Training in progress, step 900, checkpoint
Browse files- last-checkpoint/optimizer_0/.metadata +0 -0
- last-checkpoint/optimizer_0/__0_0.distcp +1 -1
- last-checkpoint/optimizer_0/__1_0.distcp +1 -1
- last-checkpoint/optimizer_0/__2_0.distcp +1 -1
- last-checkpoint/optimizer_0/__3_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/.metadata +0 -0
- last-checkpoint/pytorch_model_fsdp_0/__0_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__1_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__2_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__3_0.distcp +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +763 -3
last-checkpoint/optimizer_0/.metadata
CHANGED
|
Binary files a/last-checkpoint/optimizer_0/.metadata and b/last-checkpoint/optimizer_0/.metadata differ
|
|
|
last-checkpoint/optimizer_0/__0_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13934748
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:751cf08e904d158afe5d5c376833ae0e82507272c605040cee2892e5b08babd7
|
| 3 |
size 13934748
|
last-checkpoint/optimizer_0/__1_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13999412
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06a90cd633a16bc937002e25bf17189511b59b8e8d6bd000662a0b07bbf80e65
|
| 3 |
size 13999412
|
last-checkpoint/optimizer_0/__2_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13990904
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a2a1314f17bbfab7f1e0d63c5f4ad16ed3c81d2546ede535c552f819fd2b3ab
|
| 3 |
size 13990904
|
last-checkpoint/optimizer_0/__3_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13990904
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61ce145ff3942d4b7afa3864d5060ecb372e5343ad7abd8681f9ef04e3996a27
|
| 3 |
size 13990904
|
last-checkpoint/pytorch_model_fsdp_0/.metadata
CHANGED
|
Binary files a/last-checkpoint/pytorch_model_fsdp_0/.metadata and b/last-checkpoint/pytorch_model_fsdp_0/.metadata differ
|
|
|
last-checkpoint/pytorch_model_fsdp_0/__0_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef11d2a12800199b371850d31a6f25caf9ed6263ace1a113ebdb48708b504181
|
| 3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__1_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:146b0c5dbeeaf44008b0996c6e5385da4bd543d15c17518bdbadad49738050cf
|
| 3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__2_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ce829e48185824cbdd9332cfc8db3ea7a52a07e0542cfb9ceb1a123cf238c986
|
| 3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__3_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ebc6b5b3bf3ff4db9103e1f6c2a534b04905118e03682dcfd5b2651cf5023a9
|
| 3 |
size 6966784
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15088
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:de551c8b1c7165a88a996c1e2c9ccef2b9c716ed5dd27dbead3aed497f875baf
|
| 3 |
size 15088
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15088
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d7ac6617d0f4bedb4a49c51f3499536a49a714fbf47c2bf0f3ef31d2fc421c6
|
| 3 |
size 15088
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15088
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:52e8353d56a8d408800aeacfb35920a2641f5b231344cfc6e9685c5fff1b02a0
|
| 3 |
size 15088
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15088
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e436a7cd1ba66004593b56f4b7ea09bd56b9555226a359189f1b562635d9c755
|
| 3 |
size 15088
|
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 sha256:7754cc7953f3b6df72a3313e3aea13311d99a52571c658951fa15bc741d1fcff
|
| 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": 20,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -6099,6 +6099,766 @@
|
|
| 6099 |
"eval_samples_per_second": 5.538,
|
| 6100 |
"eval_steps_per_second": 0.19,
|
| 6101 |
"step": 800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6102 |
}
|
| 6103 |
],
|
| 6104 |
"logging_steps": 1,
|
|
@@ -6118,7 +6878,7 @@
|
|
| 6118 |
"attributes": {}
|
| 6119 |
}
|
| 6120 |
},
|
| 6121 |
-
"total_flos": 2.
|
| 6122 |
"train_batch_size": 8,
|
| 6123 |
"trial_name": null,
|
| 6124 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7627118644067796,
|
| 5 |
"eval_steps": 20,
|
| 6 |
+
"global_step": 900,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 6099 |
"eval_samples_per_second": 5.538,
|
| 6100 |
"eval_steps_per_second": 0.19,
|
| 6101 |
"step": 800
|
| 6102 |
+
},
|
| 6103 |
+
{
|
| 6104 |
+
"epoch": 0.6788135593220339,
|
| 6105 |
+
"grad_norm": 0.8634885549545288,
|
| 6106 |
+
"learning_rate": 5.6535492087643904e-06,
|
| 6107 |
+
"loss": 0.0026,
|
| 6108 |
+
"step": 801
|
| 6109 |
+
},
|
| 6110 |
+
{
|
| 6111 |
+
"epoch": 0.6796610169491526,
|
| 6112 |
+
"grad_norm": 2.032689094543457,
|
| 6113 |
+
"learning_rate": 5.626926795411447e-06,
|
| 6114 |
+
"loss": 0.01,
|
| 6115 |
+
"step": 802
|
| 6116 |
+
},
|
| 6117 |
+
{
|
| 6118 |
+
"epoch": 0.6805084745762712,
|
| 6119 |
+
"grad_norm": 0.10548463463783264,
|
| 6120 |
+
"learning_rate": 5.6003426501798665e-06,
|
| 6121 |
+
"loss": 0.0005,
|
| 6122 |
+
"step": 803
|
| 6123 |
+
},
|
| 6124 |
+
{
|
| 6125 |
+
"epoch": 0.6813559322033899,
|
| 6126 |
+
"grad_norm": 1.6854254007339478,
|
| 6127 |
+
"learning_rate": 5.573797005703601e-06,
|
| 6128 |
+
"loss": 0.0037,
|
| 6129 |
+
"step": 804
|
| 6130 |
+
},
|
| 6131 |
+
{
|
| 6132 |
+
"epoch": 0.6822033898305084,
|
| 6133 |
+
"grad_norm": 0.280429869890213,
|
| 6134 |
+
"learning_rate": 5.547290094279687e-06,
|
| 6135 |
+
"loss": 0.001,
|
| 6136 |
+
"step": 805
|
| 6137 |
+
},
|
| 6138 |
+
{
|
| 6139 |
+
"epoch": 0.6830508474576271,
|
| 6140 |
+
"grad_norm": 0.711410403251648,
|
| 6141 |
+
"learning_rate": 5.520822147866214e-06,
|
| 6142 |
+
"loss": 0.0033,
|
| 6143 |
+
"step": 806
|
| 6144 |
+
},
|
| 6145 |
+
{
|
| 6146 |
+
"epoch": 0.6838983050847458,
|
| 6147 |
+
"grad_norm": 0.2613879144191742,
|
| 6148 |
+
"learning_rate": 5.494393398080292e-06,
|
| 6149 |
+
"loss": 0.0011,
|
| 6150 |
+
"step": 807
|
| 6151 |
+
},
|
| 6152 |
+
{
|
| 6153 |
+
"epoch": 0.6847457627118644,
|
| 6154 |
+
"grad_norm": 0.8584437966346741,
|
| 6155 |
+
"learning_rate": 5.46800407619603e-06,
|
| 6156 |
+
"loss": 0.0023,
|
| 6157 |
+
"step": 808
|
| 6158 |
+
},
|
| 6159 |
+
{
|
| 6160 |
+
"epoch": 0.6855932203389831,
|
| 6161 |
+
"grad_norm": 3.826030731201172,
|
| 6162 |
+
"learning_rate": 5.441654413142508e-06,
|
| 6163 |
+
"loss": 0.0228,
|
| 6164 |
+
"step": 809
|
| 6165 |
+
},
|
| 6166 |
+
{
|
| 6167 |
+
"epoch": 0.6864406779661016,
|
| 6168 |
+
"grad_norm": 0.7557411193847656,
|
| 6169 |
+
"learning_rate": 5.415344639501754e-06,
|
| 6170 |
+
"loss": 0.0033,
|
| 6171 |
+
"step": 810
|
| 6172 |
+
},
|
| 6173 |
+
{
|
| 6174 |
+
"epoch": 0.6872881355932203,
|
| 6175 |
+
"grad_norm": 1.0392420291900635,
|
| 6176 |
+
"learning_rate": 5.38907498550674e-06,
|
| 6177 |
+
"loss": 0.0043,
|
| 6178 |
+
"step": 811
|
| 6179 |
+
},
|
| 6180 |
+
{
|
| 6181 |
+
"epoch": 0.688135593220339,
|
| 6182 |
+
"grad_norm": 2.502735137939453,
|
| 6183 |
+
"learning_rate": 5.362845681039348e-06,
|
| 6184 |
+
"loss": 0.0147,
|
| 6185 |
+
"step": 812
|
| 6186 |
+
},
|
| 6187 |
+
{
|
| 6188 |
+
"epoch": 0.6889830508474576,
|
| 6189 |
+
"grad_norm": 0.42259418964385986,
|
| 6190 |
+
"learning_rate": 5.336656955628371e-06,
|
| 6191 |
+
"loss": 0.0021,
|
| 6192 |
+
"step": 813
|
| 6193 |
+
},
|
| 6194 |
+
{
|
| 6195 |
+
"epoch": 0.6898305084745763,
|
| 6196 |
+
"grad_norm": 2.6665008068084717,
|
| 6197 |
+
"learning_rate": 5.310509038447492e-06,
|
| 6198 |
+
"loss": 0.0194,
|
| 6199 |
+
"step": 814
|
| 6200 |
+
},
|
| 6201 |
+
{
|
| 6202 |
+
"epoch": 0.690677966101695,
|
| 6203 |
+
"grad_norm": 2.9882800579071045,
|
| 6204 |
+
"learning_rate": 5.284402158313299e-06,
|
| 6205 |
+
"loss": 0.0415,
|
| 6206 |
+
"step": 815
|
| 6207 |
+
},
|
| 6208 |
+
{
|
| 6209 |
+
"epoch": 0.6915254237288135,
|
| 6210 |
+
"grad_norm": 0.5250701904296875,
|
| 6211 |
+
"learning_rate": 5.25833654368326e-06,
|
| 6212 |
+
"loss": 0.0027,
|
| 6213 |
+
"step": 816
|
| 6214 |
+
},
|
| 6215 |
+
{
|
| 6216 |
+
"epoch": 0.6923728813559322,
|
| 6217 |
+
"grad_norm": 0.46300551295280457,
|
| 6218 |
+
"learning_rate": 5.23231242265375e-06,
|
| 6219 |
+
"loss": 0.0013,
|
| 6220 |
+
"step": 817
|
| 6221 |
+
},
|
| 6222 |
+
{
|
| 6223 |
+
"epoch": 0.6932203389830508,
|
| 6224 |
+
"grad_norm": 0.7175685167312622,
|
| 6225 |
+
"learning_rate": 5.206330022958024e-06,
|
| 6226 |
+
"loss": 0.0037,
|
| 6227 |
+
"step": 818
|
| 6228 |
+
},
|
| 6229 |
+
{
|
| 6230 |
+
"epoch": 0.6940677966101695,
|
| 6231 |
+
"grad_norm": 0.36853301525115967,
|
| 6232 |
+
"learning_rate": 5.180389571964249e-06,
|
| 6233 |
+
"loss": 0.0016,
|
| 6234 |
+
"step": 819
|
| 6235 |
+
},
|
| 6236 |
+
{
|
| 6237 |
+
"epoch": 0.6949152542372882,
|
| 6238 |
+
"grad_norm": 0.713392436504364,
|
| 6239 |
+
"learning_rate": 5.1544912966735e-06,
|
| 6240 |
+
"loss": 0.0048,
|
| 6241 |
+
"step": 820
|
| 6242 |
+
},
|
| 6243 |
+
{
|
| 6244 |
+
"epoch": 0.6949152542372882,
|
| 6245 |
+
"eval_accuracy": 1.0,
|
| 6246 |
+
"eval_f1": 1.0,
|
| 6247 |
+
"eval_loss": 9.25260319490917e-05,
|
| 6248 |
+
"eval_precision": 1.0,
|
| 6249 |
+
"eval_recall": 1.0,
|
| 6250 |
+
"eval_runtime": 51.5622,
|
| 6251 |
+
"eval_samples_per_second": 5.663,
|
| 6252 |
+
"eval_steps_per_second": 0.194,
|
| 6253 |
+
"step": 820
|
| 6254 |
+
},
|
| 6255 |
+
{
|
| 6256 |
+
"epoch": 0.6957627118644067,
|
| 6257 |
+
"grad_norm": 1.1480737924575806,
|
| 6258 |
+
"learning_rate": 5.1286354237177806e-06,
|
| 6259 |
+
"loss": 0.0071,
|
| 6260 |
+
"step": 821
|
| 6261 |
+
},
|
| 6262 |
+
{
|
| 6263 |
+
"epoch": 0.6966101694915254,
|
| 6264 |
+
"grad_norm": 1.0792937278747559,
|
| 6265 |
+
"learning_rate": 5.102822179358037e-06,
|
| 6266 |
+
"loss": 0.0062,
|
| 6267 |
+
"step": 822
|
| 6268 |
+
},
|
| 6269 |
+
{
|
| 6270 |
+
"epoch": 0.6974576271186441,
|
| 6271 |
+
"grad_norm": 2.4379987716674805,
|
| 6272 |
+
"learning_rate": 5.0770517894821815e-06,
|
| 6273 |
+
"loss": 0.0301,
|
| 6274 |
+
"step": 823
|
| 6275 |
+
},
|
| 6276 |
+
{
|
| 6277 |
+
"epoch": 0.6983050847457627,
|
| 6278 |
+
"grad_norm": 1.457727313041687,
|
| 6279 |
+
"learning_rate": 5.051324479603106e-06,
|
| 6280 |
+
"loss": 0.0199,
|
| 6281 |
+
"step": 824
|
| 6282 |
+
},
|
| 6283 |
+
{
|
| 6284 |
+
"epoch": 0.6991525423728814,
|
| 6285 |
+
"grad_norm": 0.9668751955032349,
|
| 6286 |
+
"learning_rate": 5.025640474856732e-06,
|
| 6287 |
+
"loss": 0.003,
|
| 6288 |
+
"step": 825
|
| 6289 |
+
},
|
| 6290 |
+
{
|
| 6291 |
+
"epoch": 0.7,
|
| 6292 |
+
"grad_norm": 0.48410564661026,
|
| 6293 |
+
"learning_rate": 5.000000000000003e-06,
|
| 6294 |
+
"loss": 0.0013,
|
| 6295 |
+
"step": 826
|
| 6296 |
+
},
|
| 6297 |
+
{
|
| 6298 |
+
"epoch": 0.7008474576271186,
|
| 6299 |
+
"grad_norm": 0.28067731857299805,
|
| 6300 |
+
"learning_rate": 4.97440327940895e-06,
|
| 6301 |
+
"loss": 0.0012,
|
| 6302 |
+
"step": 827
|
| 6303 |
+
},
|
| 6304 |
+
{
|
| 6305 |
+
"epoch": 0.7016949152542373,
|
| 6306 |
+
"grad_norm": 0.2483496367931366,
|
| 6307 |
+
"learning_rate": 4.948850537076719e-06,
|
| 6308 |
+
"loss": 0.0007,
|
| 6309 |
+
"step": 828
|
| 6310 |
+
},
|
| 6311 |
+
{
|
| 6312 |
+
"epoch": 0.7025423728813559,
|
| 6313 |
+
"grad_norm": 1.3155626058578491,
|
| 6314 |
+
"learning_rate": 4.923341996611604e-06,
|
| 6315 |
+
"loss": 0.0077,
|
| 6316 |
+
"step": 829
|
| 6317 |
+
},
|
| 6318 |
+
{
|
| 6319 |
+
"epoch": 0.7033898305084746,
|
| 6320 |
+
"grad_norm": 0.7930201888084412,
|
| 6321 |
+
"learning_rate": 4.897877881235091e-06,
|
| 6322 |
+
"loss": 0.0027,
|
| 6323 |
+
"step": 830
|
| 6324 |
+
},
|
| 6325 |
+
{
|
| 6326 |
+
"epoch": 0.7042372881355933,
|
| 6327 |
+
"grad_norm": 0.8325580954551697,
|
| 6328 |
+
"learning_rate": 4.8724584137799235e-06,
|
| 6329 |
+
"loss": 0.0081,
|
| 6330 |
+
"step": 831
|
| 6331 |
+
},
|
| 6332 |
+
{
|
| 6333 |
+
"epoch": 0.7050847457627119,
|
| 6334 |
+
"grad_norm": 2.1520488262176514,
|
| 6335 |
+
"learning_rate": 4.847083816688123e-06,
|
| 6336 |
+
"loss": 0.0172,
|
| 6337 |
+
"step": 832
|
| 6338 |
+
},
|
| 6339 |
+
{
|
| 6340 |
+
"epoch": 0.7059322033898305,
|
| 6341 |
+
"grad_norm": 0.5673890709877014,
|
| 6342 |
+
"learning_rate": 4.821754312009063e-06,
|
| 6343 |
+
"loss": 0.0019,
|
| 6344 |
+
"step": 833
|
| 6345 |
+
},
|
| 6346 |
+
{
|
| 6347 |
+
"epoch": 0.7067796610169491,
|
| 6348 |
+
"grad_norm": 2.460827350616455,
|
| 6349 |
+
"learning_rate": 4.796470121397518e-06,
|
| 6350 |
+
"loss": 0.0238,
|
| 6351 |
+
"step": 834
|
| 6352 |
+
},
|
| 6353 |
+
{
|
| 6354 |
+
"epoch": 0.7076271186440678,
|
| 6355 |
+
"grad_norm": 1.8376518487930298,
|
| 6356 |
+
"learning_rate": 4.771231466111725e-06,
|
| 6357 |
+
"loss": 0.0136,
|
| 6358 |
+
"step": 835
|
| 6359 |
+
},
|
| 6360 |
+
{
|
| 6361 |
+
"epoch": 0.7084745762711865,
|
| 6362 |
+
"grad_norm": 1.0489223003387451,
|
| 6363 |
+
"learning_rate": 4.746038567011454e-06,
|
| 6364 |
+
"loss": 0.0035,
|
| 6365 |
+
"step": 836
|
| 6366 |
+
},
|
| 6367 |
+
{
|
| 6368 |
+
"epoch": 0.7093220338983051,
|
| 6369 |
+
"grad_norm": 2.19427227973938,
|
| 6370 |
+
"learning_rate": 4.7208916445560625e-06,
|
| 6371 |
+
"loss": 0.0212,
|
| 6372 |
+
"step": 837
|
| 6373 |
+
},
|
| 6374 |
+
{
|
| 6375 |
+
"epoch": 0.7101694915254237,
|
| 6376 |
+
"grad_norm": 2.033280849456787,
|
| 6377 |
+
"learning_rate": 4.695790918802577e-06,
|
| 6378 |
+
"loss": 0.0033,
|
| 6379 |
+
"step": 838
|
| 6380 |
+
},
|
| 6381 |
+
{
|
| 6382 |
+
"epoch": 0.7110169491525423,
|
| 6383 |
+
"grad_norm": 1.6251804828643799,
|
| 6384 |
+
"learning_rate": 4.67073660940376e-06,
|
| 6385 |
+
"loss": 0.0187,
|
| 6386 |
+
"step": 839
|
| 6387 |
+
},
|
| 6388 |
+
{
|
| 6389 |
+
"epoch": 0.711864406779661,
|
| 6390 |
+
"grad_norm": 0.2912212610244751,
|
| 6391 |
+
"learning_rate": 4.645728935606194e-06,
|
| 6392 |
+
"loss": 0.0007,
|
| 6393 |
+
"step": 840
|
| 6394 |
+
},
|
| 6395 |
+
{
|
| 6396 |
+
"epoch": 0.711864406779661,
|
| 6397 |
+
"eval_accuracy": 1.0,
|
| 6398 |
+
"eval_f1": 1.0,
|
| 6399 |
+
"eval_loss": 0.00012361357221379876,
|
| 6400 |
+
"eval_precision": 1.0,
|
| 6401 |
+
"eval_recall": 1.0,
|
| 6402 |
+
"eval_runtime": 51.0511,
|
| 6403 |
+
"eval_samples_per_second": 5.72,
|
| 6404 |
+
"eval_steps_per_second": 0.196,
|
| 6405 |
+
"step": 840
|
| 6406 |
+
},
|
| 6407 |
+
{
|
| 6408 |
+
"epoch": 0.7127118644067797,
|
| 6409 |
+
"grad_norm": 3.8638107776641846,
|
| 6410 |
+
"learning_rate": 4.620768116248362e-06,
|
| 6411 |
+
"loss": 0.0192,
|
| 6412 |
+
"step": 841
|
| 6413 |
+
},
|
| 6414 |
+
{
|
| 6415 |
+
"epoch": 0.7135593220338983,
|
| 6416 |
+
"grad_norm": 3.521618366241455,
|
| 6417 |
+
"learning_rate": 4.595854369758727e-06,
|
| 6418 |
+
"loss": 0.0239,
|
| 6419 |
+
"step": 842
|
| 6420 |
+
},
|
| 6421 |
+
{
|
| 6422 |
+
"epoch": 0.714406779661017,
|
| 6423 |
+
"grad_norm": 4.152134418487549,
|
| 6424 |
+
"learning_rate": 4.570987914153824e-06,
|
| 6425 |
+
"loss": 0.0344,
|
| 6426 |
+
"step": 843
|
| 6427 |
+
},
|
| 6428 |
+
{
|
| 6429 |
+
"epoch": 0.7152542372881356,
|
| 6430 |
+
"grad_norm": 1.025924801826477,
|
| 6431 |
+
"learning_rate": 4.546168967036352e-06,
|
| 6432 |
+
"loss": 0.0037,
|
| 6433 |
+
"step": 844
|
| 6434 |
+
},
|
| 6435 |
+
{
|
| 6436 |
+
"epoch": 0.7161016949152542,
|
| 6437 |
+
"grad_norm": 1.601256012916565,
|
| 6438 |
+
"learning_rate": 4.521397745593279e-06,
|
| 6439 |
+
"loss": 0.008,
|
| 6440 |
+
"step": 845
|
| 6441 |
+
},
|
| 6442 |
+
{
|
| 6443 |
+
"epoch": 0.7169491525423729,
|
| 6444 |
+
"grad_norm": 1.8692275285720825,
|
| 6445 |
+
"learning_rate": 4.4966744665939226e-06,
|
| 6446 |
+
"loss": 0.0072,
|
| 6447 |
+
"step": 846
|
| 6448 |
+
},
|
| 6449 |
+
{
|
| 6450 |
+
"epoch": 0.7177966101694915,
|
| 6451 |
+
"grad_norm": 2.0935654640197754,
|
| 6452 |
+
"learning_rate": 4.4719993463880695e-06,
|
| 6453 |
+
"loss": 0.0217,
|
| 6454 |
+
"step": 847
|
| 6455 |
+
},
|
| 6456 |
+
{
|
| 6457 |
+
"epoch": 0.7186440677966102,
|
| 6458 |
+
"grad_norm": 1.089236855506897,
|
| 6459 |
+
"learning_rate": 4.447372600904065e-06,
|
| 6460 |
+
"loss": 0.0095,
|
| 6461 |
+
"step": 848
|
| 6462 |
+
},
|
| 6463 |
+
{
|
| 6464 |
+
"epoch": 0.7194915254237289,
|
| 6465 |
+
"grad_norm": 1.3698294162750244,
|
| 6466 |
+
"learning_rate": 4.422794445646947e-06,
|
| 6467 |
+
"loss": 0.0075,
|
| 6468 |
+
"step": 849
|
| 6469 |
+
},
|
| 6470 |
+
{
|
| 6471 |
+
"epoch": 0.7203389830508474,
|
| 6472 |
+
"grad_norm": 1.4754582643508911,
|
| 6473 |
+
"learning_rate": 4.398265095696539e-06,
|
| 6474 |
+
"loss": 0.012,
|
| 6475 |
+
"step": 850
|
| 6476 |
+
},
|
| 6477 |
+
{
|
| 6478 |
+
"epoch": 0.7211864406779661,
|
| 6479 |
+
"grad_norm": 0.5242229104042053,
|
| 6480 |
+
"learning_rate": 4.373784765705586e-06,
|
| 6481 |
+
"loss": 0.0038,
|
| 6482 |
+
"step": 851
|
| 6483 |
+
},
|
| 6484 |
+
{
|
| 6485 |
+
"epoch": 0.7220338983050848,
|
| 6486 |
+
"grad_norm": 3.324636220932007,
|
| 6487 |
+
"learning_rate": 4.349353669897856e-06,
|
| 6488 |
+
"loss": 0.0299,
|
| 6489 |
+
"step": 852
|
| 6490 |
+
},
|
| 6491 |
+
{
|
| 6492 |
+
"epoch": 0.7228813559322034,
|
| 6493 |
+
"grad_norm": 1.4912688732147217,
|
| 6494 |
+
"learning_rate": 4.324972022066284e-06,
|
| 6495 |
+
"loss": 0.009,
|
| 6496 |
+
"step": 853
|
| 6497 |
+
},
|
| 6498 |
+
{
|
| 6499 |
+
"epoch": 0.7237288135593221,
|
| 6500 |
+
"grad_norm": 1.8088048696517944,
|
| 6501 |
+
"learning_rate": 4.300640035571085e-06,
|
| 6502 |
+
"loss": 0.0156,
|
| 6503 |
+
"step": 854
|
| 6504 |
+
},
|
| 6505 |
+
{
|
| 6506 |
+
"epoch": 0.7245762711864406,
|
| 6507 |
+
"grad_norm": 0.9810622930526733,
|
| 6508 |
+
"learning_rate": 4.276357923337903e-06,
|
| 6509 |
+
"loss": 0.0054,
|
| 6510 |
+
"step": 855
|
| 6511 |
+
},
|
| 6512 |
+
{
|
| 6513 |
+
"epoch": 0.7254237288135593,
|
| 6514 |
+
"grad_norm": 1.399301290512085,
|
| 6515 |
+
"learning_rate": 4.2521258978559324e-06,
|
| 6516 |
+
"loss": 0.0144,
|
| 6517 |
+
"step": 856
|
| 6518 |
+
},
|
| 6519 |
+
{
|
| 6520 |
+
"epoch": 0.726271186440678,
|
| 6521 |
+
"grad_norm": 1.4187586307525635,
|
| 6522 |
+
"learning_rate": 4.227944171176072e-06,
|
| 6523 |
+
"loss": 0.0111,
|
| 6524 |
+
"step": 857
|
| 6525 |
+
},
|
| 6526 |
+
{
|
| 6527 |
+
"epoch": 0.7271186440677966,
|
| 6528 |
+
"grad_norm": 1.7159850597381592,
|
| 6529 |
+
"learning_rate": 4.203812954909057e-06,
|
| 6530 |
+
"loss": 0.0124,
|
| 6531 |
+
"step": 858
|
| 6532 |
+
},
|
| 6533 |
+
{
|
| 6534 |
+
"epoch": 0.7279661016949153,
|
| 6535 |
+
"grad_norm": 1.37935471534729,
|
| 6536 |
+
"learning_rate": 4.179732460223626e-06,
|
| 6537 |
+
"loss": 0.0118,
|
| 6538 |
+
"step": 859
|
| 6539 |
+
},
|
| 6540 |
+
{
|
| 6541 |
+
"epoch": 0.7288135593220338,
|
| 6542 |
+
"grad_norm": 0.29636350274086,
|
| 6543 |
+
"learning_rate": 4.1557028978446415e-06,
|
| 6544 |
+
"loss": 0.0014,
|
| 6545 |
+
"step": 860
|
| 6546 |
+
},
|
| 6547 |
+
{
|
| 6548 |
+
"epoch": 0.7288135593220338,
|
| 6549 |
+
"eval_accuracy": 1.0,
|
| 6550 |
+
"eval_f1": 1.0,
|
| 6551 |
+
"eval_loss": 0.00012519690790213645,
|
| 6552 |
+
"eval_precision": 1.0,
|
| 6553 |
+
"eval_recall": 1.0,
|
| 6554 |
+
"eval_runtime": 50.6172,
|
| 6555 |
+
"eval_samples_per_second": 5.769,
|
| 6556 |
+
"eval_steps_per_second": 0.198,
|
| 6557 |
+
"step": 860
|
| 6558 |
+
},
|
| 6559 |
+
{
|
| 6560 |
+
"epoch": 0.7296610169491525,
|
| 6561 |
+
"grad_norm": 0.8602759838104248,
|
| 6562 |
+
"learning_rate": 4.1317244780512775e-06,
|
| 6563 |
+
"loss": 0.0092,
|
| 6564 |
+
"step": 861
|
| 6565 |
+
},
|
| 6566 |
+
{
|
| 6567 |
+
"epoch": 0.7305084745762712,
|
| 6568 |
+
"grad_norm": 2.8475470542907715,
|
| 6569 |
+
"learning_rate": 4.107797410675166e-06,
|
| 6570 |
+
"loss": 0.0388,
|
| 6571 |
+
"step": 862
|
| 6572 |
+
},
|
| 6573 |
+
{
|
| 6574 |
+
"epoch": 0.7313559322033898,
|
| 6575 |
+
"grad_norm": 1.3523766994476318,
|
| 6576 |
+
"learning_rate": 4.083921905098559e-06,
|
| 6577 |
+
"loss": 0.0068,
|
| 6578 |
+
"step": 863
|
| 6579 |
+
},
|
| 6580 |
+
{
|
| 6581 |
+
"epoch": 0.7322033898305085,
|
| 6582 |
+
"grad_norm": 1.5425556898117065,
|
| 6583 |
+
"learning_rate": 4.060098170252495e-06,
|
| 6584 |
+
"loss": 0.0116,
|
| 6585 |
+
"step": 864
|
| 6586 |
+
},
|
| 6587 |
+
{
|
| 6588 |
+
"epoch": 0.7330508474576272,
|
| 6589 |
+
"grad_norm": 0.3472643494606018,
|
| 6590 |
+
"learning_rate": 4.036326414614985e-06,
|
| 6591 |
+
"loss": 0.0016,
|
| 6592 |
+
"step": 865
|
| 6593 |
+
},
|
| 6594 |
+
{
|
| 6595 |
+
"epoch": 0.7338983050847457,
|
| 6596 |
+
"grad_norm": 0.4821791648864746,
|
| 6597 |
+
"learning_rate": 4.0126068462091705e-06,
|
| 6598 |
+
"loss": 0.0022,
|
| 6599 |
+
"step": 866
|
| 6600 |
+
},
|
| 6601 |
+
{
|
| 6602 |
+
"epoch": 0.7347457627118644,
|
| 6603 |
+
"grad_norm": 0.49641525745391846,
|
| 6604 |
+
"learning_rate": 3.988939672601509e-06,
|
| 6605 |
+
"loss": 0.0018,
|
| 6606 |
+
"step": 867
|
| 6607 |
+
},
|
| 6608 |
+
{
|
| 6609 |
+
"epoch": 0.735593220338983,
|
| 6610 |
+
"grad_norm": 1.2695690393447876,
|
| 6611 |
+
"learning_rate": 3.9653251008999615e-06,
|
| 6612 |
+
"loss": 0.0189,
|
| 6613 |
+
"step": 868
|
| 6614 |
+
},
|
| 6615 |
+
{
|
| 6616 |
+
"epoch": 0.7364406779661017,
|
| 6617 |
+
"grad_norm": 0.7525699734687805,
|
| 6618 |
+
"learning_rate": 3.941763337752177e-06,
|
| 6619 |
+
"loss": 0.0027,
|
| 6620 |
+
"step": 869
|
| 6621 |
+
},
|
| 6622 |
+
{
|
| 6623 |
+
"epoch": 0.7372881355932204,
|
| 6624 |
+
"grad_norm": 1.9671075344085693,
|
| 6625 |
+
"learning_rate": 3.918254589343683e-06,
|
| 6626 |
+
"loss": 0.0161,
|
| 6627 |
+
"step": 870
|
| 6628 |
+
},
|
| 6629 |
+
{
|
| 6630 |
+
"epoch": 0.738135593220339,
|
| 6631 |
+
"grad_norm": 0.5608029365539551,
|
| 6632 |
+
"learning_rate": 3.894799061396086e-06,
|
| 6633 |
+
"loss": 0.002,
|
| 6634 |
+
"step": 871
|
| 6635 |
+
},
|
| 6636 |
+
{
|
| 6637 |
+
"epoch": 0.7389830508474576,
|
| 6638 |
+
"grad_norm": 2.967503309249878,
|
| 6639 |
+
"learning_rate": 3.871396959165267e-06,
|
| 6640 |
+
"loss": 0.0201,
|
| 6641 |
+
"step": 872
|
| 6642 |
+
},
|
| 6643 |
+
{
|
| 6644 |
+
"epoch": 0.7398305084745763,
|
| 6645 |
+
"grad_norm": 0.9887444972991943,
|
| 6646 |
+
"learning_rate": 3.848048487439583e-06,
|
| 6647 |
+
"loss": 0.0059,
|
| 6648 |
+
"step": 873
|
| 6649 |
+
},
|
| 6650 |
+
{
|
| 6651 |
+
"epoch": 0.7406779661016949,
|
| 6652 |
+
"grad_norm": 1.1107892990112305,
|
| 6653 |
+
"learning_rate": 3.824753850538082e-06,
|
| 6654 |
+
"loss": 0.0071,
|
| 6655 |
+
"step": 874
|
| 6656 |
+
},
|
| 6657 |
+
{
|
| 6658 |
+
"epoch": 0.7415254237288136,
|
| 6659 |
+
"grad_norm": 0.9307472705841064,
|
| 6660 |
+
"learning_rate": 3.801513252308712e-06,
|
| 6661 |
+
"loss": 0.0042,
|
| 6662 |
+
"step": 875
|
| 6663 |
+
},
|
| 6664 |
+
{
|
| 6665 |
+
"epoch": 0.7423728813559322,
|
| 6666 |
+
"grad_norm": 2.223409652709961,
|
| 6667 |
+
"learning_rate": 3.778326896126533e-06,
|
| 6668 |
+
"loss": 0.0084,
|
| 6669 |
+
"step": 876
|
| 6670 |
+
},
|
| 6671 |
+
{
|
| 6672 |
+
"epoch": 0.7432203389830508,
|
| 6673 |
+
"grad_norm": 0.8251433968544006,
|
| 6674 |
+
"learning_rate": 3.755194984891943e-06,
|
| 6675 |
+
"loss": 0.0042,
|
| 6676 |
+
"step": 877
|
| 6677 |
+
},
|
| 6678 |
+
{
|
| 6679 |
+
"epoch": 0.7440677966101695,
|
| 6680 |
+
"grad_norm": 1.5690832138061523,
|
| 6681 |
+
"learning_rate": 3.7321177210288974e-06,
|
| 6682 |
+
"loss": 0.0045,
|
| 6683 |
+
"step": 878
|
| 6684 |
+
},
|
| 6685 |
+
{
|
| 6686 |
+
"epoch": 0.7449152542372881,
|
| 6687 |
+
"grad_norm": 1.1555606126785278,
|
| 6688 |
+
"learning_rate": 3.709095306483149e-06,
|
| 6689 |
+
"loss": 0.009,
|
| 6690 |
+
"step": 879
|
| 6691 |
+
},
|
| 6692 |
+
{
|
| 6693 |
+
"epoch": 0.7457627118644068,
|
| 6694 |
+
"grad_norm": 0.7731421589851379,
|
| 6695 |
+
"learning_rate": 3.6861279427204634e-06,
|
| 6696 |
+
"loss": 0.0056,
|
| 6697 |
+
"step": 880
|
| 6698 |
+
},
|
| 6699 |
+
{
|
| 6700 |
+
"epoch": 0.7457627118644068,
|
| 6701 |
+
"eval_accuracy": 1.0,
|
| 6702 |
+
"eval_f1": 1.0,
|
| 6703 |
+
"eval_loss": 9.962059266399592e-05,
|
| 6704 |
+
"eval_precision": 1.0,
|
| 6705 |
+
"eval_recall": 1.0,
|
| 6706 |
+
"eval_runtime": 52.1678,
|
| 6707 |
+
"eval_samples_per_second": 5.597,
|
| 6708 |
+
"eval_steps_per_second": 0.192,
|
| 6709 |
+
"step": 880
|
| 6710 |
+
},
|
| 6711 |
+
{
|
| 6712 |
+
"epoch": 0.7466101694915255,
|
| 6713 |
+
"grad_norm": 0.9392852783203125,
|
| 6714 |
+
"learning_rate": 3.6632158307248713e-06,
|
| 6715 |
+
"loss": 0.0043,
|
| 6716 |
+
"step": 881
|
| 6717 |
+
},
|
| 6718 |
+
{
|
| 6719 |
+
"epoch": 0.747457627118644,
|
| 6720 |
+
"grad_norm": 0.8620559573173523,
|
| 6721 |
+
"learning_rate": 3.6403591709968924e-06,
|
| 6722 |
+
"loss": 0.0049,
|
| 6723 |
+
"step": 882
|
| 6724 |
+
},
|
| 6725 |
+
{
|
| 6726 |
+
"epoch": 0.7483050847457627,
|
| 6727 |
+
"grad_norm": 1.7161260843276978,
|
| 6728 |
+
"learning_rate": 3.617558163551802e-06,
|
| 6729 |
+
"loss": 0.0101,
|
| 6730 |
+
"step": 883
|
| 6731 |
+
},
|
| 6732 |
+
{
|
| 6733 |
+
"epoch": 0.7491525423728813,
|
| 6734 |
+
"grad_norm": 0.5389106273651123,
|
| 6735 |
+
"learning_rate": 3.5948130079178666e-06,
|
| 6736 |
+
"loss": 0.0052,
|
| 6737 |
+
"step": 884
|
| 6738 |
+
},
|
| 6739 |
+
{
|
| 6740 |
+
"epoch": 0.75,
|
| 6741 |
+
"grad_norm": 1.6375874280929565,
|
| 6742 |
+
"learning_rate": 3.5721239031346067e-06,
|
| 6743 |
+
"loss": 0.0099,
|
| 6744 |
+
"step": 885
|
| 6745 |
+
},
|
| 6746 |
+
{
|
| 6747 |
+
"epoch": 0.7508474576271187,
|
| 6748 |
+
"grad_norm": 0.791953444480896,
|
| 6749 |
+
"learning_rate": 3.5494910477510445e-06,
|
| 6750 |
+
"loss": 0.0035,
|
| 6751 |
+
"step": 886
|
| 6752 |
+
},
|
| 6753 |
+
{
|
| 6754 |
+
"epoch": 0.7516949152542373,
|
| 6755 |
+
"grad_norm": 2.004249095916748,
|
| 6756 |
+
"learning_rate": 3.526914639823973e-06,
|
| 6757 |
+
"loss": 0.012,
|
| 6758 |
+
"step": 887
|
| 6759 |
+
},
|
| 6760 |
+
{
|
| 6761 |
+
"epoch": 0.752542372881356,
|
| 6762 |
+
"grad_norm": 0.7185747623443604,
|
| 6763 |
+
"learning_rate": 3.5043948769162227e-06,
|
| 6764 |
+
"loss": 0.0037,
|
| 6765 |
+
"step": 888
|
| 6766 |
+
},
|
| 6767 |
+
{
|
| 6768 |
+
"epoch": 0.7533898305084745,
|
| 6769 |
+
"grad_norm": 1.0839685201644897,
|
| 6770 |
+
"learning_rate": 3.4819319560949303e-06,
|
| 6771 |
+
"loss": 0.0043,
|
| 6772 |
+
"step": 889
|
| 6773 |
+
},
|
| 6774 |
+
{
|
| 6775 |
+
"epoch": 0.7542372881355932,
|
| 6776 |
+
"grad_norm": 1.8889586925506592,
|
| 6777 |
+
"learning_rate": 3.4595260739298174e-06,
|
| 6778 |
+
"loss": 0.0236,
|
| 6779 |
+
"step": 890
|
| 6780 |
+
},
|
| 6781 |
+
{
|
| 6782 |
+
"epoch": 0.7550847457627119,
|
| 6783 |
+
"grad_norm": 1.4316060543060303,
|
| 6784 |
+
"learning_rate": 3.437177426491468e-06,
|
| 6785 |
+
"loss": 0.0168,
|
| 6786 |
+
"step": 891
|
| 6787 |
+
},
|
| 6788 |
+
{
|
| 6789 |
+
"epoch": 0.7559322033898305,
|
| 6790 |
+
"grad_norm": 1.128692865371704,
|
| 6791 |
+
"learning_rate": 3.414886209349615e-06,
|
| 6792 |
+
"loss": 0.007,
|
| 6793 |
+
"step": 892
|
| 6794 |
+
},
|
| 6795 |
+
{
|
| 6796 |
+
"epoch": 0.7567796610169492,
|
| 6797 |
+
"grad_norm": 1.5015407800674438,
|
| 6798 |
+
"learning_rate": 3.3926526175714246e-06,
|
| 6799 |
+
"loss": 0.0063,
|
| 6800 |
+
"step": 893
|
| 6801 |
+
},
|
| 6802 |
+
{
|
| 6803 |
+
"epoch": 0.7576271186440678,
|
| 6804 |
+
"grad_norm": 0.8763653635978699,
|
| 6805 |
+
"learning_rate": 3.3704768457197933e-06,
|
| 6806 |
+
"loss": 0.0039,
|
| 6807 |
+
"step": 894
|
| 6808 |
+
},
|
| 6809 |
+
{
|
| 6810 |
+
"epoch": 0.7584745762711864,
|
| 6811 |
+
"grad_norm": 2.3314383029937744,
|
| 6812 |
+
"learning_rate": 3.3483590878516437e-06,
|
| 6813 |
+
"loss": 0.0201,
|
| 6814 |
+
"step": 895
|
| 6815 |
+
},
|
| 6816 |
+
{
|
| 6817 |
+
"epoch": 0.7593220338983051,
|
| 6818 |
+
"grad_norm": 1.774864912033081,
|
| 6819 |
+
"learning_rate": 3.3262995375162256e-06,
|
| 6820 |
+
"loss": 0.0058,
|
| 6821 |
+
"step": 896
|
| 6822 |
+
},
|
| 6823 |
+
{
|
| 6824 |
+
"epoch": 0.7601694915254237,
|
| 6825 |
+
"grad_norm": 0.5185796618461609,
|
| 6826 |
+
"learning_rate": 3.304298387753426e-06,
|
| 6827 |
+
"loss": 0.0022,
|
| 6828 |
+
"step": 897
|
| 6829 |
+
},
|
| 6830 |
+
{
|
| 6831 |
+
"epoch": 0.7610169491525424,
|
| 6832 |
+
"grad_norm": 1.1957974433898926,
|
| 6833 |
+
"learning_rate": 3.282355831092072e-06,
|
| 6834 |
+
"loss": 0.0046,
|
| 6835 |
+
"step": 898
|
| 6836 |
+
},
|
| 6837 |
+
{
|
| 6838 |
+
"epoch": 0.761864406779661,
|
| 6839 |
+
"grad_norm": 1.759142518043518,
|
| 6840 |
+
"learning_rate": 3.2604720595482598e-06,
|
| 6841 |
+
"loss": 0.0097,
|
| 6842 |
+
"step": 899
|
| 6843 |
+
},
|
| 6844 |
+
{
|
| 6845 |
+
"epoch": 0.7627118644067796,
|
| 6846 |
+
"grad_norm": 2.0010337829589844,
|
| 6847 |
+
"learning_rate": 3.2386472646236565e-06,
|
| 6848 |
+
"loss": 0.0107,
|
| 6849 |
+
"step": 900
|
| 6850 |
+
},
|
| 6851 |
+
{
|
| 6852 |
+
"epoch": 0.7627118644067796,
|
| 6853 |
+
"eval_accuracy": 1.0,
|
| 6854 |
+
"eval_f1": 1.0,
|
| 6855 |
+
"eval_loss": 9.30224996409379e-05,
|
| 6856 |
+
"eval_precision": 1.0,
|
| 6857 |
+
"eval_recall": 1.0,
|
| 6858 |
+
"eval_runtime": 49.6959,
|
| 6859 |
+
"eval_samples_per_second": 5.876,
|
| 6860 |
+
"eval_steps_per_second": 0.201,
|
| 6861 |
+
"step": 900
|
| 6862 |
}
|
| 6863 |
],
|
| 6864 |
"logging_steps": 1,
|
|
|
|
| 6878 |
"attributes": {}
|
| 6879 |
}
|
| 6880 |
},
|
| 6881 |
+
"total_flos": 2.771703818092544e+17,
|
| 6882 |
"train_batch_size": 8,
|
| 6883 |
"trial_name": null,
|
| 6884 |
"trial_params": null
|