rate": 0.0003494080590854785,
|
| 3601 |
+
"loss": 2.0728,
|
| 3602 |
+
"step": 59900
|
| 3603 |
+
},
|
| 3604 |
+
{
|
| 3605 |
+
"epoch": 6.52,
|
| 3606 |
+
"learning_rate": 0.0003483219289670902,
|
| 3607 |
+
"loss": 2.0609,
|
| 3608 |
+
"step": 60000
|
| 3609 |
+
},
|
| 3610 |
+
{
|
| 3611 |
+
"epoch": 6.53,
|
| 3612 |
+
"learning_rate": 0.0003472357988487021,
|
| 3613 |
+
"loss": 2.0301,
|
| 3614 |
+
"step": 60100
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 6.54,
|
| 3618 |
+
"learning_rate": 0.0003461496687303139,
|
| 3619 |
+
"loss": 2.0571,
|
| 3620 |
+
"step": 60200
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"epoch": 6.55,
|
| 3624 |
+
"learning_rate": 0.00034506353861192573,
|
| 3625 |
+
"loss": 2.1559,
|
| 3626 |
+
"step": 60300
|
| 3627 |
+
},
|
| 3628 |
+
{
|
| 3629 |
+
"epoch": 6.56,
|
| 3630 |
+
"learning_rate": 0.00034397740849353753,
|
| 3631 |
+
"loss": 2.1154,
|
| 3632 |
+
"step": 60400
|
| 3633 |
+
},
|
| 3634 |
+
{
|
| 3635 |
+
"epoch": 6.57,
|
| 3636 |
+
"learning_rate": 0.00034289127837514933,
|
| 3637 |
+
"loss": 2.0395,
|
| 3638 |
+
"step": 60500
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 6.58,
|
| 3642 |
+
"learning_rate": 0.0003418051482567612,
|
| 3643 |
+
"loss": 2.0714,
|
| 3644 |
+
"step": 60600
|
| 3645 |
+
},
|
| 3646 |
+
{
|
| 3647 |
+
"epoch": 6.59,
|
| 3648 |
+
"learning_rate": 0.000340719018138373,
|
| 3649 |
+
"loss": 2.0619,
|
| 3650 |
+
"step": 60700
|
| 3651 |
+
},
|
| 3652 |
+
{
|
| 3653 |
+
"epoch": 6.6,
|
| 3654 |
+
"learning_rate": 0.00033963288801998483,
|
| 3655 |
+
"loss": 2.0613,
|
| 3656 |
+
"step": 60800
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 6.61,
|
| 3660 |
+
"learning_rate": 0.00033854675790159663,
|
| 3661 |
+
"loss": 1.9728,
|
| 3662 |
+
"step": 60900
|
| 3663 |
+
},
|
| 3664 |
+
{
|
| 3665 |
+
"epoch": 6.63,
|
| 3666 |
+
"learning_rate": 0.0003374606277832085,
|
| 3667 |
+
"loss": 2.0779,
|
| 3668 |
+
"step": 61000
|
| 3669 |
+
},
|
| 3670 |
+
{
|
| 3671 |
+
"epoch": 6.64,
|
| 3672 |
+
"learning_rate": 0.00033637449766482023,
|
| 3673 |
+
"loss": 2.0838,
|
| 3674 |
+
"step": 61100
|
| 3675 |
+
},
|
| 3676 |
+
{
|
| 3677 |
+
"epoch": 6.65,
|
| 3678 |
+
"learning_rate": 0.00033528836754643203,
|
| 3679 |
+
"loss": 2.0566,
|
| 3680 |
+
"step": 61200
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 6.66,
|
| 3684 |
+
"learning_rate": 0.0003342022374280439,
|
| 3685 |
+
"loss": 2.0406,
|
| 3686 |
+
"step": 61300
|
| 3687 |
+
},
|
| 3688 |
+
{
|
| 3689 |
+
"epoch": 6.67,
|
| 3690 |
+
"learning_rate": 0.0003331161073096557,
|
| 3691 |
+
"loss": 2.0278,
|
| 3692 |
+
"step": 61400
|
| 3693 |
+
},
|
| 3694 |
+
{
|
| 3695 |
+
"epoch": 6.68,
|
| 3696 |
+
"learning_rate": 0.00033202997719126754,
|
| 3697 |
+
"loss": 2.0856,
|
| 3698 |
+
"step": 61500
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 6.69,
|
| 3702 |
+
"learning_rate": 0.00033094384707287933,
|
| 3703 |
+
"loss": 2.0846,
|
| 3704 |
+
"step": 61600
|
| 3705 |
+
},
|
| 3706 |
+
{
|
| 3707 |
+
"epoch": 6.7,
|
| 3708 |
+
"learning_rate": 0.00032985771695449113,
|
| 3709 |
+
"loss": 2.0389,
|
| 3710 |
+
"step": 61700
|
| 3711 |
+
},
|
| 3712 |
+
{
|
| 3713 |
+
"epoch": 6.71,
|
| 3714 |
+
"learning_rate": 0.000328771586836103,
|
| 3715 |
+
"loss": 2.0874,
|
| 3716 |
+
"step": 61800
|
| 3717 |
+
},
|
| 3718 |
+
{
|
| 3719 |
+
"epoch": 6.72,
|
| 3720 |
+
"learning_rate": 0.0003276854567177148,
|
| 3721 |
+
"loss": 2.0421,
|
| 3722 |
+
"step": 61900
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 6.73,
|
| 3726 |
+
"learning_rate": 0.00032659932659932664,
|
| 3727 |
+
"loss": 2.0272,
|
| 3728 |
+
"step": 62000
|
| 3729 |
+
},
|
| 3730 |
+
{
|
| 3731 |
+
"epoch": 6.74,
|
| 3732 |
+
"learning_rate": 0.00032551319648093844,
|
| 3733 |
+
"loss": 2.0566,
|
| 3734 |
+
"step": 62100
|
| 3735 |
+
},
|
| 3736 |
+
{
|
| 3737 |
+
"epoch": 6.76,
|
| 3738 |
+
"learning_rate": 0.00032442706636255024,
|
| 3739 |
+
"loss": 2.0388,
|
| 3740 |
+
"step": 62200
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 6.77,
|
| 3744 |
+
"learning_rate": 0.00032334093624416204,
|
| 3745 |
+
"loss": 2.0759,
|
| 3746 |
+
"step": 62300
|
| 3747 |
+
},
|
| 3748 |
+
{
|
| 3749 |
+
"epoch": 6.78,
|
| 3750 |
+
"learning_rate": 0.00032225480612577383,
|
| 3751 |
+
"loss": 2.1224,
|
| 3752 |
+
"step": 62400
|
| 3753 |
+
},
|
| 3754 |
+
{
|
| 3755 |
+
"epoch": 6.79,
|
| 3756 |
+
"learning_rate": 0.0003211686760073857,
|
| 3757 |
+
"loss": 2.0217,
|
| 3758 |
+
"step": 62500
|
| 3759 |
}
|
| 3760 |
],
|
| 3761 |
"max_steps": 92070,
|
| 3762 |
"num_train_epochs": 10,
|
| 3763 |
+
"total_flos": 1.2817136295965491e+17,
|
| 3764 |
"trial_name": null,
|
| 3765 |
"trial_params": null
|
| 3766 |
}
|