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/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 sha256:68eb5ed2e06aaf8ccf51155c64c6d00bc22fb8856e437e0118b64ceffe8a67cb
|
| 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:ed69f0ff9655bfaab7272f6ad7c0ae7dd2a7e69609ff879b264ac67fb89fa9f1
|
| 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 sha256:9bbd790913cc32d61c035f94711e5213bc995991b0da5ce1e10cae6c207d6aff
|
| 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 sha256:47ee402b977d2bcdbecf23c2cac2e41b1eece5636309269bebeeddbf9c83f8d2
|
| 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:a3d4686b220e210f4828731ecf8026fe744b87154e9a5102165f52041740b6ac
|
| 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:ff71901f9fc3dd0a0169fd09f191b1474c7af397ce776f7c375bb3b28eb7bd0a
|
| 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 sha256:7bd67655dcdb83e7ff22694d01a9823713b17e32ab4a1e5c80da2d423d2c37fc
|
| 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:bebb458ab33cd9a17efd02365cbba8cf0457b1be88b705a8de29ec92988fadf4
|
| 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 sha256:930942e93c4300034c378c60f4a0dab9ede67a9d39a6de67496fcfc48b953599
|
| 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 sha256:044d30e23b739bee6ba5f240459efac0778a387063ea9930d3e3133b74846187
|
| 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": 5.836,
|
| 5340 |
"eval_steps_per_second": 0.19,
|
| 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": 0.7421150278293135,
|
| 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": 5.836,
|
| 5340 |
"eval_steps_per_second": 0.19,
|
| 5341 |
"step": 700
|
| 5342 |
+
},
|
| 5343 |
+
{
|
| 5344 |
+
"epoch": 0.650278293135436,
|
| 5345 |
+
"grad_norm": 5.3951334953308105,
|
| 5346 |
+
"learning_rate": 6.5730362654554015e-06,
|
| 5347 |
+
"loss": 0.2149,
|
| 5348 |
+
"step": 701
|
| 5349 |
+
},
|
| 5350 |
+
{
|
| 5351 |
+
"epoch": 0.6512059369202227,
|
| 5352 |
+
"grad_norm": 4.401817798614502,
|
| 5353 |
+
"learning_rate": 6.542627927979772e-06,
|
| 5354 |
+
"loss": 0.2215,
|
| 5355 |
+
"step": 702
|
| 5356 |
+
},
|
| 5357 |
+
{
|
| 5358 |
+
"epoch": 0.6521335807050093,
|
| 5359 |
+
"grad_norm": 3.9779934883117676,
|
| 5360 |
+
"learning_rate": 6.5122558567011775e-06,
|
| 5361 |
+
"loss": 0.1763,
|
| 5362 |
+
"step": 703
|
| 5363 |
+
},
|
| 5364 |
+
{
|
| 5365 |
+
"epoch": 0.6530612244897959,
|
| 5366 |
+
"grad_norm": 3.7413315773010254,
|
| 5367 |
+
"learning_rate": 6.481920370208274e-06,
|
| 5368 |
+
"loss": 0.1176,
|
| 5369 |
+
"step": 704
|
| 5370 |
+
},
|
| 5371 |
+
{
|
| 5372 |
+
"epoch": 0.6539888682745826,
|
| 5373 |
+
"grad_norm": 6.082540988922119,
|
| 5374 |
+
"learning_rate": 6.4516217867059615e-06,
|
| 5375 |
+
"loss": 0.2444,
|
| 5376 |
+
"step": 705
|
| 5377 |
+
},
|
| 5378 |
+
{
|
| 5379 |
+
"epoch": 0.6549165120593692,
|
| 5380 |
+
"grad_norm": 7.003627300262451,
|
| 5381 |
+
"learning_rate": 6.421360424012039e-06,
|
| 5382 |
+
"loss": 0.1389,
|
| 5383 |
+
"step": 706
|
| 5384 |
+
},
|
| 5385 |
+
{
|
| 5386 |
+
"epoch": 0.6558441558441559,
|
| 5387 |
+
"grad_norm": 4.237902641296387,
|
| 5388 |
+
"learning_rate": 6.39113659955389e-06,
|
| 5389 |
+
"loss": 0.1767,
|
| 5390 |
+
"step": 707
|
| 5391 |
+
},
|
| 5392 |
+
{
|
| 5393 |
+
"epoch": 0.6567717996289425,
|
| 5394 |
+
"grad_norm": 3.1200437545776367,
|
| 5395 |
+
"learning_rate": 6.360950630365126e-06,
|
| 5396 |
+
"loss": 0.1127,
|
| 5397 |
+
"step": 708
|
| 5398 |
+
},
|
| 5399 |
+
{
|
| 5400 |
+
"epoch": 0.6576994434137291,
|
| 5401 |
+
"grad_norm": 4.659437656402588,
|
| 5402 |
+
"learning_rate": 6.33080283308228e-06,
|
| 5403 |
+
"loss": 0.1292,
|
| 5404 |
+
"step": 709
|
| 5405 |
+
},
|
| 5406 |
+
{
|
| 5407 |
+
"epoch": 0.6586270871985158,
|
| 5408 |
+
"grad_norm": 7.017888069152832,
|
| 5409 |
+
"learning_rate": 6.300693523941481e-06,
|
| 5410 |
+
"loss": 0.201,
|
| 5411 |
+
"step": 710
|
| 5412 |
+
},
|
| 5413 |
+
{
|
| 5414 |
+
"epoch": 0.6595547309833024,
|
| 5415 |
+
"grad_norm": 4.794656276702881,
|
| 5416 |
+
"learning_rate": 6.270623018775135e-06,
|
| 5417 |
+
"loss": 0.1835,
|
| 5418 |
+
"step": 711
|
| 5419 |
+
},
|
| 5420 |
+
{
|
| 5421 |
+
"epoch": 0.660482374768089,
|
| 5422 |
+
"grad_norm": 4.382326602935791,
|
| 5423 |
+
"learning_rate": 6.2405916330086106e-06,
|
| 5424 |
+
"loss": 0.1703,
|
| 5425 |
+
"step": 712
|
| 5426 |
+
},
|
| 5427 |
+
{
|
| 5428 |
+
"epoch": 0.6614100185528757,
|
| 5429 |
+
"grad_norm": 6.628244876861572,
|
| 5430 |
+
"learning_rate": 6.210599681656933e-06,
|
| 5431 |
+
"loss": 0.2367,
|
| 5432 |
+
"step": 713
|
| 5433 |
+
},
|
| 5434 |
+
{
|
| 5435 |
+
"epoch": 0.6623376623376623,
|
| 5436 |
+
"grad_norm": 4.854303359985352,
|
| 5437 |
+
"learning_rate": 6.180647479321484e-06,
|
| 5438 |
+
"loss": 0.1874,
|
| 5439 |
+
"step": 714
|
| 5440 |
+
},
|
| 5441 |
+
{
|
| 5442 |
+
"epoch": 0.6632653061224489,
|
| 5443 |
+
"grad_norm": 2.9809422492980957,
|
| 5444 |
+
"learning_rate": 6.1507353401866896e-06,
|
| 5445 |
+
"loss": 0.1429,
|
| 5446 |
+
"step": 715
|
| 5447 |
+
},
|
| 5448 |
+
{
|
| 5449 |
+
"epoch": 0.6641929499072357,
|
| 5450 |
+
"grad_norm": 3.899773359298706,
|
| 5451 |
+
"learning_rate": 6.120863578016736e-06,
|
| 5452 |
+
"loss": 0.1745,
|
| 5453 |
+
"step": 716
|
| 5454 |
+
},
|
| 5455 |
+
{
|
| 5456 |
+
"epoch": 0.6651205936920223,
|
| 5457 |
+
"grad_norm": 5.175771713256836,
|
| 5458 |
+
"learning_rate": 6.091032506152274e-06,
|
| 5459 |
+
"loss": 0.162,
|
| 5460 |
+
"step": 717
|
| 5461 |
+
},
|
| 5462 |
+
{
|
| 5463 |
+
"epoch": 0.6660482374768089,
|
| 5464 |
+
"grad_norm": 4.2920002937316895,
|
| 5465 |
+
"learning_rate": 6.061242437507131e-06,
|
| 5466 |
+
"loss": 0.1806,
|
| 5467 |
+
"step": 718
|
| 5468 |
+
},
|
| 5469 |
+
{
|
| 5470 |
+
"epoch": 0.6669758812615956,
|
| 5471 |
+
"grad_norm": 4.961828708648682,
|
| 5472 |
+
"learning_rate": 6.0314936845650296e-06,
|
| 5473 |
+
"loss": 0.1259,
|
| 5474 |
+
"step": 719
|
| 5475 |
+
},
|
| 5476 |
+
{
|
| 5477 |
+
"epoch": 0.6679035250463822,
|
| 5478 |
+
"grad_norm": 4.258510589599609,
|
| 5479 |
+
"learning_rate": 6.00178655937631e-06,
|
| 5480 |
+
"loss": 0.2015,
|
| 5481 |
+
"step": 720
|
| 5482 |
+
},
|
| 5483 |
+
{
|
| 5484 |
+
"epoch": 0.6679035250463822,
|
| 5485 |
+
"eval_accuracy": 0.852549889135255,
|
| 5486 |
+
"eval_f1": 0.6825775656324582,
|
| 5487 |
+
"eval_loss": 0.31203413009643555,
|
| 5488 |
+
"eval_precision": 0.8614457831325302,
|
| 5489 |
+
"eval_recall": 0.5652173913043478,
|
| 5490 |
+
"eval_runtime": 47.6328,
|
| 5491 |
+
"eval_samples_per_second": 5.794,
|
| 5492 |
+
"eval_steps_per_second": 0.189,
|
| 5493 |
+
"step": 720
|
| 5494 |
+
},
|
| 5495 |
+
{
|
| 5496 |
+
"epoch": 0.6688311688311688,
|
| 5497 |
+
"grad_norm": 4.69655179977417,
|
| 5498 |
+
"learning_rate": 5.972121373554665e-06,
|
| 5499 |
+
"loss": 0.1334,
|
| 5500 |
+
"step": 721
|
| 5501 |
+
},
|
| 5502 |
+
{
|
| 5503 |
+
"epoch": 0.6697588126159555,
|
| 5504 |
+
"grad_norm": 4.1556806564331055,
|
| 5505 |
+
"learning_rate": 5.942498438273849e-06,
|
| 5506 |
+
"loss": 0.159,
|
| 5507 |
+
"step": 722
|
| 5508 |
+
},
|
| 5509 |
+
{
|
| 5510 |
+
"epoch": 0.6706864564007421,
|
| 5511 |
+
"grad_norm": 4.261734485626221,
|
| 5512 |
+
"learning_rate": 5.912918064264441e-06,
|
| 5513 |
+
"loss": 0.14,
|
| 5514 |
+
"step": 723
|
| 5515 |
+
},
|
| 5516 |
+
{
|
| 5517 |
+
"epoch": 0.6716141001855288,
|
| 5518 |
+
"grad_norm": 5.172565460205078,
|
| 5519 |
+
"learning_rate": 5.8833805618105635e-06,
|
| 5520 |
+
"loss": 0.2259,
|
| 5521 |
+
"step": 724
|
| 5522 |
+
},
|
| 5523 |
+
{
|
| 5524 |
+
"epoch": 0.6725417439703154,
|
| 5525 |
+
"grad_norm": 5.771614074707031,
|
| 5526 |
+
"learning_rate": 5.853886240746643e-06,
|
| 5527 |
+
"loss": 0.1806,
|
| 5528 |
+
"step": 725
|
| 5529 |
+
},
|
| 5530 |
+
{
|
| 5531 |
+
"epoch": 0.673469387755102,
|
| 5532 |
+
"grad_norm": 3.933560371398926,
|
| 5533 |
+
"learning_rate": 5.82443541045415e-06,
|
| 5534 |
+
"loss": 0.1874,
|
| 5535 |
+
"step": 726
|
| 5536 |
+
},
|
| 5537 |
+
{
|
| 5538 |
+
"epoch": 0.6743970315398887,
|
| 5539 |
+
"grad_norm": 4.140106201171875,
|
| 5540 |
+
"learning_rate": 5.795028379858355e-06,
|
| 5541 |
+
"loss": 0.1252,
|
| 5542 |
+
"step": 727
|
| 5543 |
+
},
|
| 5544 |
+
{
|
| 5545 |
+
"epoch": 0.6753246753246753,
|
| 5546 |
+
"grad_norm": 4.954802513122559,
|
| 5547 |
+
"learning_rate": 5.765665457425102e-06,
|
| 5548 |
+
"loss": 0.2296,
|
| 5549 |
+
"step": 728
|
| 5550 |
+
},
|
| 5551 |
+
{
|
| 5552 |
+
"epoch": 0.6762523191094619,
|
| 5553 |
+
"grad_norm": 3.336763620376587,
|
| 5554 |
+
"learning_rate": 5.736346951157544e-06,
|
| 5555 |
+
"loss": 0.1571,
|
| 5556 |
+
"step": 729
|
| 5557 |
+
},
|
| 5558 |
+
{
|
| 5559 |
+
"epoch": 0.6771799628942486,
|
| 5560 |
+
"grad_norm": 6.951479434967041,
|
| 5561 |
+
"learning_rate": 5.707073168592943e-06,
|
| 5562 |
+
"loss": 0.227,
|
| 5563 |
+
"step": 730
|
| 5564 |
+
},
|
| 5565 |
+
{
|
| 5566 |
+
"epoch": 0.6781076066790352,
|
| 5567 |
+
"grad_norm": 6.168885231018066,
|
| 5568 |
+
"learning_rate": 5.677844416799424e-06,
|
| 5569 |
+
"loss": 0.1793,
|
| 5570 |
+
"step": 731
|
| 5571 |
+
},
|
| 5572 |
+
{
|
| 5573 |
+
"epoch": 0.6790352504638218,
|
| 5574 |
+
"grad_norm": 4.202786445617676,
|
| 5575 |
+
"learning_rate": 5.648661002372769e-06,
|
| 5576 |
+
"loss": 0.1307,
|
| 5577 |
+
"step": 732
|
| 5578 |
+
},
|
| 5579 |
+
{
|
| 5580 |
+
"epoch": 0.6799628942486086,
|
| 5581 |
+
"grad_norm": 5.333949565887451,
|
| 5582 |
+
"learning_rate": 5.619523231433177e-06,
|
| 5583 |
+
"loss": 0.1435,
|
| 5584 |
+
"step": 733
|
| 5585 |
+
},
|
| 5586 |
+
{
|
| 5587 |
+
"epoch": 0.6808905380333952,
|
| 5588 |
+
"grad_norm": 3.218813896179199,
|
| 5589 |
+
"learning_rate": 5.590431409622081e-06,
|
| 5590 |
+
"loss": 0.152,
|
| 5591 |
+
"step": 734
|
| 5592 |
+
},
|
| 5593 |
+
{
|
| 5594 |
+
"epoch": 0.6818181818181818,
|
| 5595 |
+
"grad_norm": 3.0740416049957275,
|
| 5596 |
+
"learning_rate": 5.56138584209893e-06,
|
| 5597 |
+
"loss": 0.1625,
|
| 5598 |
+
"step": 735
|
| 5599 |
+
},
|
| 5600 |
+
{
|
| 5601 |
+
"epoch": 0.6827458256029685,
|
| 5602 |
+
"grad_norm": 3.7008039951324463,
|
| 5603 |
+
"learning_rate": 5.5323868335379775e-06,
|
| 5604 |
+
"loss": 0.1361,
|
| 5605 |
+
"step": 736
|
| 5606 |
+
},
|
| 5607 |
+
{
|
| 5608 |
+
"epoch": 0.6836734693877551,
|
| 5609 |
+
"grad_norm": 3.482449769973755,
|
| 5610 |
+
"learning_rate": 5.503434688125104e-06,
|
| 5611 |
+
"loss": 0.1539,
|
| 5612 |
+
"step": 737
|
| 5613 |
+
},
|
| 5614 |
+
{
|
| 5615 |
+
"epoch": 0.6846011131725418,
|
| 5616 |
+
"grad_norm": 3.1356945037841797,
|
| 5617 |
+
"learning_rate": 5.4745297095546125e-06,
|
| 5618 |
+
"loss": 0.116,
|
| 5619 |
+
"step": 738
|
| 5620 |
+
},
|
| 5621 |
+
{
|
| 5622 |
+
"epoch": 0.6855287569573284,
|
| 5623 |
+
"grad_norm": 3.0260050296783447,
|
| 5624 |
+
"learning_rate": 5.445672201026054e-06,
|
| 5625 |
+
"loss": 0.1491,
|
| 5626 |
+
"step": 739
|
| 5627 |
+
},
|
| 5628 |
+
{
|
| 5629 |
+
"epoch": 0.686456400742115,
|
| 5630 |
+
"grad_norm": 6.386862754821777,
|
| 5631 |
+
"learning_rate": 5.416862465241033e-06,
|
| 5632 |
+
"loss": 0.2216,
|
| 5633 |
+
"step": 740
|
| 5634 |
+
},
|
| 5635 |
+
{
|
| 5636 |
+
"epoch": 0.686456400742115,
|
| 5637 |
+
"eval_accuracy": 0.8503325942350333,
|
| 5638 |
+
"eval_f1": 0.6762589928057554,
|
| 5639 |
+
"eval_loss": 0.30298492312431335,
|
| 5640 |
+
"eval_precision": 0.8597560975609756,
|
| 5641 |
+
"eval_recall": 0.5573122529644269,
|
| 5642 |
+
"eval_runtime": 47.363,
|
| 5643 |
+
"eval_samples_per_second": 5.827,
|
| 5644 |
+
"eval_steps_per_second": 0.19,
|
| 5645 |
+
"step": 740
|
| 5646 |
+
},
|
| 5647 |
+
{
|
| 5648 |
+
"epoch": 0.6873840445269017,
|
| 5649 |
+
"grad_norm": 2.5552661418914795,
|
| 5650 |
+
"learning_rate": 5.3881008044000495e-06,
|
| 5651 |
+
"loss": 0.113,
|
| 5652 |
+
"step": 741
|
| 5653 |
+
},
|
| 5654 |
+
{
|
| 5655 |
+
"epoch": 0.6883116883116883,
|
| 5656 |
+
"grad_norm": 3.5924112796783447,
|
| 5657 |
+
"learning_rate": 5.359387520199317e-06,
|
| 5658 |
+
"loss": 0.0896,
|
| 5659 |
+
"step": 742
|
| 5660 |
+
},
|
| 5661 |
+
{
|
| 5662 |
+
"epoch": 0.6892393320964749,
|
| 5663 |
+
"grad_norm": 3.5902557373046875,
|
| 5664 |
+
"learning_rate": 5.330722913827594e-06,
|
| 5665 |
+
"loss": 0.2078,
|
| 5666 |
+
"step": 743
|
| 5667 |
+
},
|
| 5668 |
+
{
|
| 5669 |
+
"epoch": 0.6901669758812616,
|
| 5670 |
+
"grad_norm": 3.3769776821136475,
|
| 5671 |
+
"learning_rate": 5.302107285963045e-06,
|
| 5672 |
+
"loss": 0.1311,
|
| 5673 |
+
"step": 744
|
| 5674 |
+
},
|
| 5675 |
+
{
|
| 5676 |
+
"epoch": 0.6910946196660482,
|
| 5677 |
+
"grad_norm": 4.20431661605835,
|
| 5678 |
+
"learning_rate": 5.273540936770059e-06,
|
| 5679 |
+
"loss": 0.1694,
|
| 5680 |
+
"step": 745
|
| 5681 |
+
},
|
| 5682 |
+
{
|
| 5683 |
+
"epoch": 0.6920222634508348,
|
| 5684 |
+
"grad_norm": 3.680614948272705,
|
| 5685 |
+
"learning_rate": 5.245024165896126e-06,
|
| 5686 |
+
"loss": 0.168,
|
| 5687 |
+
"step": 746
|
| 5688 |
+
},
|
| 5689 |
+
{
|
| 5690 |
+
"epoch": 0.6929499072356216,
|
| 5691 |
+
"grad_norm": 4.88540506362915,
|
| 5692 |
+
"learning_rate": 5.216557272468675e-06,
|
| 5693 |
+
"loss": 0.2021,
|
| 5694 |
+
"step": 747
|
| 5695 |
+
},
|
| 5696 |
+
{
|
| 5697 |
+
"epoch": 0.6938775510204082,
|
| 5698 |
+
"grad_norm": 3.159914016723633,
|
| 5699 |
+
"learning_rate": 5.18814055509195e-06,
|
| 5700 |
+
"loss": 0.1696,
|
| 5701 |
+
"step": 748
|
| 5702 |
+
},
|
| 5703 |
+
{
|
| 5704 |
+
"epoch": 0.6948051948051948,
|
| 5705 |
+
"grad_norm": 5.163516044616699,
|
| 5706 |
+
"learning_rate": 5.1597743118438725e-06,
|
| 5707 |
+
"loss": 0.1485,
|
| 5708 |
+
"step": 749
|
| 5709 |
+
},
|
| 5710 |
+
{
|
| 5711 |
+
"epoch": 0.6957328385899815,
|
| 5712 |
+
"grad_norm": 6.143499851226807,
|
| 5713 |
+
"learning_rate": 5.131458840272905e-06,
|
| 5714 |
+
"loss": 0.1769,
|
| 5715 |
+
"step": 750
|
| 5716 |
+
},
|
| 5717 |
+
{
|
| 5718 |
+
"epoch": 0.6966604823747681,
|
| 5719 |
+
"grad_norm": 3.032066822052002,
|
| 5720 |
+
"learning_rate": 5.103194437394952e-06,
|
| 5721 |
+
"loss": 0.1202,
|
| 5722 |
+
"step": 751
|
| 5723 |
+
},
|
| 5724 |
+
{
|
| 5725 |
+
"epoch": 0.6975881261595547,
|
| 5726 |
+
"grad_norm": 4.9984965324401855,
|
| 5727 |
+
"learning_rate": 5.074981399690219e-06,
|
| 5728 |
+
"loss": 0.2292,
|
| 5729 |
+
"step": 752
|
| 5730 |
+
},
|
| 5731 |
+
{
|
| 5732 |
+
"epoch": 0.6985157699443414,
|
| 5733 |
+
"grad_norm": 2.7881083488464355,
|
| 5734 |
+
"learning_rate": 5.046820023100129e-06,
|
| 5735 |
+
"loss": 0.0989,
|
| 5736 |
+
"step": 753
|
| 5737 |
+
},
|
| 5738 |
+
{
|
| 5739 |
+
"epoch": 0.699443413729128,
|
| 5740 |
+
"grad_norm": 4.462379455566406,
|
| 5741 |
+
"learning_rate": 5.018710603024187e-06,
|
| 5742 |
+
"loss": 0.1788,
|
| 5743 |
+
"step": 754
|
| 5744 |
+
},
|
| 5745 |
+
{
|
| 5746 |
+
"epoch": 0.7003710575139147,
|
| 5747 |
+
"grad_norm": 5.801154613494873,
|
| 5748 |
+
"learning_rate": 4.990653434316915e-06,
|
| 5749 |
+
"loss": 0.1807,
|
| 5750 |
+
"step": 755
|
| 5751 |
+
},
|
| 5752 |
+
{
|
| 5753 |
+
"epoch": 0.7012987012987013,
|
| 5754 |
+
"grad_norm": 4.452095031738281,
|
| 5755 |
+
"learning_rate": 4.9626488112847384e-06,
|
| 5756 |
+
"loss": 0.1348,
|
| 5757 |
+
"step": 756
|
| 5758 |
+
},
|
| 5759 |
+
{
|
| 5760 |
+
"epoch": 0.7022263450834879,
|
| 5761 |
+
"grad_norm": 4.256060600280762,
|
| 5762 |
+
"learning_rate": 4.934697027682894e-06,
|
| 5763 |
+
"loss": 0.1618,
|
| 5764 |
+
"step": 757
|
| 5765 |
+
},
|
| 5766 |
+
{
|
| 5767 |
+
"epoch": 0.7031539888682746,
|
| 5768 |
+
"grad_norm": 2.893348217010498,
|
| 5769 |
+
"learning_rate": 4.9067983767123736e-06,
|
| 5770 |
+
"loss": 0.114,
|
| 5771 |
+
"step": 758
|
| 5772 |
+
},
|
| 5773 |
+
{
|
| 5774 |
+
"epoch": 0.7040816326530612,
|
| 5775 |
+
"grad_norm": 2.409883737564087,
|
| 5776 |
+
"learning_rate": 4.878953151016816e-06,
|
| 5777 |
+
"loss": 0.0739,
|
| 5778 |
+
"step": 759
|
| 5779 |
+
},
|
| 5780 |
+
{
|
| 5781 |
+
"epoch": 0.7050092764378478,
|
| 5782 |
+
"grad_norm": 4.646303176879883,
|
| 5783 |
+
"learning_rate": 4.851161642679466e-06,
|
| 5784 |
+
"loss": 0.1936,
|
| 5785 |
+
"step": 760
|
| 5786 |
+
},
|
| 5787 |
+
{
|
| 5788 |
+
"epoch": 0.7050092764378478,
|
| 5789 |
+
"eval_accuracy": 0.8503325942350333,
|
| 5790 |
+
"eval_f1": 0.6762589928057554,
|
| 5791 |
+
"eval_loss": 0.30913296341896057,
|
| 5792 |
+
"eval_precision": 0.8597560975609756,
|
| 5793 |
+
"eval_recall": 0.5573122529644269,
|
| 5794 |
+
"eval_runtime": 47.2588,
|
| 5795 |
+
"eval_samples_per_second": 5.84,
|
| 5796 |
+
"eval_steps_per_second": 0.19,
|
| 5797 |
+
"step": 760
|
| 5798 |
+
},
|
| 5799 |
+
{
|
| 5800 |
+
"epoch": 0.7059369202226345,
|
| 5801 |
+
"grad_norm": 3.7548396587371826,
|
| 5802 |
+
"learning_rate": 4.823424143220097e-06,
|
| 5803 |
+
"loss": 0.1332,
|
| 5804 |
+
"step": 761
|
| 5805 |
+
},
|
| 5806 |
+
{
|
| 5807 |
+
"epoch": 0.7068645640074211,
|
| 5808 |
+
"grad_norm": 4.519832611083984,
|
| 5809 |
+
"learning_rate": 4.795740943591955e-06,
|
| 5810 |
+
"loss": 0.1508,
|
| 5811 |
+
"step": 762
|
| 5812 |
+
},
|
| 5813 |
+
{
|
| 5814 |
+
"epoch": 0.7077922077922078,
|
| 5815 |
+
"grad_norm": 4.223718643188477,
|
| 5816 |
+
"learning_rate": 4.7681123341787e-06,
|
| 5817 |
+
"loss": 0.2027,
|
| 5818 |
+
"step": 763
|
| 5819 |
+
},
|
| 5820 |
+
{
|
| 5821 |
+
"epoch": 0.7087198515769945,
|
| 5822 |
+
"grad_norm": 4.345351219177246,
|
| 5823 |
+
"learning_rate": 4.740538604791371e-06,
|
| 5824 |
+
"loss": 0.1641,
|
| 5825 |
+
"step": 764
|
| 5826 |
+
},
|
| 5827 |
+
{
|
| 5828 |
+
"epoch": 0.7096474953617811,
|
| 5829 |
+
"grad_norm": 6.437448024749756,
|
| 5830 |
+
"learning_rate": 4.713020044665348e-06,
|
| 5831 |
+
"loss": 0.1839,
|
| 5832 |
+
"step": 765
|
| 5833 |
+
},
|
| 5834 |
+
{
|
| 5835 |
+
"epoch": 0.7105751391465677,
|
| 5836 |
+
"grad_norm": 5.154228687286377,
|
| 5837 |
+
"learning_rate": 4.685556942457296e-06,
|
| 5838 |
+
"loss": 0.1696,
|
| 5839 |
+
"step": 766
|
| 5840 |
+
},
|
| 5841 |
+
{
|
| 5842 |
+
"epoch": 0.7115027829313544,
|
| 5843 |
+
"grad_norm": 4.437908172607422,
|
| 5844 |
+
"learning_rate": 4.65814958624217e-06,
|
| 5845 |
+
"loss": 0.1199,
|
| 5846 |
+
"step": 767
|
| 5847 |
+
},
|
| 5848 |
+
{
|
| 5849 |
+
"epoch": 0.712430426716141,
|
| 5850 |
+
"grad_norm": 4.337602615356445,
|
| 5851 |
+
"learning_rate": 4.630798263510162e-06,
|
| 5852 |
+
"loss": 0.166,
|
| 5853 |
+
"step": 768
|
| 5854 |
+
},
|
| 5855 |
+
{
|
| 5856 |
+
"epoch": 0.7133580705009277,
|
| 5857 |
+
"grad_norm": 4.633522987365723,
|
| 5858 |
+
"learning_rate": 4.60350326116371e-06,
|
| 5859 |
+
"loss": 0.1667,
|
| 5860 |
+
"step": 769
|
| 5861 |
+
},
|
| 5862 |
+
{
|
| 5863 |
+
"epoch": 0.7142857142857143,
|
| 5864 |
+
"grad_norm": 4.564393997192383,
|
| 5865 |
+
"learning_rate": 4.576264865514467e-06,
|
| 5866 |
+
"loss": 0.1996,
|
| 5867 |
+
"step": 770
|
| 5868 |
+
},
|
| 5869 |
+
{
|
| 5870 |
+
"epoch": 0.7152133580705009,
|
| 5871 |
+
"grad_norm": 5.6954803466796875,
|
| 5872 |
+
"learning_rate": 4.549083362280318e-06,
|
| 5873 |
+
"loss": 0.1227,
|
| 5874 |
+
"step": 771
|
| 5875 |
+
},
|
| 5876 |
+
{
|
| 5877 |
+
"epoch": 0.7161410018552876,
|
| 5878 |
+
"grad_norm": 2.704313278198242,
|
| 5879 |
+
"learning_rate": 4.521959036582372e-06,
|
| 5880 |
+
"loss": 0.1118,
|
| 5881 |
+
"step": 772
|
| 5882 |
+
},
|
| 5883 |
+
{
|
| 5884 |
+
"epoch": 0.7170686456400742,
|
| 5885 |
+
"grad_norm": 3.529452085494995,
|
| 5886 |
+
"learning_rate": 4.494892172941965e-06,
|
| 5887 |
+
"loss": 0.1595,
|
| 5888 |
+
"step": 773
|
| 5889 |
+
},
|
| 5890 |
+
{
|
| 5891 |
+
"epoch": 0.7179962894248608,
|
| 5892 |
+
"grad_norm": 4.921022415161133,
|
| 5893 |
+
"learning_rate": 4.467883055277696e-06,
|
| 5894 |
+
"loss": 0.1755,
|
| 5895 |
+
"step": 774
|
| 5896 |
+
},
|
| 5897 |
+
{
|
| 5898 |
+
"epoch": 0.7189239332096475,
|
| 5899 |
+
"grad_norm": 3.468756675720215,
|
| 5900 |
+
"learning_rate": 4.440931966902419e-06,
|
| 5901 |
+
"loss": 0.1198,
|
| 5902 |
+
"step": 775
|
| 5903 |
+
},
|
| 5904 |
+
{
|
| 5905 |
+
"epoch": 0.7198515769944341,
|
| 5906 |
+
"grad_norm": 3.8362793922424316,
|
| 5907 |
+
"learning_rate": 4.414039190520308e-06,
|
| 5908 |
+
"loss": 0.162,
|
| 5909 |
+
"step": 776
|
| 5910 |
+
},
|
| 5911 |
+
{
|
| 5912 |
+
"epoch": 0.7207792207792207,
|
| 5913 |
+
"grad_norm": 3.9787826538085938,
|
| 5914 |
+
"learning_rate": 4.3872050082238535e-06,
|
| 5915 |
+
"loss": 0.1539,
|
| 5916 |
+
"step": 777
|
| 5917 |
+
},
|
| 5918 |
+
{
|
| 5919 |
+
"epoch": 0.7217068645640075,
|
| 5920 |
+
"grad_norm": 4.675662517547607,
|
| 5921 |
+
"learning_rate": 4.360429701490935e-06,
|
| 5922 |
+
"loss": 0.1961,
|
| 5923 |
+
"step": 778
|
| 5924 |
+
},
|
| 5925 |
+
{
|
| 5926 |
+
"epoch": 0.7226345083487941,
|
| 5927 |
+
"grad_norm": 3.907136917114258,
|
| 5928 |
+
"learning_rate": 4.3337135511818514e-06,
|
| 5929 |
+
"loss": 0.1566,
|
| 5930 |
+
"step": 779
|
| 5931 |
+
},
|
| 5932 |
+
{
|
| 5933 |
+
"epoch": 0.7235621521335807,
|
| 5934 |
+
"grad_norm": 4.647920608520508,
|
| 5935 |
+
"learning_rate": 4.307056837536373e-06,
|
| 5936 |
+
"loss": 0.135,
|
| 5937 |
+
"step": 780
|
| 5938 |
+
},
|
| 5939 |
+
{
|
| 5940 |
+
"epoch": 0.7235621521335807,
|
| 5941 |
+
"eval_accuracy": 0.852549889135255,
|
| 5942 |
+
"eval_f1": 0.6855791962174941,
|
| 5943 |
+
"eval_loss": 0.30233433842658997,
|
| 5944 |
+
"eval_precision": 0.8529411764705882,
|
| 5945 |
+
"eval_recall": 0.5731225296442688,
|
| 5946 |
+
"eval_runtime": 49.2875,
|
| 5947 |
+
"eval_samples_per_second": 5.6,
|
| 5948 |
+
"eval_steps_per_second": 0.183,
|
| 5949 |
+
"step": 780
|
| 5950 |
+
},
|
| 5951 |
+
{
|
| 5952 |
+
"epoch": 0.7244897959183674,
|
| 5953 |
+
"grad_norm": 4.066539287567139,
|
| 5954 |
+
"learning_rate": 4.280459840170818e-06,
|
| 5955 |
+
"loss": 0.098,
|
| 5956 |
+
"step": 781
|
| 5957 |
+
},
|
| 5958 |
+
{
|
| 5959 |
+
"epoch": 0.725417439703154,
|
| 5960 |
+
"grad_norm": 4.152294635772705,
|
| 5961 |
+
"learning_rate": 4.2539228380750955e-06,
|
| 5962 |
+
"loss": 0.1492,
|
| 5963 |
+
"step": 782
|
| 5964 |
+
},
|
| 5965 |
+
{
|
| 5966 |
+
"epoch": 0.7263450834879406,
|
| 5967 |
+
"grad_norm": 4.5321831703186035,
|
| 5968 |
+
"learning_rate": 4.2274461096098085e-06,
|
| 5969 |
+
"loss": 0.1727,
|
| 5970 |
+
"step": 783
|
| 5971 |
+
},
|
| 5972 |
+
{
|
| 5973 |
+
"epoch": 0.7272727272727273,
|
| 5974 |
+
"grad_norm": 4.524892807006836,
|
| 5975 |
+
"learning_rate": 4.201029932503303e-06,
|
| 5976 |
+
"loss": 0.1529,
|
| 5977 |
+
"step": 784
|
| 5978 |
+
},
|
| 5979 |
+
{
|
| 5980 |
+
"epoch": 0.7282003710575139,
|
| 5981 |
+
"grad_norm": 6.373286724090576,
|
| 5982 |
+
"learning_rate": 4.17467458384878e-06,
|
| 5983 |
+
"loss": 0.1686,
|
| 5984 |
+
"step": 785
|
| 5985 |
+
},
|
| 5986 |
+
{
|
| 5987 |
+
"epoch": 0.7291280148423006,
|
| 5988 |
+
"grad_norm": 3.8836734294891357,
|
| 5989 |
+
"learning_rate": 4.14838034010138e-06,
|
| 5990 |
+
"loss": 0.1148,
|
| 5991 |
+
"step": 786
|
| 5992 |
+
},
|
| 5993 |
+
{
|
| 5994 |
+
"epoch": 0.7300556586270872,
|
| 5995 |
+
"grad_norm": 4.612453937530518,
|
| 5996 |
+
"learning_rate": 4.12214747707527e-06,
|
| 5997 |
+
"loss": 0.1585,
|
| 5998 |
+
"step": 787
|
| 5999 |
+
},
|
| 6000 |
+
{
|
| 6001 |
+
"epoch": 0.7309833024118738,
|
| 6002 |
+
"grad_norm": 3.490913152694702,
|
| 6003 |
+
"learning_rate": 4.095976269940777e-06,
|
| 6004 |
+
"loss": 0.1045,
|
| 6005 |
+
"step": 788
|
| 6006 |
+
},
|
| 6007 |
+
{
|
| 6008 |
+
"epoch": 0.7319109461966605,
|
| 6009 |
+
"grad_norm": 4.676249027252197,
|
| 6010 |
+
"learning_rate": 4.069866993221473e-06,
|
| 6011 |
+
"loss": 0.2105,
|
| 6012 |
+
"step": 789
|
| 6013 |
+
},
|
| 6014 |
+
{
|
| 6015 |
+
"epoch": 0.7328385899814471,
|
| 6016 |
+
"grad_norm": 4.428783416748047,
|
| 6017 |
+
"learning_rate": 4.043819920791322e-06,
|
| 6018 |
+
"loss": 0.1764,
|
| 6019 |
+
"step": 790
|
| 6020 |
+
},
|
| 6021 |
+
{
|
| 6022 |
+
"epoch": 0.7337662337662337,
|
| 6023 |
+
"grad_norm": 4.41617488861084,
|
| 6024 |
+
"learning_rate": 4.017835325871781e-06,
|
| 6025 |
+
"loss": 0.1616,
|
| 6026 |
+
"step": 791
|
| 6027 |
+
},
|
| 6028 |
+
{
|
| 6029 |
+
"epoch": 0.7346938775510204,
|
| 6030 |
+
"grad_norm": 3.8224010467529297,
|
| 6031 |
+
"learning_rate": 3.991913481028965e-06,
|
| 6032 |
+
"loss": 0.2009,
|
| 6033 |
+
"step": 792
|
| 6034 |
+
},
|
| 6035 |
+
{
|
| 6036 |
+
"epoch": 0.735621521335807,
|
| 6037 |
+
"grad_norm": 4.631185531616211,
|
| 6038 |
+
"learning_rate": 3.966054658170754e-06,
|
| 6039 |
+
"loss": 0.1481,
|
| 6040 |
+
"step": 793
|
| 6041 |
+
},
|
| 6042 |
+
{
|
| 6043 |
+
"epoch": 0.7365491651205937,
|
| 6044 |
+
"grad_norm": 4.32712984085083,
|
| 6045 |
+
"learning_rate": 3.940259128543967e-06,
|
| 6046 |
+
"loss": 0.1977,
|
| 6047 |
+
"step": 794
|
| 6048 |
+
},
|
| 6049 |
+
{
|
| 6050 |
+
"epoch": 0.7374768089053804,
|
| 6051 |
+
"grad_norm": 2.936823606491089,
|
| 6052 |
+
"learning_rate": 3.914527162731498e-06,
|
| 6053 |
+
"loss": 0.1197,
|
| 6054 |
+
"step": 795
|
| 6055 |
+
},
|
| 6056 |
+
{
|
| 6057 |
+
"epoch": 0.738404452690167,
|
| 6058 |
+
"grad_norm": 6.375387191772461,
|
| 6059 |
+
"learning_rate": 3.888859030649498e-06,
|
| 6060 |
+
"loss": 0.2587,
|
| 6061 |
+
"step": 796
|
| 6062 |
+
},
|
| 6063 |
+
{
|
| 6064 |
+
"epoch": 0.7393320964749536,
|
| 6065 |
+
"grad_norm": 5.587421894073486,
|
| 6066 |
+
"learning_rate": 3.863255001544526e-06,
|
| 6067 |
+
"loss": 0.2271,
|
| 6068 |
+
"step": 797
|
| 6069 |
+
},
|
| 6070 |
+
{
|
| 6071 |
+
"epoch": 0.7402597402597403,
|
| 6072 |
+
"grad_norm": 5.898324966430664,
|
| 6073 |
+
"learning_rate": 3.837715343990727e-06,
|
| 6074 |
+
"loss": 0.2026,
|
| 6075 |
+
"step": 798
|
| 6076 |
+
},
|
| 6077 |
+
{
|
| 6078 |
+
"epoch": 0.7411873840445269,
|
| 6079 |
+
"grad_norm": 6.89965295791626,
|
| 6080 |
+
"learning_rate": 3.81224032588703e-06,
|
| 6081 |
+
"loss": 0.222,
|
| 6082 |
+
"step": 799
|
| 6083 |
+
},
|
| 6084 |
+
{
|
| 6085 |
+
"epoch": 0.7421150278293135,
|
| 6086 |
+
"grad_norm": 4.845470428466797,
|
| 6087 |
+
"learning_rate": 3.7868302144543146e-06,
|
| 6088 |
+
"loss": 0.1332,
|
| 6089 |
+
"step": 800
|
| 6090 |
+
},
|
| 6091 |
+
{
|
| 6092 |
+
"epoch": 0.7421150278293135,
|
| 6093 |
+
"eval_accuracy": 0.843680709534368,
|
| 6094 |
+
"eval_f1": 0.6466165413533834,
|
| 6095 |
+
"eval_loss": 0.32069557905197144,
|
| 6096 |
+
"eval_precision": 0.8835616438356164,
|
| 6097 |
+
"eval_recall": 0.5098814229249012,
|
| 6098 |
+
"eval_runtime": 47.5467,
|
| 6099 |
+
"eval_samples_per_second": 5.805,
|
| 6100 |
+
"eval_steps_per_second": 0.189,
|
| 6101 |
+
"step": 800
|
| 6102 |
}
|
| 6103 |
],
|
| 6104 |
"logging_steps": 1,
|
|
|
|
| 6118 |
"attributes": {}
|
| 6119 |
}
|
| 6120 |
},
|
| 6121 |
+
"total_flos": 2.5585969473467187e+17,
|
| 6122 |
"train_batch_size": 8,
|
| 6123 |
"trial_name": null,
|
| 6124 |
"trial_params": null
|