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 sha256:8e408bfee6bc720f2626f42236cb7ed47eed15851394e64745280cda47346a24
|
| 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:eff316f99046dc065f6573097972dede0370afbf132b27b8c0122d95e707ef46
|
| 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:ff619163df21cd3fce301b8caa1204b04f12d8929b430f9ba8a93cc2b633db21
|
| 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:c9f0b1099d24eb7f394ea4f3fe171409b994687feddd6582b56703c5558366fc
|
| 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:ddc7713c77dfea2848b11f758cffc94b65fdb54736e77c0647082559d10aa06b
|
| 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 sha256:6610841a05b9998513c9700bc4bead2bfbd262e59da6e4197d08b8fa080a9641
|
| 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:f09233c56315737a8a0656ed9c80d92a6963808314b0fe48bf44cb8c6799ef3a
|
| 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:bebfedc8c1e6754606faa59b7c45e93ee3e4c7ec2913e3893c4695781b7892e7
|
| 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 sha256:6f2fd9d1d3847bee68df39de96a06913e37dc3cacd6dcaa01e654f56e2f4eb49
|
| 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 sha256:584d56bb430fe8df3c24eeab1822b6d753b2090cb92990956e81b8f8e3c6e416
|
| 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 sha256:283174a53dfb12f541f1223f29dfd78957ef99fb6b3f708ac21ff4aa6e7733c3
|
| 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:f4f93a55b6907505dff041e23ee75d98392142bc1e2e39401947dac1e4fb011a
|
| 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 sha256:75ef6331ac4b2e8cf5bcc3f43391a2f41a0430eee842180c387f3d81fdad2fdc
|
| 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": 5.75,
|
| 780 |
"eval_steps_per_second": 0.193,
|
| 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": 0.13531799729364005,
|
| 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": 5.75,
|
| 780 |
"eval_steps_per_second": 0.193,
|
| 781 |
"step": 100
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.06833558863328823,
|
| 785 |
+
"grad_norm": 2.3830792903900146,
|
| 786 |
+
"learning_rate": 1.364864864864865e-05,
|
| 787 |
+
"loss": 0.2756,
|
| 788 |
+
"step": 101
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.06901217861975643,
|
| 792 |
+
"grad_norm": 2.8554539680480957,
|
| 793 |
+
"learning_rate": 1.3783783783783784e-05,
|
| 794 |
+
"loss": 0.3233,
|
| 795 |
+
"step": 102
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.06968876860622462,
|
| 799 |
+
"grad_norm": 3.331234931945801,
|
| 800 |
+
"learning_rate": 1.391891891891892e-05,
|
| 801 |
+
"loss": 0.3524,
|
| 802 |
+
"step": 103
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.07036535859269283,
|
| 806 |
+
"grad_norm": 2.8779256343841553,
|
| 807 |
+
"learning_rate": 1.4054054054054055e-05,
|
| 808 |
+
"loss": 0.323,
|
| 809 |
+
"step": 104
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.07104194857916103,
|
| 813 |
+
"grad_norm": 2.8393092155456543,
|
| 814 |
+
"learning_rate": 1.4189189189189189e-05,
|
| 815 |
+
"loss": 0.328,
|
| 816 |
+
"step": 105
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.07171853856562922,
|
| 820 |
+
"grad_norm": 3.7622110843658447,
|
| 821 |
+
"learning_rate": 1.4324324324324326e-05,
|
| 822 |
+
"loss": 0.3449,
|
| 823 |
+
"step": 106
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.07239512855209743,
|
| 827 |
+
"grad_norm": 3.734447479248047,
|
| 828 |
+
"learning_rate": 1.4459459459459462e-05,
|
| 829 |
+
"loss": 0.3367,
|
| 830 |
+
"step": 107
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.07307171853856563,
|
| 834 |
+
"grad_norm": 4.105041980743408,
|
| 835 |
+
"learning_rate": 1.4594594594594596e-05,
|
| 836 |
+
"loss": 0.3038,
|
| 837 |
+
"step": 108
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.07374830852503383,
|
| 841 |
+
"grad_norm": 3.9254539012908936,
|
| 842 |
+
"learning_rate": 1.4729729729729731e-05,
|
| 843 |
+
"loss": 0.2617,
|
| 844 |
+
"step": 109
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.07442489851150202,
|
| 848 |
+
"grad_norm": 5.182884693145752,
|
| 849 |
+
"learning_rate": 1.4864864864864865e-05,
|
| 850 |
+
"loss": 0.3423,
|
| 851 |
+
"step": 110
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.07510148849797023,
|
| 855 |
+
"grad_norm": 3.852728843688965,
|
| 856 |
+
"learning_rate": 1.5000000000000002e-05,
|
| 857 |
+
"loss": 0.246,
|
| 858 |
+
"step": 111
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.07577807848443843,
|
| 862 |
+
"grad_norm": 3.291020631790161,
|
| 863 |
+
"learning_rate": 1.5135135135135138e-05,
|
| 864 |
+
"loss": 0.3383,
|
| 865 |
+
"step": 112
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.07645466847090664,
|
| 869 |
+
"grad_norm": 5.644819259643555,
|
| 870 |
+
"learning_rate": 1.527027027027027e-05,
|
| 871 |
+
"loss": 0.2452,
|
| 872 |
+
"step": 113
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.07713125845737483,
|
| 876 |
+
"grad_norm": 6.728042125701904,
|
| 877 |
+
"learning_rate": 1.540540540540541e-05,
|
| 878 |
+
"loss": 0.2767,
|
| 879 |
+
"step": 114
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.07780784844384303,
|
| 883 |
+
"grad_norm": 4.200859546661377,
|
| 884 |
+
"learning_rate": 1.554054054054054e-05,
|
| 885 |
+
"loss": 0.2707,
|
| 886 |
+
"step": 115
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.07848443843031123,
|
| 890 |
+
"grad_norm": 3.9574716091156006,
|
| 891 |
+
"learning_rate": 1.5675675675675676e-05,
|
| 892 |
+
"loss": 0.2733,
|
| 893 |
+
"step": 116
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.07916102841677942,
|
| 897 |
+
"grad_norm": 3.50284743309021,
|
| 898 |
+
"learning_rate": 1.581081081081081e-05,
|
| 899 |
+
"loss": 0.2615,
|
| 900 |
+
"step": 117
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.07983761840324763,
|
| 904 |
+
"grad_norm": 7.720501899719238,
|
| 905 |
+
"learning_rate": 1.5945945945945947e-05,
|
| 906 |
+
"loss": 0.2353,
|
| 907 |
+
"step": 118
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.08051420838971583,
|
| 911 |
+
"grad_norm": 5.794226169586182,
|
| 912 |
+
"learning_rate": 1.6081081081081083e-05,
|
| 913 |
+
"loss": 0.2454,
|
| 914 |
+
"step": 119
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.08119079837618404,
|
| 918 |
+
"grad_norm": 6.7274250984191895,
|
| 919 |
+
"learning_rate": 1.6216216216216218e-05,
|
| 920 |
+
"loss": 0.2948,
|
| 921 |
+
"step": 120
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.08119079837618404,
|
| 925 |
+
"eval_accuracy": 0.7678899082568807,
|
| 926 |
+
"eval_f1": 0.3990498812351544,
|
| 927 |
+
"eval_loss": 0.7271434664726257,
|
| 928 |
+
"eval_precision": 0.6829268292682927,
|
| 929 |
+
"eval_recall": 0.28187919463087246,
|
| 930 |
+
"eval_runtime": 51.8636,
|
| 931 |
+
"eval_samples_per_second": 5.746,
|
| 932 |
+
"eval_steps_per_second": 0.193,
|
| 933 |
+
"step": 120
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.08186738836265223,
|
| 937 |
+
"grad_norm": 4.321250915527344,
|
| 938 |
+
"learning_rate": 1.6351351351351354e-05,
|
| 939 |
+
"loss": 0.2774,
|
| 940 |
+
"step": 121
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.08254397834912043,
|
| 944 |
+
"grad_norm": 5.205666542053223,
|
| 945 |
+
"learning_rate": 1.648648648648649e-05,
|
| 946 |
+
"loss": 0.254,
|
| 947 |
+
"step": 122
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.08322056833558863,
|
| 951 |
+
"grad_norm": 4.166099548339844,
|
| 952 |
+
"learning_rate": 1.662162162162162e-05,
|
| 953 |
+
"loss": 0.2455,
|
| 954 |
+
"step": 123
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.08389715832205684,
|
| 958 |
+
"grad_norm": 5.376754283905029,
|
| 959 |
+
"learning_rate": 1.6756756756756757e-05,
|
| 960 |
+
"loss": 0.2982,
|
| 961 |
+
"step": 124
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.08457374830852503,
|
| 965 |
+
"grad_norm": 5.893986225128174,
|
| 966 |
+
"learning_rate": 1.6891891891891896e-05,
|
| 967 |
+
"loss": 0.2632,
|
| 968 |
+
"step": 125
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.08525033829499323,
|
| 972 |
+
"grad_norm": 5.461335182189941,
|
| 973 |
+
"learning_rate": 1.7027027027027028e-05,
|
| 974 |
+
"loss": 0.1964,
|
| 975 |
+
"step": 126
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.08592692828146144,
|
| 979 |
+
"grad_norm": 8.870018005371094,
|
| 980 |
+
"learning_rate": 1.7162162162162163e-05,
|
| 981 |
+
"loss": 0.3057,
|
| 982 |
+
"step": 127
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.08660351826792964,
|
| 986 |
+
"grad_norm": 3.8947367668151855,
|
| 987 |
+
"learning_rate": 1.72972972972973e-05,
|
| 988 |
+
"loss": 0.2715,
|
| 989 |
+
"step": 128
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.08728010825439783,
|
| 993 |
+
"grad_norm": 4.829451084136963,
|
| 994 |
+
"learning_rate": 1.7432432432432434e-05,
|
| 995 |
+
"loss": 0.2395,
|
| 996 |
+
"step": 129
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.08795669824086604,
|
| 1000 |
+
"grad_norm": 3.4110400676727295,
|
| 1001 |
+
"learning_rate": 1.756756756756757e-05,
|
| 1002 |
+
"loss": 0.2363,
|
| 1003 |
+
"step": 130
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.08863328822733424,
|
| 1007 |
+
"grad_norm": 3.4218814373016357,
|
| 1008 |
+
"learning_rate": 1.7702702702702702e-05,
|
| 1009 |
+
"loss": 0.2343,
|
| 1010 |
+
"step": 131
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.08930987821380243,
|
| 1014 |
+
"grad_norm": 4.7118425369262695,
|
| 1015 |
+
"learning_rate": 1.783783783783784e-05,
|
| 1016 |
+
"loss": 0.2438,
|
| 1017 |
+
"step": 132
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.08998646820027063,
|
| 1021 |
+
"grad_norm": 5.201712608337402,
|
| 1022 |
+
"learning_rate": 1.7972972972972976e-05,
|
| 1023 |
+
"loss": 0.3213,
|
| 1024 |
+
"step": 133
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.09066305818673884,
|
| 1028 |
+
"grad_norm": 8.192056655883789,
|
| 1029 |
+
"learning_rate": 1.8108108108108108e-05,
|
| 1030 |
+
"loss": 0.3043,
|
| 1031 |
+
"step": 134
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.09133964817320704,
|
| 1035 |
+
"grad_norm": 4.3292694091796875,
|
| 1036 |
+
"learning_rate": 1.8243243243243244e-05,
|
| 1037 |
+
"loss": 0.3077,
|
| 1038 |
+
"step": 135
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.09201623815967523,
|
| 1042 |
+
"grad_norm": 9.865090370178223,
|
| 1043 |
+
"learning_rate": 1.8378378378378383e-05,
|
| 1044 |
+
"loss": 0.2584,
|
| 1045 |
+
"step": 136
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.09269282814614344,
|
| 1049 |
+
"grad_norm": 7.474611759185791,
|
| 1050 |
+
"learning_rate": 1.8513513513513515e-05,
|
| 1051 |
+
"loss": 0.2391,
|
| 1052 |
+
"step": 137
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.09336941813261164,
|
| 1056 |
+
"grad_norm": 3.1332149505615234,
|
| 1057 |
+
"learning_rate": 1.864864864864865e-05,
|
| 1058 |
+
"loss": 0.1335,
|
| 1059 |
+
"step": 138
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.09404600811907984,
|
| 1063 |
+
"grad_norm": 4.493014335632324,
|
| 1064 |
+
"learning_rate": 1.8783783783783786e-05,
|
| 1065 |
+
"loss": 0.2344,
|
| 1066 |
+
"step": 139
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.09472259810554803,
|
| 1070 |
+
"grad_norm": 5.90848970413208,
|
| 1071 |
+
"learning_rate": 1.891891891891892e-05,
|
| 1072 |
+
"loss": 0.2329,
|
| 1073 |
+
"step": 140
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.09472259810554803,
|
| 1077 |
+
"eval_accuracy": 0.7697247706422018,
|
| 1078 |
+
"eval_f1": 0.35475578406169667,
|
| 1079 |
+
"eval_loss": 0.5965576767921448,
|
| 1080 |
+
"eval_precision": 0.7582417582417582,
|
| 1081 |
+
"eval_recall": 0.23154362416107382,
|
| 1082 |
+
"eval_runtime": 51.8498,
|
| 1083 |
+
"eval_samples_per_second": 5.747,
|
| 1084 |
+
"eval_steps_per_second": 0.193,
|
| 1085 |
+
"step": 140
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.09539918809201624,
|
| 1089 |
+
"grad_norm": 3.0857784748077393,
|
| 1090 |
+
"learning_rate": 1.9054054054054057e-05,
|
| 1091 |
+
"loss": 0.1699,
|
| 1092 |
+
"step": 141
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 0.09607577807848444,
|
| 1096 |
+
"grad_norm": 3.591951370239258,
|
| 1097 |
+
"learning_rate": 1.918918918918919e-05,
|
| 1098 |
+
"loss": 0.2166,
|
| 1099 |
+
"step": 142
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.09675236806495263,
|
| 1103 |
+
"grad_norm": 5.5953826904296875,
|
| 1104 |
+
"learning_rate": 1.9324324324324328e-05,
|
| 1105 |
+
"loss": 0.1826,
|
| 1106 |
+
"step": 143
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.09742895805142084,
|
| 1110 |
+
"grad_norm": 4.522704601287842,
|
| 1111 |
+
"learning_rate": 1.9459459459459463e-05,
|
| 1112 |
+
"loss": 0.2288,
|
| 1113 |
+
"step": 144
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.09810554803788904,
|
| 1117 |
+
"grad_norm": 2.501812219619751,
|
| 1118 |
+
"learning_rate": 1.9594594594594595e-05,
|
| 1119 |
+
"loss": 0.1412,
|
| 1120 |
+
"step": 145
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.09878213802435724,
|
| 1124 |
+
"grad_norm": 6.26653528213501,
|
| 1125 |
+
"learning_rate": 1.972972972972973e-05,
|
| 1126 |
+
"loss": 0.2609,
|
| 1127 |
+
"step": 146
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 0.09945872801082543,
|
| 1131 |
+
"grad_norm": 13.06122875213623,
|
| 1132 |
+
"learning_rate": 1.9864864864864866e-05,
|
| 1133 |
+
"loss": 0.3233,
|
| 1134 |
+
"step": 147
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.10013531799729364,
|
| 1138 |
+
"grad_norm": 4.477540493011475,
|
| 1139 |
+
"learning_rate": 2e-05,
|
| 1140 |
+
"loss": 0.2679,
|
| 1141 |
+
"step": 148
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.10081190798376184,
|
| 1145 |
+
"grad_norm": 5.897082328796387,
|
| 1146 |
+
"learning_rate": 1.9999972102437076e-05,
|
| 1147 |
+
"loss": 0.2136,
|
| 1148 |
+
"step": 149
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.10148849797023005,
|
| 1152 |
+
"grad_norm": 4.226516246795654,
|
| 1153 |
+
"learning_rate": 1.9999888409903948e-05,
|
| 1154 |
+
"loss": 0.22,
|
| 1155 |
+
"step": 150
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.10216508795669824,
|
| 1159 |
+
"grad_norm": 6.373837471008301,
|
| 1160 |
+
"learning_rate": 1.9999748922867592e-05,
|
| 1161 |
+
"loss": 0.2117,
|
| 1162 |
+
"step": 151
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.10284167794316644,
|
| 1166 |
+
"grad_norm": 4.057104110717773,
|
| 1167 |
+
"learning_rate": 1.9999553642106267e-05,
|
| 1168 |
+
"loss": 0.2398,
|
| 1169 |
+
"step": 152
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 0.10351826792963464,
|
| 1173 |
+
"grad_norm": 6.765925884246826,
|
| 1174 |
+
"learning_rate": 1.9999302568709548e-05,
|
| 1175 |
+
"loss": 0.2921,
|
| 1176 |
+
"step": 153
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.10419485791610285,
|
| 1180 |
+
"grad_norm": 11.143022537231445,
|
| 1181 |
+
"learning_rate": 1.9998995704078305e-05,
|
| 1182 |
+
"loss": 0.2496,
|
| 1183 |
+
"step": 154
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.10487144790257104,
|
| 1187 |
+
"grad_norm": 7.253014087677002,
|
| 1188 |
+
"learning_rate": 1.9998633049924693e-05,
|
| 1189 |
+
"loss": 0.1869,
|
| 1190 |
+
"step": 155
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.10554803788903924,
|
| 1194 |
+
"grad_norm": 9.102387428283691,
|
| 1195 |
+
"learning_rate": 1.9998214608272136e-05,
|
| 1196 |
+
"loss": 0.2344,
|
| 1197 |
+
"step": 156
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.10622462787550745,
|
| 1201 |
+
"grad_norm": 3.9514195919036865,
|
| 1202 |
+
"learning_rate": 1.9997740381455348e-05,
|
| 1203 |
+
"loss": 0.2364,
|
| 1204 |
+
"step": 157
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.10690121786197564,
|
| 1208 |
+
"grad_norm": 5.509130954742432,
|
| 1209 |
+
"learning_rate": 1.9997210372120276e-05,
|
| 1210 |
+
"loss": 0.2863,
|
| 1211 |
+
"step": 158
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 0.10757780784844384,
|
| 1215 |
+
"grad_norm": 3.954360246658325,
|
| 1216 |
+
"learning_rate": 1.9996624583224112e-05,
|
| 1217 |
+
"loss": 0.1248,
|
| 1218 |
+
"step": 159
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 0.10825439783491204,
|
| 1222 |
+
"grad_norm": 3.0605578422546387,
|
| 1223 |
+
"learning_rate": 1.999598301803528e-05,
|
| 1224 |
+
"loss": 0.1726,
|
| 1225 |
+
"step": 160
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 0.10825439783491204,
|
| 1229 |
+
"eval_accuracy": 0.763302752293578,
|
| 1230 |
+
"eval_f1": 0.31382978723404253,
|
| 1231 |
+
"eval_loss": 0.5946537256240845,
|
| 1232 |
+
"eval_precision": 0.7564102564102564,
|
| 1233 |
+
"eval_recall": 0.19798657718120805,
|
| 1234 |
+
"eval_runtime": 51.9229,
|
| 1235 |
+
"eval_samples_per_second": 5.739,
|
| 1236 |
+
"eval_steps_per_second": 0.193,
|
| 1237 |
+
"step": 160
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.10893098782138025,
|
| 1241 |
+
"grad_norm": 4.9909281730651855,
|
| 1242 |
+
"learning_rate": 1.9995285680133393e-05,
|
| 1243 |
+
"loss": 0.2449,
|
| 1244 |
+
"step": 161
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.10960757780784844,
|
| 1248 |
+
"grad_norm": 2.7885420322418213,
|
| 1249 |
+
"learning_rate": 1.999453257340926e-05,
|
| 1250 |
+
"loss": 0.1239,
|
| 1251 |
+
"step": 162
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.11028416779431664,
|
| 1255 |
+
"grad_norm": 4.381866931915283,
|
| 1256 |
+
"learning_rate": 1.9993723702064852e-05,
|
| 1257 |
+
"loss": 0.146,
|
| 1258 |
+
"step": 163
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 0.11096075778078485,
|
| 1262 |
+
"grad_norm": 7.0832109451293945,
|
| 1263 |
+
"learning_rate": 1.9992859070613275e-05,
|
| 1264 |
+
"loss": 0.2178,
|
| 1265 |
+
"step": 164
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 0.11163734776725305,
|
| 1269 |
+
"grad_norm": 4.502629280090332,
|
| 1270 |
+
"learning_rate": 1.9991938683878746e-05,
|
| 1271 |
+
"loss": 0.2039,
|
| 1272 |
+
"step": 165
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 0.11231393775372124,
|
| 1276 |
+
"grad_norm": 3.46604323387146,
|
| 1277 |
+
"learning_rate": 1.9990962546996583e-05,
|
| 1278 |
+
"loss": 0.1235,
|
| 1279 |
+
"step": 166
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 0.11299052774018944,
|
| 1283 |
+
"grad_norm": 2.314317464828491,
|
| 1284 |
+
"learning_rate": 1.9989930665413148e-05,
|
| 1285 |
+
"loss": 0.1033,
|
| 1286 |
+
"step": 167
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.11366711772665765,
|
| 1290 |
+
"grad_norm": 5.851840019226074,
|
| 1291 |
+
"learning_rate": 1.998884304488584e-05,
|
| 1292 |
+
"loss": 0.2414,
|
| 1293 |
+
"step": 168
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 0.11434370771312584,
|
| 1297 |
+
"grad_norm": 6.2724714279174805,
|
| 1298 |
+
"learning_rate": 1.998769969148305e-05,
|
| 1299 |
+
"loss": 0.2474,
|
| 1300 |
+
"step": 169
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 0.11502029769959404,
|
| 1304 |
+
"grad_norm": 3.0591259002685547,
|
| 1305 |
+
"learning_rate": 1.9986500611584133e-05,
|
| 1306 |
+
"loss": 0.1661,
|
| 1307 |
+
"step": 170
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 0.11569688768606225,
|
| 1311 |
+
"grad_norm": 4.147556304931641,
|
| 1312 |
+
"learning_rate": 1.9985245811879372e-05,
|
| 1313 |
+
"loss": 0.1855,
|
| 1314 |
+
"step": 171
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.11637347767253045,
|
| 1318 |
+
"grad_norm": 4.872109413146973,
|
| 1319 |
+
"learning_rate": 1.9983935299369934e-05,
|
| 1320 |
+
"loss": 0.2505,
|
| 1321 |
+
"step": 172
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.11705006765899864,
|
| 1325 |
+
"grad_norm": 10.929080963134766,
|
| 1326 |
+
"learning_rate": 1.9982569081367844e-05,
|
| 1327 |
+
"loss": 0.238,
|
| 1328 |
+
"step": 173
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.11772665764546685,
|
| 1332 |
+
"grad_norm": 9.166586875915527,
|
| 1333 |
+
"learning_rate": 1.998114716549593e-05,
|
| 1334 |
+
"loss": 0.2415,
|
| 1335 |
+
"step": 174
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 0.11840324763193505,
|
| 1339 |
+
"grad_norm": 4.646167278289795,
|
| 1340 |
+
"learning_rate": 1.997966955968779e-05,
|
| 1341 |
+
"loss": 0.1264,
|
| 1342 |
+
"step": 175
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 0.11907983761840325,
|
| 1346 |
+
"grad_norm": 4.666916847229004,
|
| 1347 |
+
"learning_rate": 1.9978136272187745e-05,
|
| 1348 |
+
"loss": 0.178,
|
| 1349 |
+
"step": 176
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 0.11975642760487144,
|
| 1353 |
+
"grad_norm": 7.303848743438721,
|
| 1354 |
+
"learning_rate": 1.9976547311550796e-05,
|
| 1355 |
+
"loss": 0.2303,
|
| 1356 |
+
"step": 177
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 0.12043301759133965,
|
| 1360 |
+
"grad_norm": 5.617541313171387,
|
| 1361 |
+
"learning_rate": 1.997490268664256e-05,
|
| 1362 |
+
"loss": 0.1295,
|
| 1363 |
+
"step": 178
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 0.12110960757780785,
|
| 1367 |
+
"grad_norm": 7.912723541259766,
|
| 1368 |
+
"learning_rate": 1.9973202406639247e-05,
|
| 1369 |
+
"loss": 0.2137,
|
| 1370 |
+
"step": 179
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.12178619756427606,
|
| 1374 |
+
"grad_norm": 3.9384965896606445,
|
| 1375 |
+
"learning_rate": 1.997144648102759e-05,
|
| 1376 |
+
"loss": 0.1085,
|
| 1377 |
+
"step": 180
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 0.12178619756427606,
|
| 1381 |
+
"eval_accuracy": 0.7678899082568807,
|
| 1382 |
+
"eval_f1": 0.3394255874673629,
|
| 1383 |
+
"eval_loss": 0.558770477771759,
|
| 1384 |
+
"eval_precision": 0.7647058823529411,
|
| 1385 |
+
"eval_recall": 0.2181208053691275,
|
| 1386 |
+
"eval_runtime": 52.2836,
|
| 1387 |
+
"eval_samples_per_second": 5.7,
|
| 1388 |
+
"eval_steps_per_second": 0.191,
|
| 1389 |
+
"step": 180
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.12246278755074425,
|
| 1393 |
+
"grad_norm": 4.896997928619385,
|
| 1394 |
+
"learning_rate": 1.99696349196048e-05,
|
| 1395 |
+
"loss": 0.2525,
|
| 1396 |
+
"step": 181
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.12313937753721245,
|
| 1400 |
+
"grad_norm": 2.2250826358795166,
|
| 1401 |
+
"learning_rate": 1.9967767732478506e-05,
|
| 1402 |
+
"loss": 0.1442,
|
| 1403 |
+
"step": 182
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.12381596752368065,
|
| 1407 |
+
"grad_norm": 5.748762607574463,
|
| 1408 |
+
"learning_rate": 1.99658449300667e-05,
|
| 1409 |
+
"loss": 0.3173,
|
| 1410 |
+
"step": 183
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.12449255751014884,
|
| 1414 |
+
"grad_norm": 3.4051263332366943,
|
| 1415 |
+
"learning_rate": 1.9963866523097683e-05,
|
| 1416 |
+
"loss": 0.2134,
|
| 1417 |
+
"step": 184
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.12516914749661706,
|
| 1421 |
+
"grad_norm": 3.8892011642456055,
|
| 1422 |
+
"learning_rate": 1.9961832522610004e-05,
|
| 1423 |
+
"loss": 0.2136,
|
| 1424 |
+
"step": 185
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.12584573748308525,
|
| 1428 |
+
"grad_norm": 5.042850017547607,
|
| 1429 |
+
"learning_rate": 1.9959742939952393e-05,
|
| 1430 |
+
"loss": 0.1986,
|
| 1431 |
+
"step": 186
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.12652232746955344,
|
| 1435 |
+
"grad_norm": 7.566000461578369,
|
| 1436 |
+
"learning_rate": 1.99575977867837e-05,
|
| 1437 |
+
"loss": 0.2481,
|
| 1438 |
+
"step": 187
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.12719891745602166,
|
| 1442 |
+
"grad_norm": 5.193778991699219,
|
| 1443 |
+
"learning_rate": 1.995539707507284e-05,
|
| 1444 |
+
"loss": 0.2304,
|
| 1445 |
+
"step": 188
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.12787550744248985,
|
| 1449 |
+
"grad_norm": 4.714810371398926,
|
| 1450 |
+
"learning_rate": 1.99531408170987e-05,
|
| 1451 |
+
"loss": 0.2234,
|
| 1452 |
+
"step": 189
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.12855209742895804,
|
| 1456 |
+
"grad_norm": 4.679834842681885,
|
| 1457 |
+
"learning_rate": 1.9950829025450116e-05,
|
| 1458 |
+
"loss": 0.2152,
|
| 1459 |
+
"step": 190
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.12922868741542626,
|
| 1463 |
+
"grad_norm": 2.8689143657684326,
|
| 1464 |
+
"learning_rate": 1.994846171302575e-05,
|
| 1465 |
+
"loss": 0.1938,
|
| 1466 |
+
"step": 191
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.12990527740189445,
|
| 1470 |
+
"grad_norm": 3.1976468563079834,
|
| 1471 |
+
"learning_rate": 1.9946038893034045e-05,
|
| 1472 |
+
"loss": 0.1858,
|
| 1473 |
+
"step": 192
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.13058186738836267,
|
| 1477 |
+
"grad_norm": 3.2573113441467285,
|
| 1478 |
+
"learning_rate": 1.994356057899317e-05,
|
| 1479 |
+
"loss": 0.1333,
|
| 1480 |
+
"step": 193
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.13125845737483086,
|
| 1484 |
+
"grad_norm": 6.062759876251221,
|
| 1485 |
+
"learning_rate": 1.9941026784730898e-05,
|
| 1486 |
+
"loss": 0.2143,
|
| 1487 |
+
"step": 194
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.13193504736129905,
|
| 1491 |
+
"grad_norm": 3.474382162094116,
|
| 1492 |
+
"learning_rate": 1.9938437524384572e-05,
|
| 1493 |
+
"loss": 0.2385,
|
| 1494 |
+
"step": 195
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.13261163734776726,
|
| 1498 |
+
"grad_norm": 4.171142101287842,
|
| 1499 |
+
"learning_rate": 1.9935792812400997e-05,
|
| 1500 |
+
"loss": 0.2212,
|
| 1501 |
+
"step": 196
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.13328822733423545,
|
| 1505 |
+
"grad_norm": 2.72599720954895,
|
| 1506 |
+
"learning_rate": 1.9933092663536384e-05,
|
| 1507 |
+
"loss": 0.1579,
|
| 1508 |
+
"step": 197
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.13396481732070364,
|
| 1512 |
+
"grad_norm": 6.6125102043151855,
|
| 1513 |
+
"learning_rate": 1.9930337092856243e-05,
|
| 1514 |
+
"loss": 0.2187,
|
| 1515 |
+
"step": 198
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.13464140730717186,
|
| 1519 |
+
"grad_norm": 2.2951035499572754,
|
| 1520 |
+
"learning_rate": 1.9927526115735315e-05,
|
| 1521 |
+
"loss": 0.1567,
|
| 1522 |
+
"step": 199
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.13531799729364005,
|
| 1526 |
+
"grad_norm": 4.760623931884766,
|
| 1527 |
+
"learning_rate": 1.9924659747857485e-05,
|
| 1528 |
+
"loss": 0.2326,
|
| 1529 |
+
"step": 200
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.13531799729364005,
|
| 1533 |
+
"eval_accuracy": 0.7623853211009174,
|
| 1534 |
+
"eval_f1": 0.3508771929824561,
|
| 1535 |
+
"eval_loss": 0.5019528865814209,
|
| 1536 |
+
"eval_precision": 0.693069306930693,
|
| 1537 |
+
"eval_recall": 0.2348993288590604,
|
| 1538 |
+
"eval_runtime": 51.9146,
|
| 1539 |
+
"eval_samples_per_second": 5.74,
|
| 1540 |
+
"eval_steps_per_second": 0.193,
|
| 1541 |
+
"step": 200
|
| 1542 |
}
|
| 1543 |
],
|
| 1544 |
"logging_steps": 1,
|
|
|
|
| 1558 |
"attributes": {}
|
| 1559 |
}
|
| 1560 |
},
|
| 1561 |
+
"total_flos": 6.053734571520819e+16,
|
| 1562 |
"train_batch_size": 8,
|
| 1563 |
"trial_name": null,
|
| 1564 |
"trial_params": null
|