Training in progress, step 200, 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:c0ba6fb33cfa34a3a19c1c859523ba5b8ee34e4ce14cd7ee85604eeb2a478122
|
| 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:af5a9ded1ac1ad15369c22168c9bdc24120369b807b0236304fc238cd01770cd
|
| 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:aac30f6108157475f0f5a525a5d713d9caa83e68412d6e9feee34fa1c788d678
|
| 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 sha256:39a8a076cc77e594caeb9e94a5de64b7a427c01ec5dd10b1dbe76fa77717e2cd
|
| 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 sha256:28d745959ce06db61825f40fb63ea63b9e62f268815aaab31b17f3705247564b
|
| 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 sha256:1c5600d73cdd63622761d1592bb37fa01c39c9bcc957af85fe4bd2e4cd01fabc
|
| 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:be56560930f7a89d468b530340537eec5b918174080267a8cc6186d4978acf89
|
| 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 sha256:851b0a400aa71d9fc3d83e0e2570f4bbeaf98efc2e51c1f18c4d64aa51f39304
|
| 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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:805afa176b455b67a891f7c63c255879dd3a372d6c9fa2140f3c0a2149d52710
|
| 3 |
size 14960
|
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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:656385b8033d1cc9de4c8239cf888e2d83a5db8f95016de71e971858eab1c195
|
| 3 |
size 14960
|
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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a4775b283f1cbab74e1bfc47bfbe045632e0a9c46d8f354762f3216e862bf61
|
| 3 |
size 14960
|
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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ecbc04b6bcc44f7032a40edb9b3c06e3acf5ba0f1fb508b9a44802995aad5b9
|
| 3 |
size 14960
|
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 sha256:88d5a351fddcb4718730dd82c69354176cd179de4c82fa6d41e0282fb5e2ab11
|
| 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,
|
|
@@ -779,6 +779,766 @@
|
|
| 779 |
"eval_samples_per_second": 6.352,
|
| 780 |
"eval_steps_per_second": 0.24,
|
| 781 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 782 |
}
|
| 783 |
],
|
| 784 |
"logging_steps": 1,
|
|
@@ -798,7 +1558,7 @@
|
|
| 798 |
"attributes": {}
|
| 799 |
}
|
| 800 |
},
|
| 801 |
-
"total_flos":
|
| 802 |
"train_batch_size": 8,
|
| 803 |
"trial_name": null,
|
| 804 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.5747126436781609,
|
| 5 |
"eval_steps": 20,
|
| 6 |
+
"global_step": 200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 779 |
"eval_samples_per_second": 6.352,
|
| 780 |
"eval_steps_per_second": 0.24,
|
| 781 |
"step": 100
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.29022988505747127,
|
| 785 |
+
"grad_norm": 2.5615780353546143,
|
| 786 |
+
"learning_rate": 1.7884911347740556e-05,
|
| 787 |
+
"loss": 0.3328,
|
| 788 |
+
"step": 101
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.29310344827586204,
|
| 792 |
+
"grad_norm": 2.6771366596221924,
|
| 793 |
+
"learning_rate": 1.782278280813882e-05,
|
| 794 |
+
"loss": 0.3584,
|
| 795 |
+
"step": 102
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.2959770114942529,
|
| 799 |
+
"grad_norm": 2.411428451538086,
|
| 800 |
+
"learning_rate": 1.775986619145697e-05,
|
| 801 |
+
"loss": 0.3463,
|
| 802 |
+
"step": 103
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.2988505747126437,
|
| 806 |
+
"grad_norm": 1.991093397140503,
|
| 807 |
+
"learning_rate": 1.7696167835994927e-05,
|
| 808 |
+
"loss": 0.2636,
|
| 809 |
+
"step": 104
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.3017241379310345,
|
| 813 |
+
"grad_norm": 2.979641914367676,
|
| 814 |
+
"learning_rate": 1.7631694158805945e-05,
|
| 815 |
+
"loss": 0.2833,
|
| 816 |
+
"step": 105
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.3045977011494253,
|
| 820 |
+
"grad_norm": 2.2680490016937256,
|
| 821 |
+
"learning_rate": 1.7566451655050197e-05,
|
| 822 |
+
"loss": 0.2915,
|
| 823 |
+
"step": 106
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.3074712643678161,
|
| 827 |
+
"grad_norm": 2.6766483783721924,
|
| 828 |
+
"learning_rate": 1.7500446897340408e-05,
|
| 829 |
+
"loss": 0.3172,
|
| 830 |
+
"step": 107
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.3103448275862069,
|
| 834 |
+
"grad_norm": 2.766521692276001,
|
| 835 |
+
"learning_rate": 1.7433686535079736e-05,
|
| 836 |
+
"loss": 0.3347,
|
| 837 |
+
"step": 108
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.3132183908045977,
|
| 841 |
+
"grad_norm": 2.0438175201416016,
|
| 842 |
+
"learning_rate": 1.736617729379191e-05,
|
| 843 |
+
"loss": 0.3315,
|
| 844 |
+
"step": 109
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.3160919540229885,
|
| 848 |
+
"grad_norm": 2.5706722736358643,
|
| 849 |
+
"learning_rate": 1.7297925974443675e-05,
|
| 850 |
+
"loss": 0.2903,
|
| 851 |
+
"step": 110
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.31896551724137934,
|
| 855 |
+
"grad_norm": 6.830801010131836,
|
| 856 |
+
"learning_rate": 1.7228939452759666e-05,
|
| 857 |
+
"loss": 0.372,
|
| 858 |
+
"step": 111
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.3218390804597701,
|
| 862 |
+
"grad_norm": 2.375408411026001,
|
| 863 |
+
"learning_rate": 1.7159224678529734e-05,
|
| 864 |
+
"loss": 0.2875,
|
| 865 |
+
"step": 112
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.32471264367816094,
|
| 869 |
+
"grad_norm": 2.542205572128296,
|
| 870 |
+
"learning_rate": 1.7088788674908817e-05,
|
| 871 |
+
"loss": 0.3327,
|
| 872 |
+
"step": 113
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.3275862068965517,
|
| 876 |
+
"grad_norm": 2.645517587661743,
|
| 877 |
+
"learning_rate": 1.7017638537709426e-05,
|
| 878 |
+
"loss": 0.3225,
|
| 879 |
+
"step": 114
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.33045977011494254,
|
| 883 |
+
"grad_norm": 3.9527246952056885,
|
| 884 |
+
"learning_rate": 1.6945781434686783e-05,
|
| 885 |
+
"loss": 0.3683,
|
| 886 |
+
"step": 115
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.3333333333333333,
|
| 890 |
+
"grad_norm": 3.478126049041748,
|
| 891 |
+
"learning_rate": 1.6873224604816753e-05,
|
| 892 |
+
"loss": 0.3448,
|
| 893 |
+
"step": 116
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.33620689655172414,
|
| 897 |
+
"grad_norm": 4.451388359069824,
|
| 898 |
+
"learning_rate": 1.679997535756657e-05,
|
| 899 |
+
"loss": 0.2846,
|
| 900 |
+
"step": 117
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.3390804597701149,
|
| 904 |
+
"grad_norm": 2.1524624824523926,
|
| 905 |
+
"learning_rate": 1.672604107215848e-05,
|
| 906 |
+
"loss": 0.273,
|
| 907 |
+
"step": 118
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.34195402298850575,
|
| 911 |
+
"grad_norm": 2.950127601623535,
|
| 912 |
+
"learning_rate": 1.6651429196826337e-05,
|
| 913 |
+
"loss": 0.3886,
|
| 914 |
+
"step": 119
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.3448275862068966,
|
| 918 |
+
"grad_norm": 3.5067648887634277,
|
| 919 |
+
"learning_rate": 1.6576147248065268e-05,
|
| 920 |
+
"loss": 0.3822,
|
| 921 |
+
"step": 120
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.3448275862068966,
|
| 925 |
+
"eval_accuracy": 0.8514851485148515,
|
| 926 |
+
"eval_f1": 0.7297297297297297,
|
| 927 |
+
"eval_loss": 0.35510221123695374,
|
| 928 |
+
"eval_precision": 0.6982758620689655,
|
| 929 |
+
"eval_recall": 0.7641509433962265,
|
| 930 |
+
"eval_runtime": 16.7999,
|
| 931 |
+
"eval_samples_per_second": 6.31,
|
| 932 |
+
"eval_steps_per_second": 0.238,
|
| 933 |
+
"step": 120
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.34770114942528735,
|
| 937 |
+
"grad_norm": 4.016488552093506,
|
| 938 |
+
"learning_rate": 1.6500202809874446e-05,
|
| 939 |
+
"loss": 0.3354,
|
| 940 |
+
"step": 121
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.3505747126436782,
|
| 944 |
+
"grad_norm": 2.8852052688598633,
|
| 945 |
+
"learning_rate": 1.6423603532993074e-05,
|
| 946 |
+
"loss": 0.3143,
|
| 947 |
+
"step": 122
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.35344827586206895,
|
| 951 |
+
"grad_norm": 3.2638821601867676,
|
| 952 |
+
"learning_rate": 1.634635713412964e-05,
|
| 953 |
+
"loss": 0.3273,
|
| 954 |
+
"step": 123
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.3563218390804598,
|
| 958 |
+
"grad_norm": 2.6508638858795166,
|
| 959 |
+
"learning_rate": 1.626847139518452e-05,
|
| 960 |
+
"loss": 0.3539,
|
| 961 |
+
"step": 124
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.35919540229885055,
|
| 965 |
+
"grad_norm": 3.1945559978485107,
|
| 966 |
+
"learning_rate": 1.618995416246601e-05,
|
| 967 |
+
"loss": 0.3577,
|
| 968 |
+
"step": 125
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.3620689655172414,
|
| 972 |
+
"grad_norm": 3.6181535720825195,
|
| 973 |
+
"learning_rate": 1.6110813345899914e-05,
|
| 974 |
+
"loss": 0.3194,
|
| 975 |
+
"step": 126
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.3649425287356322,
|
| 979 |
+
"grad_norm": 3.1202199459075928,
|
| 980 |
+
"learning_rate": 1.6031056918232642e-05,
|
| 981 |
+
"loss": 0.3014,
|
| 982 |
+
"step": 127
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.367816091954023,
|
| 986 |
+
"grad_norm": 3.069596529006958,
|
| 987 |
+
"learning_rate": 1.595069291422807e-05,
|
| 988 |
+
"loss": 0.2934,
|
| 989 |
+
"step": 128
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.3706896551724138,
|
| 993 |
+
"grad_norm": 2.2219157218933105,
|
| 994 |
+
"learning_rate": 1.586972942985807e-05,
|
| 995 |
+
"loss": 0.2801,
|
| 996 |
+
"step": 129
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.3735632183908046,
|
| 1000 |
+
"grad_norm": 2.305501699447632,
|
| 1001 |
+
"learning_rate": 1.5788174621486936e-05,
|
| 1002 |
+
"loss": 0.2895,
|
| 1003 |
+
"step": 130
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.3764367816091954,
|
| 1007 |
+
"grad_norm": 2.0006630420684814,
|
| 1008 |
+
"learning_rate": 1.570603670504969e-05,
|
| 1009 |
+
"loss": 0.2614,
|
| 1010 |
+
"step": 131
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3793103448275862,
|
| 1014 |
+
"grad_norm": 2.4915575981140137,
|
| 1015 |
+
"learning_rate": 1.5623323955224404e-05,
|
| 1016 |
+
"loss": 0.3215,
|
| 1017 |
+
"step": 132
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.382183908045977,
|
| 1021 |
+
"grad_norm": 2.3519020080566406,
|
| 1022 |
+
"learning_rate": 1.5540044704598588e-05,
|
| 1023 |
+
"loss": 0.291,
|
| 1024 |
+
"step": 133
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3850574712643678,
|
| 1028 |
+
"grad_norm": 2.4606783390045166,
|
| 1029 |
+
"learning_rate": 1.5456207342829777e-05,
|
| 1030 |
+
"loss": 0.3641,
|
| 1031 |
+
"step": 134
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3879310344827586,
|
| 1035 |
+
"grad_norm": 2.4096121788024902,
|
| 1036 |
+
"learning_rate": 1.5371820315800316e-05,
|
| 1037 |
+
"loss": 0.3071,
|
| 1038 |
+
"step": 135
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.39080459770114945,
|
| 1042 |
+
"grad_norm": 2.9145760536193848,
|
| 1043 |
+
"learning_rate": 1.5286892124766546e-05,
|
| 1044 |
+
"loss": 0.2811,
|
| 1045 |
+
"step": 136
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.3936781609195402,
|
| 1049 |
+
"grad_norm": 1.954529881477356,
|
| 1050 |
+
"learning_rate": 1.5201431325502332e-05,
|
| 1051 |
+
"loss": 0.3137,
|
| 1052 |
+
"step": 137
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.39655172413793105,
|
| 1056 |
+
"grad_norm": 3.2401654720306396,
|
| 1057 |
+
"learning_rate": 1.5115446527437193e-05,
|
| 1058 |
+
"loss": 0.3052,
|
| 1059 |
+
"step": 138
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3994252873563218,
|
| 1063 |
+
"grad_norm": 1.9580965042114258,
|
| 1064 |
+
"learning_rate": 1.5028946392788934e-05,
|
| 1065 |
+
"loss": 0.2898,
|
| 1066 |
+
"step": 139
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.40229885057471265,
|
| 1070 |
+
"grad_norm": 3.239868640899658,
|
| 1071 |
+
"learning_rate": 1.4941939635691036e-05,
|
| 1072 |
+
"loss": 0.3955,
|
| 1073 |
+
"step": 140
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.40229885057471265,
|
| 1077 |
+
"eval_accuracy": 0.8589108910891089,
|
| 1078 |
+
"eval_f1": 0.7135678391959799,
|
| 1079 |
+
"eval_loss": 0.34418779611587524,
|
| 1080 |
+
"eval_precision": 0.7634408602150538,
|
| 1081 |
+
"eval_recall": 0.6698113207547169,
|
| 1082 |
+
"eval_runtime": 16.4081,
|
| 1083 |
+
"eval_samples_per_second": 6.46,
|
| 1084 |
+
"eval_steps_per_second": 0.244,
|
| 1085 |
+
"step": 140
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.4051724137931034,
|
| 1089 |
+
"grad_norm": 4.790642738342285,
|
| 1090 |
+
"learning_rate": 1.4854435021314766e-05,
|
| 1091 |
+
"loss": 0.3939,
|
| 1092 |
+
"step": 141
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 0.40804597701149425,
|
| 1096 |
+
"grad_norm": 1.9184012413024902,
|
| 1097 |
+
"learning_rate": 1.4766441364986162e-05,
|
| 1098 |
+
"loss": 0.2496,
|
| 1099 |
+
"step": 142
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.4109195402298851,
|
| 1103 |
+
"grad_norm": 3.910475969314575,
|
| 1104 |
+
"learning_rate": 1.467796753129797e-05,
|
| 1105 |
+
"loss": 0.3206,
|
| 1106 |
+
"step": 143
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.41379310344827586,
|
| 1110 |
+
"grad_norm": 5.272514820098877,
|
| 1111 |
+
"learning_rate": 1.4589022433216616e-05,
|
| 1112 |
+
"loss": 0.3299,
|
| 1113 |
+
"step": 144
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.4166666666666667,
|
| 1117 |
+
"grad_norm": 4.5233941078186035,
|
| 1118 |
+
"learning_rate": 1.4499615031184297e-05,
|
| 1119 |
+
"loss": 0.3089,
|
| 1120 |
+
"step": 145
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.41954022988505746,
|
| 1124 |
+
"grad_norm": 3.6632847785949707,
|
| 1125 |
+
"learning_rate": 1.4409754332216303e-05,
|
| 1126 |
+
"loss": 0.3436,
|
| 1127 |
+
"step": 146
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 0.4224137931034483,
|
| 1131 |
+
"grad_norm": 2.548621892929077,
|
| 1132 |
+
"learning_rate": 1.431944938899363e-05,
|
| 1133 |
+
"loss": 0.3434,
|
| 1134 |
+
"step": 147
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.42528735632183906,
|
| 1138 |
+
"grad_norm": 2.903876304626465,
|
| 1139 |
+
"learning_rate": 1.4228709298950998e-05,
|
| 1140 |
+
"loss": 0.2672,
|
| 1141 |
+
"step": 148
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.4281609195402299,
|
| 1145 |
+
"grad_norm": 4.375998020172119,
|
| 1146 |
+
"learning_rate": 1.4137543203360382e-05,
|
| 1147 |
+
"loss": 0.2741,
|
| 1148 |
+
"step": 149
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.43103448275862066,
|
| 1152 |
+
"grad_norm": 4.29938268661499,
|
| 1153 |
+
"learning_rate": 1.4045960286410093e-05,
|
| 1154 |
+
"loss": 0.3264,
|
| 1155 |
+
"step": 150
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.4339080459770115,
|
| 1159 |
+
"grad_norm": 2.3729660511016846,
|
| 1160 |
+
"learning_rate": 1.395396977427955e-05,
|
| 1161 |
+
"loss": 0.3135,
|
| 1162 |
+
"step": 151
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.4367816091954023,
|
| 1166 |
+
"grad_norm": 4.978923797607422,
|
| 1167 |
+
"learning_rate": 1.3861580934209832e-05,
|
| 1168 |
+
"loss": 0.3174,
|
| 1169 |
+
"step": 152
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 0.4396551724137931,
|
| 1173 |
+
"grad_norm": 2.483069658279419,
|
| 1174 |
+
"learning_rate": 1.376880307357009e-05,
|
| 1175 |
+
"loss": 0.2683,
|
| 1176 |
+
"step": 153
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.4425287356321839,
|
| 1180 |
+
"grad_norm": 2.2193140983581543,
|
| 1181 |
+
"learning_rate": 1.3675645538919884e-05,
|
| 1182 |
+
"loss": 0.257,
|
| 1183 |
+
"step": 154
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.4454022988505747,
|
| 1187 |
+
"grad_norm": 2.5293915271759033,
|
| 1188 |
+
"learning_rate": 1.3582117715067628e-05,
|
| 1189 |
+
"loss": 0.3204,
|
| 1190 |
+
"step": 155
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.4482758620689655,
|
| 1194 |
+
"grad_norm": 5.482168674468994,
|
| 1195 |
+
"learning_rate": 1.3488229024125142e-05,
|
| 1196 |
+
"loss": 0.3334,
|
| 1197 |
+
"step": 156
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.4511494252873563,
|
| 1201 |
+
"grad_norm": 3.027650833129883,
|
| 1202 |
+
"learning_rate": 1.3393988924558445e-05,
|
| 1203 |
+
"loss": 0.2637,
|
| 1204 |
+
"step": 157
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.4540229885057471,
|
| 1208 |
+
"grad_norm": 5.8336262702941895,
|
| 1209 |
+
"learning_rate": 1.3299406910234917e-05,
|
| 1210 |
+
"loss": 0.3285,
|
| 1211 |
+
"step": 158
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 0.45689655172413796,
|
| 1215 |
+
"grad_norm": 3.0736873149871826,
|
| 1216 |
+
"learning_rate": 1.3204492509466862e-05,
|
| 1217 |
+
"loss": 0.3032,
|
| 1218 |
+
"step": 159
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 0.45977011494252873,
|
| 1222 |
+
"grad_norm": 4.618353366851807,
|
| 1223 |
+
"learning_rate": 1.3109255284051615e-05,
|
| 1224 |
+
"loss": 0.34,
|
| 1225 |
+
"step": 160
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 0.45977011494252873,
|
| 1229 |
+
"eval_accuracy": 0.8613861386138614,
|
| 1230 |
+
"eval_f1": 0.7431192660550459,
|
| 1231 |
+
"eval_loss": 0.3399461805820465,
|
| 1232 |
+
"eval_precision": 0.7232142857142857,
|
| 1233 |
+
"eval_recall": 0.7641509433962265,
|
| 1234 |
+
"eval_runtime": 16.589,
|
| 1235 |
+
"eval_samples_per_second": 6.39,
|
| 1236 |
+
"eval_steps_per_second": 0.241,
|
| 1237 |
+
"step": 160
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.46264367816091956,
|
| 1241 |
+
"grad_norm": 1.9849704504013062,
|
| 1242 |
+
"learning_rate": 1.3013704828308276e-05,
|
| 1243 |
+
"loss": 0.2715,
|
| 1244 |
+
"step": 161
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.46551724137931033,
|
| 1248 |
+
"grad_norm": 3.116058111190796,
|
| 1249 |
+
"learning_rate": 1.2917850768111171e-05,
|
| 1250 |
+
"loss": 0.3093,
|
| 1251 |
+
"step": 162
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.46839080459770116,
|
| 1255 |
+
"grad_norm": 3.0582470893859863,
|
| 1256 |
+
"learning_rate": 1.282170275992012e-05,
|
| 1257 |
+
"loss": 0.3502,
|
| 1258 |
+
"step": 163
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 0.47126436781609193,
|
| 1262 |
+
"grad_norm": 5.603944301605225,
|
| 1263 |
+
"learning_rate": 1.2725270489807637e-05,
|
| 1264 |
+
"loss": 0.3453,
|
| 1265 |
+
"step": 164
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 0.47413793103448276,
|
| 1269 |
+
"grad_norm": 3.758445978164673,
|
| 1270 |
+
"learning_rate": 1.2628563672483147e-05,
|
| 1271 |
+
"loss": 0.3542,
|
| 1272 |
+
"step": 165
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 0.47701149425287354,
|
| 1276 |
+
"grad_norm": 2.5959854125976562,
|
| 1277 |
+
"learning_rate": 1.2531592050314308e-05,
|
| 1278 |
+
"loss": 0.3346,
|
| 1279 |
+
"step": 166
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 0.47988505747126436,
|
| 1283 |
+
"grad_norm": 2.4197230339050293,
|
| 1284 |
+
"learning_rate": 1.2434365392345553e-05,
|
| 1285 |
+
"loss": 0.2817,
|
| 1286 |
+
"step": 167
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.4827586206896552,
|
| 1290 |
+
"grad_norm": 2.748478412628174,
|
| 1291 |
+
"learning_rate": 1.2336893493313946e-05,
|
| 1292 |
+
"loss": 0.3511,
|
| 1293 |
+
"step": 168
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 0.48563218390804597,
|
| 1297 |
+
"grad_norm": 3.117748975753784,
|
| 1298 |
+
"learning_rate": 1.223918617266245e-05,
|
| 1299 |
+
"loss": 0.3724,
|
| 1300 |
+
"step": 169
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 0.4885057471264368,
|
| 1304 |
+
"grad_norm": 2.6176235675811768,
|
| 1305 |
+
"learning_rate": 1.2141253273550698e-05,
|
| 1306 |
+
"loss": 0.3454,
|
| 1307 |
+
"step": 170
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 0.49137931034482757,
|
| 1311 |
+
"grad_norm": 2.648324966430664,
|
| 1312 |
+
"learning_rate": 1.2043104661863386e-05,
|
| 1313 |
+
"loss": 0.2713,
|
| 1314 |
+
"step": 171
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.4942528735632184,
|
| 1318 |
+
"grad_norm": 2.6428864002227783,
|
| 1319 |
+
"learning_rate": 1.1944750225216363e-05,
|
| 1320 |
+
"loss": 0.2618,
|
| 1321 |
+
"step": 172
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.49712643678160917,
|
| 1325 |
+
"grad_norm": 2.6188530921936035,
|
| 1326 |
+
"learning_rate": 1.1846199871960557e-05,
|
| 1327 |
+
"loss": 0.3443,
|
| 1328 |
+
"step": 173
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.5,
|
| 1332 |
+
"grad_norm": 2.8425047397613525,
|
| 1333 |
+
"learning_rate": 1.1747463530183781e-05,
|
| 1334 |
+
"loss": 0.3323,
|
| 1335 |
+
"step": 174
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 0.5028735632183908,
|
| 1339 |
+
"grad_norm": 3.0745861530303955,
|
| 1340 |
+
"learning_rate": 1.1648551146710557e-05,
|
| 1341 |
+
"loss": 0.2854,
|
| 1342 |
+
"step": 175
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 0.5057471264367817,
|
| 1346 |
+
"grad_norm": 3.5819411277770996,
|
| 1347 |
+
"learning_rate": 1.1549472686100079e-05,
|
| 1348 |
+
"loss": 0.3851,
|
| 1349 |
+
"step": 176
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 0.5086206896551724,
|
| 1353 |
+
"grad_norm": 2.7981903553009033,
|
| 1354 |
+
"learning_rate": 1.145023812964237e-05,
|
| 1355 |
+
"loss": 0.288,
|
| 1356 |
+
"step": 177
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 0.5114942528735632,
|
| 1360 |
+
"grad_norm": 2.600273847579956,
|
| 1361 |
+
"learning_rate": 1.1350857474352734e-05,
|
| 1362 |
+
"loss": 0.2982,
|
| 1363 |
+
"step": 178
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 0.514367816091954,
|
| 1367 |
+
"grad_norm": 2.207815170288086,
|
| 1368 |
+
"learning_rate": 1.1251340731964664e-05,
|
| 1369 |
+
"loss": 0.2697,
|
| 1370 |
+
"step": 179
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.5172413793103449,
|
| 1374 |
+
"grad_norm": 3.180569887161255,
|
| 1375 |
+
"learning_rate": 1.1151697927921242e-05,
|
| 1376 |
+
"loss": 0.2897,
|
| 1377 |
+
"step": 180
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 0.5172413793103449,
|
| 1381 |
+
"eval_accuracy": 0.8613861386138614,
|
| 1382 |
+
"eval_f1": 0.7333333333333333,
|
| 1383 |
+
"eval_loss": 0.3244224786758423,
|
| 1384 |
+
"eval_precision": 0.7403846153846154,
|
| 1385 |
+
"eval_recall": 0.7264150943396226,
|
| 1386 |
+
"eval_runtime": 16.4883,
|
| 1387 |
+
"eval_samples_per_second": 6.429,
|
| 1388 |
+
"eval_steps_per_second": 0.243,
|
| 1389 |
+
"step": 180
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.5201149425287356,
|
| 1393 |
+
"grad_norm": 4.220306396484375,
|
| 1394 |
+
"learning_rate": 1.1051939100365154e-05,
|
| 1395 |
+
"loss": 0.254,
|
| 1396 |
+
"step": 181
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.5229885057471264,
|
| 1400 |
+
"grad_norm": 2.567113161087036,
|
| 1401 |
+
"learning_rate": 1.0952074299127451e-05,
|
| 1402 |
+
"loss": 0.2826,
|
| 1403 |
+
"step": 182
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.5258620689655172,
|
| 1407 |
+
"grad_norm": 2.881145715713501,
|
| 1408 |
+
"learning_rate": 1.0852113584715103e-05,
|
| 1409 |
+
"loss": 0.2882,
|
| 1410 |
+
"step": 183
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.5287356321839081,
|
| 1414 |
+
"grad_norm": 2.523293972015381,
|
| 1415 |
+
"learning_rate": 1.0752067027297486e-05,
|
| 1416 |
+
"loss": 0.3224,
|
| 1417 |
+
"step": 184
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.5316091954022989,
|
| 1421 |
+
"grad_norm": 3.2287135124206543,
|
| 1422 |
+
"learning_rate": 1.065194470569193e-05,
|
| 1423 |
+
"loss": 0.3054,
|
| 1424 |
+
"step": 185
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.5344827586206896,
|
| 1428 |
+
"grad_norm": 2.670992851257324,
|
| 1429 |
+
"learning_rate": 1.0551756706348331e-05,
|
| 1430 |
+
"loss": 0.3404,
|
| 1431 |
+
"step": 186
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.5373563218390804,
|
| 1435 |
+
"grad_norm": 2.8072383403778076,
|
| 1436 |
+
"learning_rate": 1.0451513122333042e-05,
|
| 1437 |
+
"loss": 0.3289,
|
| 1438 |
+
"step": 187
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.5402298850574713,
|
| 1442 |
+
"grad_norm": 2.7612991333007812,
|
| 1443 |
+
"learning_rate": 1.035122405231209e-05,
|
| 1444 |
+
"loss": 0.2705,
|
| 1445 |
+
"step": 188
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.5431034482758621,
|
| 1449 |
+
"grad_norm": 3.2553176879882812,
|
| 1450 |
+
"learning_rate": 1.0250899599533833e-05,
|
| 1451 |
+
"loss": 0.3088,
|
| 1452 |
+
"step": 189
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.5459770114942529,
|
| 1456 |
+
"grad_norm": 3.1975409984588623,
|
| 1457 |
+
"learning_rate": 1.0150549870811108e-05,
|
| 1458 |
+
"loss": 0.4086,
|
| 1459 |
+
"step": 190
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.5488505747126436,
|
| 1463 |
+
"grad_norm": 3.0886051654815674,
|
| 1464 |
+
"learning_rate": 1.0050184975503104e-05,
|
| 1465 |
+
"loss": 0.3177,
|
| 1466 |
+
"step": 191
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.5517241379310345,
|
| 1470 |
+
"grad_norm": 4.40545654296875,
|
| 1471 |
+
"learning_rate": 9.949815024496901e-06,
|
| 1472 |
+
"loss": 0.346,
|
| 1473 |
+
"step": 192
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.5545977011494253,
|
| 1477 |
+
"grad_norm": 4.658652305603027,
|
| 1478 |
+
"learning_rate": 9.849450129188895e-06,
|
| 1479 |
+
"loss": 0.2821,
|
| 1480 |
+
"step": 193
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.5574712643678161,
|
| 1484 |
+
"grad_norm": 7.797382831573486,
|
| 1485 |
+
"learning_rate": 9.74910040046617e-06,
|
| 1486 |
+
"loss": 0.3194,
|
| 1487 |
+
"step": 194
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.5603448275862069,
|
| 1491 |
+
"grad_norm": 3.3024234771728516,
|
| 1492 |
+
"learning_rate": 9.648775947687914e-06,
|
| 1493 |
+
"loss": 0.2912,
|
| 1494 |
+
"step": 195
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.5632183908045977,
|
| 1498 |
+
"grad_norm": 2.7685601711273193,
|
| 1499 |
+
"learning_rate": 9.548486877666963e-06,
|
| 1500 |
+
"loss": 0.3019,
|
| 1501 |
+
"step": 196
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.5660919540229885,
|
| 1505 |
+
"grad_norm": 3.0222320556640625,
|
| 1506 |
+
"learning_rate": 9.448243293651676e-06,
|
| 1507 |
+
"loss": 0.2963,
|
| 1508 |
+
"step": 197
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.5689655172413793,
|
| 1512 |
+
"grad_norm": 3.1808133125305176,
|
| 1513 |
+
"learning_rate": 9.348055294308074e-06,
|
| 1514 |
+
"loss": 0.3107,
|
| 1515 |
+
"step": 198
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.5718390804597702,
|
| 1519 |
+
"grad_norm": 2.3709192276000977,
|
| 1520 |
+
"learning_rate": 9.247932972702514e-06,
|
| 1521 |
+
"loss": 0.294,
|
| 1522 |
+
"step": 199
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.5747126436781609,
|
| 1526 |
+
"grad_norm": 2.9564688205718994,
|
| 1527 |
+
"learning_rate": 9.147886415284903e-06,
|
| 1528 |
+
"loss": 0.2599,
|
| 1529 |
+
"step": 200
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.5747126436781609,
|
| 1533 |
+
"eval_accuracy": 0.8638613861386139,
|
| 1534 |
+
"eval_f1": 0.7417840375586855,
|
| 1535 |
+
"eval_loss": 0.32250022888183594,
|
| 1536 |
+
"eval_precision": 0.7383177570093458,
|
| 1537 |
+
"eval_recall": 0.7452830188679245,
|
| 1538 |
+
"eval_runtime": 16.3331,
|
| 1539 |
+
"eval_samples_per_second": 6.49,
|
| 1540 |
+
"eval_steps_per_second": 0.245,
|
| 1541 |
+
"step": 200
|
| 1542 |
}
|
| 1543 |
],
|
| 1544 |
"logging_steps": 1,
|
|
|
|
| 1558 |
"attributes": {}
|
| 1559 |
}
|
| 1560 |
},
|
| 1561 |
+
"total_flos": 6.331748017661542e+16,
|
| 1562 |
"train_batch_size": 8,
|
| 1563 |
"trial_name": null,
|
| 1564 |
"trial_params": null
|