Training in progress, step 800, 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 sha256:1780f9c1185f675cd6d51e82c1e80ca4f3e247aa087747947454cf738a5d4c69
|
| 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:f5c1cde5d8aa7aeb944e3e90997bd082e98d4ca837e32246771d3c564f3795b8
|
| 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:b83c11bec463dc59bf896173c919dd90bf0348066e44adb9b0e543295330a1bf
|
| 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:c8fd391b9627b47091cd87960049ac26042dde18861c1d7e32eb8c1118797bb8
|
| 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:fe0b23737c45689a0672974138d2e4bf3250efbe4a01af660359c0407af9bb21
|
| 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 sha256:10e2f0e46255077284fd1723812107259e93b0ddedaac70ce788c07045f96453
|
| 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 sha256:75eb0c742069dd550c1d5ea0a95a5d6f378d2b8b113ac677ef2cdd84c0e57d0b
|
| 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:fcb86dc908dae8fcffc64fbd5f755a25779848603655d0b14f508407e36d14b3
|
| 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 sha256:1bbdc461acddb68778ff93a3106863b49db625d394bc9812c902dbf4a3f0decb
|
| 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 sha256:84aacbbc43943b2ed2d3eeb03d889744c56f76845a7d96e49fa7bff5da5be3b6
|
| 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:fa47b6b07773b1a1fc22394cf526cb8aba69b558d0c5e9d7eeb7487fded1884b
|
| 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:ae426b79fc7504b3415de1b203e94c3df18eeaa89bc726cbd2ff1e2fdd6c95ce
|
| 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 sha256:8412f5078d5372d361046df0c90d05d473909fc8ab1aa58766d872c8114b5c70
|
| 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,
|
|
@@ -5339,6 +5339,766 @@
|
|
| 5339 |
"eval_samples_per_second": 5.866,
|
| 5340 |
"eval_steps_per_second": 0.201,
|
| 5341 |
"step": 700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5342 |
}
|
| 5343 |
],
|
| 5344 |
"logging_steps": 1,
|
|
@@ -5358,7 +6118,7 @@
|
|
| 5358 |
"attributes": {}
|
| 5359 |
}
|
| 5360 |
},
|
| 5361 |
-
"total_flos": 2.
|
| 5362 |
"train_batch_size": 8,
|
| 5363 |
"trial_name": null,
|
| 5364 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.6779661016949152,
|
| 5 |
"eval_steps": 20,
|
| 6 |
+
"global_step": 800,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 5339 |
"eval_samples_per_second": 5.866,
|
| 5340 |
"eval_steps_per_second": 0.201,
|
| 5341 |
"step": 700
|
| 5342 |
+
},
|
| 5343 |
+
{
|
| 5344 |
+
"epoch": 0.5940677966101695,
|
| 5345 |
+
"grad_norm": 1.3996844291687012,
|
| 5346 |
+
"learning_rate": 8.46780298998537e-06,
|
| 5347 |
+
"loss": 0.0051,
|
| 5348 |
+
"step": 701
|
| 5349 |
+
},
|
| 5350 |
+
{
|
| 5351 |
+
"epoch": 0.5949152542372881,
|
| 5352 |
+
"grad_norm": 1.469482421875,
|
| 5353 |
+
"learning_rate": 8.438577183006448e-06,
|
| 5354 |
+
"loss": 0.0096,
|
| 5355 |
+
"step": 702
|
| 5356 |
+
},
|
| 5357 |
+
{
|
| 5358 |
+
"epoch": 0.5957627118644068,
|
| 5359 |
+
"grad_norm": 0.796106219291687,
|
| 5360 |
+
"learning_rate": 8.409365039809282e-06,
|
| 5361 |
+
"loss": 0.006,
|
| 5362 |
+
"step": 703
|
| 5363 |
+
},
|
| 5364 |
+
{
|
| 5365 |
+
"epoch": 0.5966101694915255,
|
| 5366 |
+
"grad_norm": 1.5384269952774048,
|
| 5367 |
+
"learning_rate": 8.380166816025043e-06,
|
| 5368 |
+
"loss": 0.0144,
|
| 5369 |
+
"step": 704
|
| 5370 |
+
},
|
| 5371 |
+
{
|
| 5372 |
+
"epoch": 0.597457627118644,
|
| 5373 |
+
"grad_norm": 2.1261978149414062,
|
| 5374 |
+
"learning_rate": 8.350982767163105e-06,
|
| 5375 |
+
"loss": 0.0185,
|
| 5376 |
+
"step": 705
|
| 5377 |
+
},
|
| 5378 |
+
{
|
| 5379 |
+
"epoch": 0.5983050847457627,
|
| 5380 |
+
"grad_norm": 2.251471757888794,
|
| 5381 |
+
"learning_rate": 8.321813148608784e-06,
|
| 5382 |
+
"loss": 0.0158,
|
| 5383 |
+
"step": 706
|
| 5384 |
+
},
|
| 5385 |
+
{
|
| 5386 |
+
"epoch": 0.5991525423728814,
|
| 5387 |
+
"grad_norm": 0.6852215528488159,
|
| 5388 |
+
"learning_rate": 8.292658215621139e-06,
|
| 5389 |
+
"loss": 0.0025,
|
| 5390 |
+
"step": 707
|
| 5391 |
+
},
|
| 5392 |
+
{
|
| 5393 |
+
"epoch": 0.6,
|
| 5394 |
+
"grad_norm": 0.9804059863090515,
|
| 5395 |
+
"learning_rate": 8.263518223330698e-06,
|
| 5396 |
+
"loss": 0.0028,
|
| 5397 |
+
"step": 708
|
| 5398 |
+
},
|
| 5399 |
+
{
|
| 5400 |
+
"epoch": 0.6008474576271187,
|
| 5401 |
+
"grad_norm": 1.716261863708496,
|
| 5402 |
+
"learning_rate": 8.234393426737258e-06,
|
| 5403 |
+
"loss": 0.0095,
|
| 5404 |
+
"step": 709
|
| 5405 |
+
},
|
| 5406 |
+
{
|
| 5407 |
+
"epoch": 0.6016949152542372,
|
| 5408 |
+
"grad_norm": 2.618445634841919,
|
| 5409 |
+
"learning_rate": 8.205284080707634e-06,
|
| 5410 |
+
"loss": 0.0172,
|
| 5411 |
+
"step": 710
|
| 5412 |
+
},
|
| 5413 |
+
{
|
| 5414 |
+
"epoch": 0.6025423728813559,
|
| 5415 |
+
"grad_norm": 0.9949232935905457,
|
| 5416 |
+
"learning_rate": 8.176190439973442e-06,
|
| 5417 |
+
"loss": 0.0044,
|
| 5418 |
+
"step": 711
|
| 5419 |
+
},
|
| 5420 |
+
{
|
| 5421 |
+
"epoch": 0.6033898305084746,
|
| 5422 |
+
"grad_norm": 0.9794982075691223,
|
| 5423 |
+
"learning_rate": 8.147112759128859e-06,
|
| 5424 |
+
"loss": 0.0046,
|
| 5425 |
+
"step": 712
|
| 5426 |
+
},
|
| 5427 |
+
{
|
| 5428 |
+
"epoch": 0.6042372881355932,
|
| 5429 |
+
"grad_norm": 1.576850414276123,
|
| 5430 |
+
"learning_rate": 8.118051292628394e-06,
|
| 5431 |
+
"loss": 0.0127,
|
| 5432 |
+
"step": 713
|
| 5433 |
+
},
|
| 5434 |
+
{
|
| 5435 |
+
"epoch": 0.6050847457627119,
|
| 5436 |
+
"grad_norm": 1.6285077333450317,
|
| 5437 |
+
"learning_rate": 8.089006294784679e-06,
|
| 5438 |
+
"loss": 0.0077,
|
| 5439 |
+
"step": 714
|
| 5440 |
+
},
|
| 5441 |
+
{
|
| 5442 |
+
"epoch": 0.6059322033898306,
|
| 5443 |
+
"grad_norm": 2.055169105529785,
|
| 5444 |
+
"learning_rate": 8.059978019766219e-06,
|
| 5445 |
+
"loss": 0.0078,
|
| 5446 |
+
"step": 715
|
| 5447 |
+
},
|
| 5448 |
+
{
|
| 5449 |
+
"epoch": 0.6067796610169491,
|
| 5450 |
+
"grad_norm": 1.1716010570526123,
|
| 5451 |
+
"learning_rate": 8.030966721595186e-06,
|
| 5452 |
+
"loss": 0.0042,
|
| 5453 |
+
"step": 716
|
| 5454 |
+
},
|
| 5455 |
+
{
|
| 5456 |
+
"epoch": 0.6076271186440678,
|
| 5457 |
+
"grad_norm": 0.2800714075565338,
|
| 5458 |
+
"learning_rate": 8.001972654145194e-06,
|
| 5459 |
+
"loss": 0.0012,
|
| 5460 |
+
"step": 717
|
| 5461 |
+
},
|
| 5462 |
+
{
|
| 5463 |
+
"epoch": 0.6084745762711864,
|
| 5464 |
+
"grad_norm": 1.2018545866012573,
|
| 5465 |
+
"learning_rate": 7.972996071139065e-06,
|
| 5466 |
+
"loss": 0.009,
|
| 5467 |
+
"step": 718
|
| 5468 |
+
},
|
| 5469 |
+
{
|
| 5470 |
+
"epoch": 0.6093220338983051,
|
| 5471 |
+
"grad_norm": 3.2295126914978027,
|
| 5472 |
+
"learning_rate": 7.944037226146623e-06,
|
| 5473 |
+
"loss": 0.0188,
|
| 5474 |
+
"step": 719
|
| 5475 |
+
},
|
| 5476 |
+
{
|
| 5477 |
+
"epoch": 0.6101694915254238,
|
| 5478 |
+
"grad_norm": 2.7130823135375977,
|
| 5479 |
+
"learning_rate": 7.915096372582467e-06,
|
| 5480 |
+
"loss": 0.014,
|
| 5481 |
+
"step": 720
|
| 5482 |
+
},
|
| 5483 |
+
{
|
| 5484 |
+
"epoch": 0.6101694915254238,
|
| 5485 |
+
"eval_accuracy": 1.0,
|
| 5486 |
+
"eval_f1": 1.0,
|
| 5487 |
+
"eval_loss": 9.377366950502619e-05,
|
| 5488 |
+
"eval_precision": 1.0,
|
| 5489 |
+
"eval_recall": 1.0,
|
| 5490 |
+
"eval_runtime": 51.3568,
|
| 5491 |
+
"eval_samples_per_second": 5.686,
|
| 5492 |
+
"eval_steps_per_second": 0.195,
|
| 5493 |
+
"step": 720
|
| 5494 |
+
},
|
| 5495 |
+
{
|
| 5496 |
+
"epoch": 0.6110169491525423,
|
| 5497 |
+
"grad_norm": 0.14015014469623566,
|
| 5498 |
+
"learning_rate": 7.886173763703757e-06,
|
| 5499 |
+
"loss": 0.0007,
|
| 5500 |
+
"step": 721
|
| 5501 |
+
},
|
| 5502 |
+
{
|
| 5503 |
+
"epoch": 0.611864406779661,
|
| 5504 |
+
"grad_norm": 1.3559490442276,
|
| 5505 |
+
"learning_rate": 7.857269652607995e-06,
|
| 5506 |
+
"loss": 0.005,
|
| 5507 |
+
"step": 722
|
| 5508 |
+
},
|
| 5509 |
+
{
|
| 5510 |
+
"epoch": 0.6127118644067797,
|
| 5511 |
+
"grad_norm": 1.6353223323822021,
|
| 5512 |
+
"learning_rate": 7.828384292230814e-06,
|
| 5513 |
+
"loss": 0.0081,
|
| 5514 |
+
"step": 723
|
| 5515 |
+
},
|
| 5516 |
+
{
|
| 5517 |
+
"epoch": 0.6135593220338983,
|
| 5518 |
+
"grad_norm": 1.1183819770812988,
|
| 5519 |
+
"learning_rate": 7.799517935343757e-06,
|
| 5520 |
+
"loss": 0.0034,
|
| 5521 |
+
"step": 724
|
| 5522 |
+
},
|
| 5523 |
+
{
|
| 5524 |
+
"epoch": 0.614406779661017,
|
| 5525 |
+
"grad_norm": 1.9186286926269531,
|
| 5526 |
+
"learning_rate": 7.770670834552074e-06,
|
| 5527 |
+
"loss": 0.0153,
|
| 5528 |
+
"step": 725
|
| 5529 |
+
},
|
| 5530 |
+
{
|
| 5531 |
+
"epoch": 0.6152542372881356,
|
| 5532 |
+
"grad_norm": 1.465635895729065,
|
| 5533 |
+
"learning_rate": 7.741843242292508e-06,
|
| 5534 |
+
"loss": 0.0108,
|
| 5535 |
+
"step": 726
|
| 5536 |
+
},
|
| 5537 |
+
{
|
| 5538 |
+
"epoch": 0.6161016949152542,
|
| 5539 |
+
"grad_norm": 1.6631442308425903,
|
| 5540 |
+
"learning_rate": 7.713035410831086e-06,
|
| 5541 |
+
"loss": 0.0155,
|
| 5542 |
+
"step": 727
|
| 5543 |
+
},
|
| 5544 |
+
{
|
| 5545 |
+
"epoch": 0.6169491525423729,
|
| 5546 |
+
"grad_norm": 0.37416231632232666,
|
| 5547 |
+
"learning_rate": 7.68424759226091e-06,
|
| 5548 |
+
"loss": 0.0011,
|
| 5549 |
+
"step": 728
|
| 5550 |
+
},
|
| 5551 |
+
{
|
| 5552 |
+
"epoch": 0.6177966101694915,
|
| 5553 |
+
"grad_norm": 1.5809506177902222,
|
| 5554 |
+
"learning_rate": 7.655480038499953e-06,
|
| 5555 |
+
"loss": 0.0071,
|
| 5556 |
+
"step": 729
|
| 5557 |
+
},
|
| 5558 |
+
{
|
| 5559 |
+
"epoch": 0.6186440677966102,
|
| 5560 |
+
"grad_norm": 1.0575535297393799,
|
| 5561 |
+
"learning_rate": 7.626733001288852e-06,
|
| 5562 |
+
"loss": 0.003,
|
| 5563 |
+
"step": 730
|
| 5564 |
+
},
|
| 5565 |
+
{
|
| 5566 |
+
"epoch": 0.6194915254237288,
|
| 5567 |
+
"grad_norm": 1.1455744504928589,
|
| 5568 |
+
"learning_rate": 7.598006732188708e-06,
|
| 5569 |
+
"loss": 0.007,
|
| 5570 |
+
"step": 731
|
| 5571 |
+
},
|
| 5572 |
+
{
|
| 5573 |
+
"epoch": 0.6203389830508474,
|
| 5574 |
+
"grad_norm": 1.1559722423553467,
|
| 5575 |
+
"learning_rate": 7.569301482578885e-06,
|
| 5576 |
+
"loss": 0.0043,
|
| 5577 |
+
"step": 732
|
| 5578 |
+
},
|
| 5579 |
+
{
|
| 5580 |
+
"epoch": 0.6211864406779661,
|
| 5581 |
+
"grad_norm": 4.226150035858154,
|
| 5582 |
+
"learning_rate": 7.540617503654804e-06,
|
| 5583 |
+
"loss": 0.0244,
|
| 5584 |
+
"step": 733
|
| 5585 |
+
},
|
| 5586 |
+
{
|
| 5587 |
+
"epoch": 0.6220338983050847,
|
| 5588 |
+
"grad_norm": 1.66443932056427,
|
| 5589 |
+
"learning_rate": 7.511955046425754e-06,
|
| 5590 |
+
"loss": 0.0073,
|
| 5591 |
+
"step": 734
|
| 5592 |
+
},
|
| 5593 |
+
{
|
| 5594 |
+
"epoch": 0.6228813559322034,
|
| 5595 |
+
"grad_norm": 4.203097820281982,
|
| 5596 |
+
"learning_rate": 7.483314361712686e-06,
|
| 5597 |
+
"loss": 0.0236,
|
| 5598 |
+
"step": 735
|
| 5599 |
+
},
|
| 5600 |
+
{
|
| 5601 |
+
"epoch": 0.6237288135593221,
|
| 5602 |
+
"grad_norm": 3.4578402042388916,
|
| 5603 |
+
"learning_rate": 7.454695700146024e-06,
|
| 5604 |
+
"loss": 0.0224,
|
| 5605 |
+
"step": 736
|
| 5606 |
+
},
|
| 5607 |
+
{
|
| 5608 |
+
"epoch": 0.6245762711864407,
|
| 5609 |
+
"grad_norm": 2.495051383972168,
|
| 5610 |
+
"learning_rate": 7.42609931216348e-06,
|
| 5611 |
+
"loss": 0.0303,
|
| 5612 |
+
"step": 737
|
| 5613 |
+
},
|
| 5614 |
+
{
|
| 5615 |
+
"epoch": 0.6254237288135593,
|
| 5616 |
+
"grad_norm": 0.11093101650476456,
|
| 5617 |
+
"learning_rate": 7.397525448007838e-06,
|
| 5618 |
+
"loss": 0.0005,
|
| 5619 |
+
"step": 738
|
| 5620 |
+
},
|
| 5621 |
+
{
|
| 5622 |
+
"epoch": 0.6262711864406779,
|
| 5623 |
+
"grad_norm": 0.8444148302078247,
|
| 5624 |
+
"learning_rate": 7.368974357724789e-06,
|
| 5625 |
+
"loss": 0.0038,
|
| 5626 |
+
"step": 739
|
| 5627 |
+
},
|
| 5628 |
+
{
|
| 5629 |
+
"epoch": 0.6271186440677966,
|
| 5630 |
+
"grad_norm": 0.6768697500228882,
|
| 5631 |
+
"learning_rate": 7.3404462911607325e-06,
|
| 5632 |
+
"loss": 0.0042,
|
| 5633 |
+
"step": 740
|
| 5634 |
+
},
|
| 5635 |
+
{
|
| 5636 |
+
"epoch": 0.6271186440677966,
|
| 5637 |
+
"eval_accuracy": 1.0,
|
| 5638 |
+
"eval_f1": 1.0,
|
| 5639 |
+
"eval_loss": 9.098863665712997e-05,
|
| 5640 |
+
"eval_precision": 1.0,
|
| 5641 |
+
"eval_recall": 1.0,
|
| 5642 |
+
"eval_runtime": 51.0217,
|
| 5643 |
+
"eval_samples_per_second": 5.723,
|
| 5644 |
+
"eval_steps_per_second": 0.196,
|
| 5645 |
+
"step": 740
|
| 5646 |
+
},
|
| 5647 |
+
{
|
| 5648 |
+
"epoch": 0.6279661016949153,
|
| 5649 |
+
"grad_norm": 0.5370987057685852,
|
| 5650 |
+
"learning_rate": 7.3119414979605865e-06,
|
| 5651 |
+
"loss": 0.0043,
|
| 5652 |
+
"step": 741
|
| 5653 |
+
},
|
| 5654 |
+
{
|
| 5655 |
+
"epoch": 0.6288135593220339,
|
| 5656 |
+
"grad_norm": 2.2939460277557373,
|
| 5657 |
+
"learning_rate": 7.283460227565614e-06,
|
| 5658 |
+
"loss": 0.0163,
|
| 5659 |
+
"step": 742
|
| 5660 |
+
},
|
| 5661 |
+
{
|
| 5662 |
+
"epoch": 0.6296610169491526,
|
| 5663 |
+
"grad_norm": 0.1840486228466034,
|
| 5664 |
+
"learning_rate": 7.255002729211227e-06,
|
| 5665 |
+
"loss": 0.0014,
|
| 5666 |
+
"step": 743
|
| 5667 |
+
},
|
| 5668 |
+
{
|
| 5669 |
+
"epoch": 0.6305084745762712,
|
| 5670 |
+
"grad_norm": 2.645730495452881,
|
| 5671 |
+
"learning_rate": 7.226569251924811e-06,
|
| 5672 |
+
"loss": 0.0305,
|
| 5673 |
+
"step": 744
|
| 5674 |
+
},
|
| 5675 |
+
{
|
| 5676 |
+
"epoch": 0.6313559322033898,
|
| 5677 |
+
"grad_norm": 2.7763075828552246,
|
| 5678 |
+
"learning_rate": 7.198160044523555e-06,
|
| 5679 |
+
"loss": 0.0297,
|
| 5680 |
+
"step": 745
|
| 5681 |
+
},
|
| 5682 |
+
{
|
| 5683 |
+
"epoch": 0.6322033898305085,
|
| 5684 |
+
"grad_norm": 1.471494436264038,
|
| 5685 |
+
"learning_rate": 7.169775355612261e-06,
|
| 5686 |
+
"loss": 0.0038,
|
| 5687 |
+
"step": 746
|
| 5688 |
+
},
|
| 5689 |
+
{
|
| 5690 |
+
"epoch": 0.6330508474576271,
|
| 5691 |
+
"grad_norm": 1.1910088062286377,
|
| 5692 |
+
"learning_rate": 7.141415433581169e-06,
|
| 5693 |
+
"loss": 0.0051,
|
| 5694 |
+
"step": 747
|
| 5695 |
+
},
|
| 5696 |
+
{
|
| 5697 |
+
"epoch": 0.6338983050847458,
|
| 5698 |
+
"grad_norm": 0.6888835430145264,
|
| 5699 |
+
"learning_rate": 7.113080526603793e-06,
|
| 5700 |
+
"loss": 0.0018,
|
| 5701 |
+
"step": 748
|
| 5702 |
+
},
|
| 5703 |
+
{
|
| 5704 |
+
"epoch": 0.6347457627118644,
|
| 5705 |
+
"grad_norm": 0.5418266654014587,
|
| 5706 |
+
"learning_rate": 7.0847708826347415e-06,
|
| 5707 |
+
"loss": 0.0023,
|
| 5708 |
+
"step": 749
|
| 5709 |
+
},
|
| 5710 |
+
{
|
| 5711 |
+
"epoch": 0.635593220338983,
|
| 5712 |
+
"grad_norm": 0.5402665138244629,
|
| 5713 |
+
"learning_rate": 7.056486749407552e-06,
|
| 5714 |
+
"loss": 0.0013,
|
| 5715 |
+
"step": 750
|
| 5716 |
+
},
|
| 5717 |
+
{
|
| 5718 |
+
"epoch": 0.6364406779661017,
|
| 5719 |
+
"grad_norm": 0.44419825077056885,
|
| 5720 |
+
"learning_rate": 7.028228374432523e-06,
|
| 5721 |
+
"loss": 0.0018,
|
| 5722 |
+
"step": 751
|
| 5723 |
+
},
|
| 5724 |
+
{
|
| 5725 |
+
"epoch": 0.6372881355932203,
|
| 5726 |
+
"grad_norm": 2.0415663719177246,
|
| 5727 |
+
"learning_rate": 6.9999960049945406e-06,
|
| 5728 |
+
"loss": 0.0136,
|
| 5729 |
+
"step": 752
|
| 5730 |
+
},
|
| 5731 |
+
{
|
| 5732 |
+
"epoch": 0.638135593220339,
|
| 5733 |
+
"grad_norm": 1.5157843828201294,
|
| 5734 |
+
"learning_rate": 6.971789888150926e-06,
|
| 5735 |
+
"loss": 0.0225,
|
| 5736 |
+
"step": 753
|
| 5737 |
+
},
|
| 5738 |
+
{
|
| 5739 |
+
"epoch": 0.6389830508474577,
|
| 5740 |
+
"grad_norm": 0.30365321040153503,
|
| 5741 |
+
"learning_rate": 6.943610270729261e-06,
|
| 5742 |
+
"loss": 0.0013,
|
| 5743 |
+
"step": 754
|
| 5744 |
+
},
|
| 5745 |
+
{
|
| 5746 |
+
"epoch": 0.6398305084745762,
|
| 5747 |
+
"grad_norm": 1.8797694444656372,
|
| 5748 |
+
"learning_rate": 6.915457399325243e-06,
|
| 5749 |
+
"loss": 0.0126,
|
| 5750 |
+
"step": 755
|
| 5751 |
+
},
|
| 5752 |
+
{
|
| 5753 |
+
"epoch": 0.6406779661016949,
|
| 5754 |
+
"grad_norm": 1.3082208633422852,
|
| 5755 |
+
"learning_rate": 6.8873315203005155e-06,
|
| 5756 |
+
"loss": 0.0186,
|
| 5757 |
+
"step": 756
|
| 5758 |
+
},
|
| 5759 |
+
{
|
| 5760 |
+
"epoch": 0.6415254237288136,
|
| 5761 |
+
"grad_norm": 0.4574929475784302,
|
| 5762 |
+
"learning_rate": 6.859232879780515e-06,
|
| 5763 |
+
"loss": 0.0033,
|
| 5764 |
+
"step": 757
|
| 5765 |
+
},
|
| 5766 |
+
{
|
| 5767 |
+
"epoch": 0.6423728813559322,
|
| 5768 |
+
"grad_norm": 1.4100733995437622,
|
| 5769 |
+
"learning_rate": 6.831161723652313e-06,
|
| 5770 |
+
"loss": 0.0054,
|
| 5771 |
+
"step": 758
|
| 5772 |
+
},
|
| 5773 |
+
{
|
| 5774 |
+
"epoch": 0.6432203389830509,
|
| 5775 |
+
"grad_norm": 0.8795613646507263,
|
| 5776 |
+
"learning_rate": 6.803118297562479e-06,
|
| 5777 |
+
"loss": 0.006,
|
| 5778 |
+
"step": 759
|
| 5779 |
+
},
|
| 5780 |
+
{
|
| 5781 |
+
"epoch": 0.6440677966101694,
|
| 5782 |
+
"grad_norm": 3.974163293838501,
|
| 5783 |
+
"learning_rate": 6.775102846914912e-06,
|
| 5784 |
+
"loss": 0.0418,
|
| 5785 |
+
"step": 760
|
| 5786 |
+
},
|
| 5787 |
+
{
|
| 5788 |
+
"epoch": 0.6440677966101694,
|
| 5789 |
+
"eval_accuracy": 1.0,
|
| 5790 |
+
"eval_f1": 1.0,
|
| 5791 |
+
"eval_loss": 0.0002802866802085191,
|
| 5792 |
+
"eval_precision": 1.0,
|
| 5793 |
+
"eval_recall": 1.0,
|
| 5794 |
+
"eval_runtime": 50.7257,
|
| 5795 |
+
"eval_samples_per_second": 5.756,
|
| 5796 |
+
"eval_steps_per_second": 0.197,
|
| 5797 |
+
"step": 760
|
| 5798 |
+
},
|
| 5799 |
+
{
|
| 5800 |
+
"epoch": 0.6449152542372881,
|
| 5801 |
+
"grad_norm": 2.1430978775024414,
|
| 5802 |
+
"learning_rate": 6.747115616868711e-06,
|
| 5803 |
+
"loss": 0.0171,
|
| 5804 |
+
"step": 761
|
| 5805 |
+
},
|
| 5806 |
+
{
|
| 5807 |
+
"epoch": 0.6457627118644068,
|
| 5808 |
+
"grad_norm": 0.2186567485332489,
|
| 5809 |
+
"learning_rate": 6.719156852336015e-06,
|
| 5810 |
+
"loss": 0.0009,
|
| 5811 |
+
"step": 762
|
| 5812 |
+
},
|
| 5813 |
+
{
|
| 5814 |
+
"epoch": 0.6466101694915254,
|
| 5815 |
+
"grad_norm": 2.305494546890259,
|
| 5816 |
+
"learning_rate": 6.691226797979864e-06,
|
| 5817 |
+
"loss": 0.0223,
|
| 5818 |
+
"step": 763
|
| 5819 |
+
},
|
| 5820 |
+
{
|
| 5821 |
+
"epoch": 0.6474576271186441,
|
| 5822 |
+
"grad_norm": 0.28418537974357605,
|
| 5823 |
+
"learning_rate": 6.66332569821206e-06,
|
| 5824 |
+
"loss": 0.0013,
|
| 5825 |
+
"step": 764
|
| 5826 |
+
},
|
| 5827 |
+
{
|
| 5828 |
+
"epoch": 0.6483050847457628,
|
| 5829 |
+
"grad_norm": 0.9447827935218811,
|
| 5830 |
+
"learning_rate": 6.635453797191034e-06,
|
| 5831 |
+
"loss": 0.0048,
|
| 5832 |
+
"step": 765
|
| 5833 |
+
},
|
| 5834 |
+
{
|
| 5835 |
+
"epoch": 0.6491525423728813,
|
| 5836 |
+
"grad_norm": 1.4569226503372192,
|
| 5837 |
+
"learning_rate": 6.607611338819697e-06,
|
| 5838 |
+
"loss": 0.0032,
|
| 5839 |
+
"step": 766
|
| 5840 |
+
},
|
| 5841 |
+
{
|
| 5842 |
+
"epoch": 0.65,
|
| 5843 |
+
"grad_norm": 0.2803967297077179,
|
| 5844 |
+
"learning_rate": 6.579798566743314e-06,
|
| 5845 |
+
"loss": 0.0011,
|
| 5846 |
+
"step": 767
|
| 5847 |
+
},
|
| 5848 |
+
{
|
| 5849 |
+
"epoch": 0.6508474576271186,
|
| 5850 |
+
"grad_norm": 1.287703037261963,
|
| 5851 |
+
"learning_rate": 6.552015724347367e-06,
|
| 5852 |
+
"loss": 0.0057,
|
| 5853 |
+
"step": 768
|
| 5854 |
+
},
|
| 5855 |
+
{
|
| 5856 |
+
"epoch": 0.6516949152542373,
|
| 5857 |
+
"grad_norm": 1.7164057493209839,
|
| 5858 |
+
"learning_rate": 6.52426305475543e-06,
|
| 5859 |
+
"loss": 0.0123,
|
| 5860 |
+
"step": 769
|
| 5861 |
+
},
|
| 5862 |
+
{
|
| 5863 |
+
"epoch": 0.652542372881356,
|
| 5864 |
+
"grad_norm": 2.8528084754943848,
|
| 5865 |
+
"learning_rate": 6.4965408008270355e-06,
|
| 5866 |
+
"loss": 0.0127,
|
| 5867 |
+
"step": 770
|
| 5868 |
+
},
|
| 5869 |
+
{
|
| 5870 |
+
"epoch": 0.6533898305084745,
|
| 5871 |
+
"grad_norm": 2.0749473571777344,
|
| 5872 |
+
"learning_rate": 6.4688492051555604e-06,
|
| 5873 |
+
"loss": 0.0145,
|
| 5874 |
+
"step": 771
|
| 5875 |
+
},
|
| 5876 |
+
{
|
| 5877 |
+
"epoch": 0.6542372881355932,
|
| 5878 |
+
"grad_norm": 3.287358522415161,
|
| 5879 |
+
"learning_rate": 6.441188510066092e-06,
|
| 5880 |
+
"loss": 0.0254,
|
| 5881 |
+
"step": 772
|
| 5882 |
+
},
|
| 5883 |
+
{
|
| 5884 |
+
"epoch": 0.6550847457627119,
|
| 5885 |
+
"grad_norm": 0.4311177432537079,
|
| 5886 |
+
"learning_rate": 6.413558957613309e-06,
|
| 5887 |
+
"loss": 0.0012,
|
| 5888 |
+
"step": 773
|
| 5889 |
+
},
|
| 5890 |
+
{
|
| 5891 |
+
"epoch": 0.6559322033898305,
|
| 5892 |
+
"grad_norm": 1.6251966953277588,
|
| 5893 |
+
"learning_rate": 6.385960789579364e-06,
|
| 5894 |
+
"loss": 0.008,
|
| 5895 |
+
"step": 774
|
| 5896 |
+
},
|
| 5897 |
+
{
|
| 5898 |
+
"epoch": 0.6567796610169492,
|
| 5899 |
+
"grad_norm": 2.233088493347168,
|
| 5900 |
+
"learning_rate": 6.358394247471779e-06,
|
| 5901 |
+
"loss": 0.0203,
|
| 5902 |
+
"step": 775
|
| 5903 |
+
},
|
| 5904 |
+
{
|
| 5905 |
+
"epoch": 0.6576271186440678,
|
| 5906 |
+
"grad_norm": 0.6515169143676758,
|
| 5907 |
+
"learning_rate": 6.330859572521311e-06,
|
| 5908 |
+
"loss": 0.0032,
|
| 5909 |
+
"step": 776
|
| 5910 |
+
},
|
| 5911 |
+
{
|
| 5912 |
+
"epoch": 0.6584745762711864,
|
| 5913 |
+
"grad_norm": 0.9649152159690857,
|
| 5914 |
+
"learning_rate": 6.303357005679858e-06,
|
| 5915 |
+
"loss": 0.0057,
|
| 5916 |
+
"step": 777
|
| 5917 |
+
},
|
| 5918 |
+
{
|
| 5919 |
+
"epoch": 0.6593220338983051,
|
| 5920 |
+
"grad_norm": 0.40915530920028687,
|
| 5921 |
+
"learning_rate": 6.275886787618339e-06,
|
| 5922 |
+
"loss": 0.0017,
|
| 5923 |
+
"step": 778
|
| 5924 |
+
},
|
| 5925 |
+
{
|
| 5926 |
+
"epoch": 0.6601694915254237,
|
| 5927 |
+
"grad_norm": 2.0900630950927734,
|
| 5928 |
+
"learning_rate": 6.2484491587246096e-06,
|
| 5929 |
+
"loss": 0.0091,
|
| 5930 |
+
"step": 779
|
| 5931 |
+
},
|
| 5932 |
+
{
|
| 5933 |
+
"epoch": 0.6610169491525424,
|
| 5934 |
+
"grad_norm": 0.4685976803302765,
|
| 5935 |
+
"learning_rate": 6.221044359101317e-06,
|
| 5936 |
+
"loss": 0.0024,
|
| 5937 |
+
"step": 780
|
| 5938 |
+
},
|
| 5939 |
+
{
|
| 5940 |
+
"epoch": 0.6610169491525424,
|
| 5941 |
+
"eval_accuracy": 1.0,
|
| 5942 |
+
"eval_f1": 1.0,
|
| 5943 |
+
"eval_loss": 0.00018300920783076435,
|
| 5944 |
+
"eval_precision": 1.0,
|
| 5945 |
+
"eval_recall": 1.0,
|
| 5946 |
+
"eval_runtime": 50.7091,
|
| 5947 |
+
"eval_samples_per_second": 5.758,
|
| 5948 |
+
"eval_steps_per_second": 0.197,
|
| 5949 |
+
"step": 780
|
| 5950 |
+
},
|
| 5951 |
+
{
|
| 5952 |
+
"epoch": 0.661864406779661,
|
| 5953 |
+
"grad_norm": 0.4373491108417511,
|
| 5954 |
+
"learning_rate": 6.193672628563845e-06,
|
| 5955 |
+
"loss": 0.0016,
|
| 5956 |
+
"step": 781
|
| 5957 |
+
},
|
| 5958 |
+
{
|
| 5959 |
+
"epoch": 0.6627118644067796,
|
| 5960 |
+
"grad_norm": 0.18425188958644867,
|
| 5961 |
+
"learning_rate": 6.166334206638186e-06,
|
| 5962 |
+
"loss": 0.0007,
|
| 5963 |
+
"step": 782
|
| 5964 |
+
},
|
| 5965 |
+
{
|
| 5966 |
+
"epoch": 0.6635593220338983,
|
| 5967 |
+
"grad_norm": 1.5032649040222168,
|
| 5968 |
+
"learning_rate": 6.139029332558855e-06,
|
| 5969 |
+
"loss": 0.0087,
|
| 5970 |
+
"step": 783
|
| 5971 |
+
},
|
| 5972 |
+
{
|
| 5973 |
+
"epoch": 0.6644067796610169,
|
| 5974 |
+
"grad_norm": 1.4531553983688354,
|
| 5975 |
+
"learning_rate": 6.111758245266795e-06,
|
| 5976 |
+
"loss": 0.0055,
|
| 5977 |
+
"step": 784
|
| 5978 |
+
},
|
| 5979 |
+
{
|
| 5980 |
+
"epoch": 0.6652542372881356,
|
| 5981 |
+
"grad_norm": 1.0665950775146484,
|
| 5982 |
+
"learning_rate": 6.084521183407286e-06,
|
| 5983 |
+
"loss": 0.0075,
|
| 5984 |
+
"step": 785
|
| 5985 |
+
},
|
| 5986 |
+
{
|
| 5987 |
+
"epoch": 0.6661016949152543,
|
| 5988 |
+
"grad_norm": 2.243398666381836,
|
| 5989 |
+
"learning_rate": 6.057318385327857e-06,
|
| 5990 |
+
"loss": 0.0172,
|
| 5991 |
+
"step": 786
|
| 5992 |
+
},
|
| 5993 |
+
{
|
| 5994 |
+
"epoch": 0.6669491525423729,
|
| 5995 |
+
"grad_norm": 1.1581579446792603,
|
| 5996 |
+
"learning_rate": 6.030150089076199e-06,
|
| 5997 |
+
"loss": 0.0046,
|
| 5998 |
+
"step": 787
|
| 5999 |
+
},
|
| 6000 |
+
{
|
| 6001 |
+
"epoch": 0.6677966101694915,
|
| 6002 |
+
"grad_norm": 2.128571033477783,
|
| 6003 |
+
"learning_rate": 6.0030165323980794e-06,
|
| 6004 |
+
"loss": 0.0052,
|
| 6005 |
+
"step": 788
|
| 6006 |
+
},
|
| 6007 |
+
{
|
| 6008 |
+
"epoch": 0.6686440677966101,
|
| 6009 |
+
"grad_norm": 3.439826011657715,
|
| 6010 |
+
"learning_rate": 5.975917952735272e-06,
|
| 6011 |
+
"loss": 0.0174,
|
| 6012 |
+
"step": 789
|
| 6013 |
+
},
|
| 6014 |
+
{
|
| 6015 |
+
"epoch": 0.6694915254237288,
|
| 6016 |
+
"grad_norm": 0.6685582995414734,
|
| 6017 |
+
"learning_rate": 5.948854587223465e-06,
|
| 6018 |
+
"loss": 0.0034,
|
| 6019 |
+
"step": 790
|
| 6020 |
+
},
|
| 6021 |
+
{
|
| 6022 |
+
"epoch": 0.6703389830508475,
|
| 6023 |
+
"grad_norm": 0.4205131530761719,
|
| 6024 |
+
"learning_rate": 5.921826672690199e-06,
|
| 6025 |
+
"loss": 0.0016,
|
| 6026 |
+
"step": 791
|
| 6027 |
+
},
|
| 6028 |
+
{
|
| 6029 |
+
"epoch": 0.6711864406779661,
|
| 6030 |
+
"grad_norm": 0.8802849650382996,
|
| 6031 |
+
"learning_rate": 5.894834445652777e-06,
|
| 6032 |
+
"loss": 0.0041,
|
| 6033 |
+
"step": 792
|
| 6034 |
+
},
|
| 6035 |
+
{
|
| 6036 |
+
"epoch": 0.6720338983050848,
|
| 6037 |
+
"grad_norm": 2.2352685928344727,
|
| 6038 |
+
"learning_rate": 5.867878142316221e-06,
|
| 6039 |
+
"loss": 0.0108,
|
| 6040 |
+
"step": 793
|
| 6041 |
+
},
|
| 6042 |
+
{
|
| 6043 |
+
"epoch": 0.6728813559322034,
|
| 6044 |
+
"grad_norm": 2.6779541969299316,
|
| 6045 |
+
"learning_rate": 5.840957998571177e-06,
|
| 6046 |
+
"loss": 0.0305,
|
| 6047 |
+
"step": 794
|
| 6048 |
+
},
|
| 6049 |
+
{
|
| 6050 |
+
"epoch": 0.673728813559322,
|
| 6051 |
+
"grad_norm": 2.999138116836548,
|
| 6052 |
+
"learning_rate": 5.814074249991876e-06,
|
| 6053 |
+
"loss": 0.0156,
|
| 6054 |
+
"step": 795
|
| 6055 |
+
},
|
| 6056 |
+
{
|
| 6057 |
+
"epoch": 0.6745762711864407,
|
| 6058 |
+
"grad_norm": 0.4623313546180725,
|
| 6059 |
+
"learning_rate": 5.787227131834043e-06,
|
| 6060 |
+
"loss": 0.0013,
|
| 6061 |
+
"step": 796
|
| 6062 |
+
},
|
| 6063 |
+
{
|
| 6064 |
+
"epoch": 0.6754237288135593,
|
| 6065 |
+
"grad_norm": 0.4670167863368988,
|
| 6066 |
+
"learning_rate": 5.7604168790328774e-06,
|
| 6067 |
+
"loss": 0.0016,
|
| 6068 |
+
"step": 797
|
| 6069 |
+
},
|
| 6070 |
+
{
|
| 6071 |
+
"epoch": 0.676271186440678,
|
| 6072 |
+
"grad_norm": 1.1071865558624268,
|
| 6073 |
+
"learning_rate": 5.733643726200958e-06,
|
| 6074 |
+
"loss": 0.0055,
|
| 6075 |
+
"step": 798
|
| 6076 |
+
},
|
| 6077 |
+
{
|
| 6078 |
+
"epoch": 0.6771186440677966,
|
| 6079 |
+
"grad_norm": 1.0425511598587036,
|
| 6080 |
+
"learning_rate": 5.706907907626216e-06,
|
| 6081 |
+
"loss": 0.0115,
|
| 6082 |
+
"step": 799
|
| 6083 |
+
},
|
| 6084 |
+
{
|
| 6085 |
+
"epoch": 0.6779661016949152,
|
| 6086 |
+
"grad_norm": 0.9477194547653198,
|
| 6087 |
+
"learning_rate": 5.680209657269871e-06,
|
| 6088 |
+
"loss": 0.0039,
|
| 6089 |
+
"step": 800
|
| 6090 |
+
},
|
| 6091 |
+
{
|
| 6092 |
+
"epoch": 0.6779661016949152,
|
| 6093 |
+
"eval_accuracy": 1.0,
|
| 6094 |
+
"eval_f1": 1.0,
|
| 6095 |
+
"eval_loss": 0.0001500351936556399,
|
| 6096 |
+
"eval_precision": 1.0,
|
| 6097 |
+
"eval_recall": 1.0,
|
| 6098 |
+
"eval_runtime": 52.726,
|
| 6099 |
+
"eval_samples_per_second": 5.538,
|
| 6100 |
+
"eval_steps_per_second": 0.19,
|
| 6101 |
+
"step": 800
|
| 6102 |
}
|
| 6103 |
],
|
| 6104 |
"logging_steps": 1,
|
|
|
|
| 6118 |
"attributes": {}
|
| 6119 |
}
|
| 6120 |
},
|
| 6121 |
+
"total_flos": 2.46322680561664e+17,
|
| 6122 |
"train_batch_size": 8,
|
| 6123 |
"trial_name": null,
|
| 6124 |
"trial_params": null
|