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 sha256:35f51831b2b098cdce1336c36fdb466a2549cbaa1f8a57f3dfb51b4a2a5bf371
|
| 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:a789181143a79739789063cd00a232ab9f16e3bca19ecdc66bcebfc70abdf7f0
|
| 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 sha256:6b06c737fb3780906c6db6f49888f41e1ff147cd36f721e2ec559502e5722dcf
|
| 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 sha256:95524888ea03dc8db342a6452b79ac2dc498646d4c1397845f5c61de5e72a273
|
| 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 sha256:405be8d53a873909641aaba4d30e01e797a1e6db0878263ef451a17ff9e941b9
|
| 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 sha256:4da83fa371fc825b162abc2365a97c78bdd6c68b3c8715678d0f9f6a05017b53
|
| 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:cda0931130711d5820d481d7b5cc9a36c4df6219fb59d6ebb68f7ab10a011c4b
|
| 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:d48786f7f9663e086296dbd832d3f41b07c50d093d8d13185ba7c06b778eba15
|
| 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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad1d2ce7a6ea45f3182b7421bc96713b2844cd0ec18a52bec861802d753d23df
|
| 3 |
size 15024
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bde33c3dc9d4b5847aa5e82a41ef1d715b6cab5c6f68c90d9c12b98c9395b5a1
|
| 3 |
size 15024
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fcb9b1e0f74c5f2631e58aded928e5d64789892339a1cd1a1bb054b2a8717bf3
|
| 3 |
size 15024
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e93d6940d870db9ebba78cee7722d0384b494610e71e7f8b2e22bb0fd8e406ed
|
| 3 |
size 15024
|
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 sha256:2a024530de56227bb3ef9eb28b732e8ef3d765c77ebd0a0c5bc59f62e1682f1a
|
| 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": 5.193,
|
| 780 |
"eval_steps_per_second": 0.169,
|
| 781 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 782 |
}
|
| 783 |
],
|
| 784 |
"logging_steps": 1,
|
|
@@ -798,7 +1558,7 @@
|
|
| 798 |
"attributes": {}
|
| 799 |
}
|
| 800 |
},
|
| 801 |
-
"total_flos":
|
| 802 |
"train_batch_size": 6,
|
| 803 |
"trial_name": null,
|
| 804 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.2510460251046025,
|
| 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": 5.193,
|
| 780 |
"eval_steps_per_second": 0.169,
|
| 781 |
"step": 100
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.12677824267782425,
|
| 785 |
+
"grad_norm": 3.9941301345825195,
|
| 786 |
+
"learning_rate": 1.9957579575884978e-05,
|
| 787 |
+
"loss": 0.3364,
|
| 788 |
+
"step": 101
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.1280334728033473,
|
| 792 |
+
"grad_norm": 3.4411442279815674,
|
| 793 |
+
"learning_rate": 1.995344655617815e-05,
|
| 794 |
+
"loss": 0.3481,
|
| 795 |
+
"step": 102
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.1292887029288703,
|
| 799 |
+
"grad_norm": 3.1757569313049316,
|
| 800 |
+
"learning_rate": 1.9949121913975275e-05,
|
| 801 |
+
"loss": 0.3447,
|
| 802 |
+
"step": 103
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.1305439330543933,
|
| 806 |
+
"grad_norm": 5.263054847717285,
|
| 807 |
+
"learning_rate": 1.994460573253382e-05,
|
| 808 |
+
"loss": 0.3705,
|
| 809 |
+
"step": 104
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.13179916317991633,
|
| 813 |
+
"grad_norm": 4.291073322296143,
|
| 814 |
+
"learning_rate": 1.9939898098798736e-05,
|
| 815 |
+
"loss": 0.349,
|
| 816 |
+
"step": 105
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.13305439330543933,
|
| 820 |
+
"grad_norm": 4.688785076141357,
|
| 821 |
+
"learning_rate": 1.9934999103400797e-05,
|
| 822 |
+
"loss": 0.2573,
|
| 823 |
+
"step": 106
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.13430962343096234,
|
| 827 |
+
"grad_norm": 3.483659505844116,
|
| 828 |
+
"learning_rate": 1.992990884065484e-05,
|
| 829 |
+
"loss": 0.2812,
|
| 830 |
+
"step": 107
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.13556485355648534,
|
| 834 |
+
"grad_norm": 5.222522258758545,
|
| 835 |
+
"learning_rate": 1.9924627408557963e-05,
|
| 836 |
+
"loss": 0.3208,
|
| 837 |
+
"step": 108
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.13682008368200838,
|
| 841 |
+
"grad_norm": 3.1946051120758057,
|
| 842 |
+
"learning_rate": 1.991915490878763e-05,
|
| 843 |
+
"loss": 0.4041,
|
| 844 |
+
"step": 109
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.13807531380753138,
|
| 848 |
+
"grad_norm": 2.834019899368286,
|
| 849 |
+
"learning_rate": 1.9913491446699715e-05,
|
| 850 |
+
"loss": 0.2989,
|
| 851 |
+
"step": 110
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.13933054393305438,
|
| 855 |
+
"grad_norm": 4.4058380126953125,
|
| 856 |
+
"learning_rate": 1.9907637131326475e-05,
|
| 857 |
+
"loss": 0.3247,
|
| 858 |
+
"step": 111
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.14058577405857742,
|
| 862 |
+
"grad_norm": 4.437101364135742,
|
| 863 |
+
"learning_rate": 1.9901592075374447e-05,
|
| 864 |
+
"loss": 0.3487,
|
| 865 |
+
"step": 112
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.14184100418410042,
|
| 869 |
+
"grad_norm": 3.1267802715301514,
|
| 870 |
+
"learning_rate": 1.989535639522229e-05,
|
| 871 |
+
"loss": 0.2741,
|
| 872 |
+
"step": 113
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.14309623430962343,
|
| 876 |
+
"grad_norm": 3.8325576782226562,
|
| 877 |
+
"learning_rate": 1.988893021091853e-05,
|
| 878 |
+
"loss": 0.385,
|
| 879 |
+
"step": 114
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.14435146443514643,
|
| 883 |
+
"grad_norm": 4.569618225097656,
|
| 884 |
+
"learning_rate": 1.9882313646179247e-05,
|
| 885 |
+
"loss": 0.3595,
|
| 886 |
+
"step": 115
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.14560669456066946,
|
| 890 |
+
"grad_norm": 5.401278972625732,
|
| 891 |
+
"learning_rate": 1.9875506828385723e-05,
|
| 892 |
+
"loss": 0.2875,
|
| 893 |
+
"step": 116
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.14686192468619247,
|
| 897 |
+
"grad_norm": 3.8402180671691895,
|
| 898 |
+
"learning_rate": 1.9868509888581945e-05,
|
| 899 |
+
"loss": 0.3079,
|
| 900 |
+
"step": 117
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.14811715481171547,
|
| 904 |
+
"grad_norm": 3.5551564693450928,
|
| 905 |
+
"learning_rate": 1.986132296147212e-05,
|
| 906 |
+
"loss": 0.3157,
|
| 907 |
+
"step": 118
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.1493723849372385,
|
| 911 |
+
"grad_norm": 9.136929512023926,
|
| 912 |
+
"learning_rate": 1.9853946185418056e-05,
|
| 913 |
+
"loss": 0.4209,
|
| 914 |
+
"step": 119
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.1506276150627615,
|
| 918 |
+
"grad_norm": 3.290203332901001,
|
| 919 |
+
"learning_rate": 1.9846379702436518e-05,
|
| 920 |
+
"loss": 0.2959,
|
| 921 |
+
"step": 120
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.1506276150627615,
|
| 925 |
+
"eval_accuracy": 0.8211920529801324,
|
| 926 |
+
"eval_f1": 0.6197183098591549,
|
| 927 |
+
"eval_loss": 0.37509337067604065,
|
| 928 |
+
"eval_precision": 0.7674418604651163,
|
| 929 |
+
"eval_recall": 0.5196850393700787,
|
| 930 |
+
"eval_runtime": 52.5731,
|
| 931 |
+
"eval_samples_per_second": 5.269,
|
| 932 |
+
"eval_steps_per_second": 0.171,
|
| 933 |
+
"step": 120
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.15188284518828452,
|
| 937 |
+
"grad_norm": 4.515352725982666,
|
| 938 |
+
"learning_rate": 1.983862365819648e-05,
|
| 939 |
+
"loss": 0.3283,
|
| 940 |
+
"step": 121
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.15313807531380752,
|
| 944 |
+
"grad_norm": 3.97063946723938,
|
| 945 |
+
"learning_rate": 1.9830678202016324e-05,
|
| 946 |
+
"loss": 0.3505,
|
| 947 |
+
"step": 122
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.15439330543933055,
|
| 951 |
+
"grad_norm": 4.553818225860596,
|
| 952 |
+
"learning_rate": 1.982254348686097e-05,
|
| 953 |
+
"loss": 0.313,
|
| 954 |
+
"step": 123
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.15564853556485356,
|
| 958 |
+
"grad_norm": 3.5846359729766846,
|
| 959 |
+
"learning_rate": 1.981421966933893e-05,
|
| 960 |
+
"loss": 0.35,
|
| 961 |
+
"step": 124
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.15690376569037656,
|
| 965 |
+
"grad_norm": 5.479614734649658,
|
| 966 |
+
"learning_rate": 1.9805706909699283e-05,
|
| 967 |
+
"loss": 0.3134,
|
| 968 |
+
"step": 125
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.1581589958158996,
|
| 972 |
+
"grad_norm": 3.6926157474517822,
|
| 973 |
+
"learning_rate": 1.9797005371828603e-05,
|
| 974 |
+
"loss": 0.3659,
|
| 975 |
+
"step": 126
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.1594142259414226,
|
| 979 |
+
"grad_norm": 4.4174957275390625,
|
| 980 |
+
"learning_rate": 1.97881152232478e-05,
|
| 981 |
+
"loss": 0.3069,
|
| 982 |
+
"step": 127
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.1606694560669456,
|
| 986 |
+
"grad_norm": 2.855861186981201,
|
| 987 |
+
"learning_rate": 1.9779036635108892e-05,
|
| 988 |
+
"loss": 0.2748,
|
| 989 |
+
"step": 128
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.1619246861924686,
|
| 993 |
+
"grad_norm": 3.4113943576812744,
|
| 994 |
+
"learning_rate": 1.976976978219171e-05,
|
| 995 |
+
"loss": 0.2942,
|
| 996 |
+
"step": 129
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.16317991631799164,
|
| 1000 |
+
"grad_norm": 2.8706114292144775,
|
| 1001 |
+
"learning_rate": 1.9760314842900537e-05,
|
| 1002 |
+
"loss": 0.26,
|
| 1003 |
+
"step": 130
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.16443514644351465,
|
| 1007 |
+
"grad_norm": 3.3289883136749268,
|
| 1008 |
+
"learning_rate": 1.975067199926067e-05,
|
| 1009 |
+
"loss": 0.2942,
|
| 1010 |
+
"step": 131
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.16569037656903765,
|
| 1014 |
+
"grad_norm": 3.1963343620300293,
|
| 1015 |
+
"learning_rate": 1.9740841436914917e-05,
|
| 1016 |
+
"loss": 0.3404,
|
| 1017 |
+
"step": 132
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.16694560669456068,
|
| 1021 |
+
"grad_norm": 4.106410026550293,
|
| 1022 |
+
"learning_rate": 1.9730823345120024e-05,
|
| 1023 |
+
"loss": 0.3645,
|
| 1024 |
+
"step": 133
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.1682008368200837,
|
| 1028 |
+
"grad_norm": 3.587475299835205,
|
| 1029 |
+
"learning_rate": 1.9720617916743022e-05,
|
| 1030 |
+
"loss": 0.2905,
|
| 1031 |
+
"step": 134
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.1694560669456067,
|
| 1035 |
+
"grad_norm": 4.643335819244385,
|
| 1036 |
+
"learning_rate": 1.971022534825754e-05,
|
| 1037 |
+
"loss": 0.3199,
|
| 1038 |
+
"step": 135
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.1707112970711297,
|
| 1042 |
+
"grad_norm": 3.8745625019073486,
|
| 1043 |
+
"learning_rate": 1.9699645839739987e-05,
|
| 1044 |
+
"loss": 0.3276,
|
| 1045 |
+
"step": 136
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.17196652719665273,
|
| 1049 |
+
"grad_norm": 4.443915367126465,
|
| 1050 |
+
"learning_rate": 1.9688879594865726e-05,
|
| 1051 |
+
"loss": 0.1989,
|
| 1052 |
+
"step": 137
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.17322175732217573,
|
| 1056 |
+
"grad_norm": 3.165154218673706,
|
| 1057 |
+
"learning_rate": 1.9677926820905143e-05,
|
| 1058 |
+
"loss": 0.2877,
|
| 1059 |
+
"step": 138
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.17447698744769874,
|
| 1063 |
+
"grad_norm": 3.396127462387085,
|
| 1064 |
+
"learning_rate": 1.9666787728719664e-05,
|
| 1065 |
+
"loss": 0.2869,
|
| 1066 |
+
"step": 139
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.17573221757322174,
|
| 1070 |
+
"grad_norm": 4.032714366912842,
|
| 1071 |
+
"learning_rate": 1.9655462532757677e-05,
|
| 1072 |
+
"loss": 0.336,
|
| 1073 |
+
"step": 140
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.17573221757322174,
|
| 1077 |
+
"eval_accuracy": 0.8278145695364238,
|
| 1078 |
+
"eval_f1": 0.6060606060606061,
|
| 1079 |
+
"eval_loss": 0.3764040172100067,
|
| 1080 |
+
"eval_precision": 0.8450704225352113,
|
| 1081 |
+
"eval_recall": 0.47244094488188976,
|
| 1082 |
+
"eval_runtime": 52.6872,
|
| 1083 |
+
"eval_samples_per_second": 5.257,
|
| 1084 |
+
"eval_steps_per_second": 0.171,
|
| 1085 |
+
"step": 140
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.17698744769874478,
|
| 1089 |
+
"grad_norm": 2.6727371215820312,
|
| 1090 |
+
"learning_rate": 1.9643951451050428e-05,
|
| 1091 |
+
"loss": 0.2636,
|
| 1092 |
+
"step": 141
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 0.17824267782426778,
|
| 1096 |
+
"grad_norm": 3.8816864490509033,
|
| 1097 |
+
"learning_rate": 1.9632254705207813e-05,
|
| 1098 |
+
"loss": 0.3208,
|
| 1099 |
+
"step": 142
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.17949790794979079,
|
| 1103 |
+
"grad_norm": 3.4616892337799072,
|
| 1104 |
+
"learning_rate": 1.9620372520414098e-05,
|
| 1105 |
+
"loss": 0.3218,
|
| 1106 |
+
"step": 143
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.18075313807531382,
|
| 1110 |
+
"grad_norm": 4.056252479553223,
|
| 1111 |
+
"learning_rate": 1.9608305125423608e-05,
|
| 1112 |
+
"loss": 0.2844,
|
| 1113 |
+
"step": 144
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.18200836820083682,
|
| 1117 |
+
"grad_norm": 5.902234077453613,
|
| 1118 |
+
"learning_rate": 1.9596052752556308e-05,
|
| 1119 |
+
"loss": 0.2497,
|
| 1120 |
+
"step": 145
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.18326359832635983,
|
| 1124 |
+
"grad_norm": 4.0488996505737305,
|
| 1125 |
+
"learning_rate": 1.958361563769333e-05,
|
| 1126 |
+
"loss": 0.2764,
|
| 1127 |
+
"step": 146
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 0.18451882845188283,
|
| 1131 |
+
"grad_norm": 4.619633197784424,
|
| 1132 |
+
"learning_rate": 1.957099402027244e-05,
|
| 1133 |
+
"loss": 0.3775,
|
| 1134 |
+
"step": 147
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.18577405857740587,
|
| 1138 |
+
"grad_norm": 4.491790294647217,
|
| 1139 |
+
"learning_rate": 1.9558188143283425e-05,
|
| 1140 |
+
"loss": 0.4185,
|
| 1141 |
+
"step": 148
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.18702928870292887,
|
| 1145 |
+
"grad_norm": 9.393437385559082,
|
| 1146 |
+
"learning_rate": 1.954519825326341e-05,
|
| 1147 |
+
"loss": 0.292,
|
| 1148 |
+
"step": 149
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.18828451882845187,
|
| 1152 |
+
"grad_norm": 9.774816513061523,
|
| 1153 |
+
"learning_rate": 1.9532024600292115e-05,
|
| 1154 |
+
"loss": 0.341,
|
| 1155 |
+
"step": 150
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.1895397489539749,
|
| 1159 |
+
"grad_norm": 9.051419258117676,
|
| 1160 |
+
"learning_rate": 1.9518667437987045e-05,
|
| 1161 |
+
"loss": 0.3125,
|
| 1162 |
+
"step": 151
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.1907949790794979,
|
| 1166 |
+
"grad_norm": 4.726169586181641,
|
| 1167 |
+
"learning_rate": 1.9505127023498603e-05,
|
| 1168 |
+
"loss": 0.3283,
|
| 1169 |
+
"step": 152
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 0.19205020920502092,
|
| 1173 |
+
"grad_norm": 3.818352222442627,
|
| 1174 |
+
"learning_rate": 1.9491403617505134e-05,
|
| 1175 |
+
"loss": 0.2696,
|
| 1176 |
+
"step": 153
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.19330543933054392,
|
| 1180 |
+
"grad_norm": 4.901086330413818,
|
| 1181 |
+
"learning_rate": 1.9477497484207922e-05,
|
| 1182 |
+
"loss": 0.2927,
|
| 1183 |
+
"step": 154
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.19456066945606695,
|
| 1187 |
+
"grad_norm": 2.7958414554595947,
|
| 1188 |
+
"learning_rate": 1.9463408891326088e-05,
|
| 1189 |
+
"loss": 0.2544,
|
| 1190 |
+
"step": 155
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.19581589958158996,
|
| 1194 |
+
"grad_norm": 3.541666030883789,
|
| 1195 |
+
"learning_rate": 1.9449138110091444e-05,
|
| 1196 |
+
"loss": 0.2723,
|
| 1197 |
+
"step": 156
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.19707112970711296,
|
| 1201 |
+
"grad_norm": 4.369930744171143,
|
| 1202 |
+
"learning_rate": 1.9434685415243267e-05,
|
| 1203 |
+
"loss": 0.3121,
|
| 1204 |
+
"step": 157
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.198326359832636,
|
| 1208 |
+
"grad_norm": 4.061751842498779,
|
| 1209 |
+
"learning_rate": 1.9420051085023006e-05,
|
| 1210 |
+
"loss": 0.3238,
|
| 1211 |
+
"step": 158
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 0.199581589958159,
|
| 1215 |
+
"grad_norm": 5.1077446937561035,
|
| 1216 |
+
"learning_rate": 1.940523540116895e-05,
|
| 1217 |
+
"loss": 0.2935,
|
| 1218 |
+
"step": 159
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 0.200836820083682,
|
| 1222 |
+
"grad_norm": 3.7316763401031494,
|
| 1223 |
+
"learning_rate": 1.9390238648910765e-05,
|
| 1224 |
+
"loss": 0.3239,
|
| 1225 |
+
"step": 160
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 0.200836820083682,
|
| 1229 |
+
"eval_accuracy": 0.8200883002207505,
|
| 1230 |
+
"eval_f1": 0.5788113695090439,
|
| 1231 |
+
"eval_loss": 0.3607686161994934,
|
| 1232 |
+
"eval_precision": 0.8421052631578947,
|
| 1233 |
+
"eval_recall": 0.4409448818897638,
|
| 1234 |
+
"eval_runtime": 53.735,
|
| 1235 |
+
"eval_samples_per_second": 5.155,
|
| 1236 |
+
"eval_steps_per_second": 0.167,
|
| 1237 |
+
"step": 160
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.202092050209205,
|
| 1241 |
+
"grad_norm": 4.719494819641113,
|
| 1242 |
+
"learning_rate": 1.9375061116964032e-05,
|
| 1243 |
+
"loss": 0.3164,
|
| 1244 |
+
"step": 161
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.20334728033472804,
|
| 1248 |
+
"grad_norm": 3.245194435119629,
|
| 1249 |
+
"learning_rate": 1.935970309752469e-05,
|
| 1250 |
+
"loss": 0.2923,
|
| 1251 |
+
"step": 162
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.20460251046025105,
|
| 1255 |
+
"grad_norm": 4.244296550750732,
|
| 1256 |
+
"learning_rate": 1.9344164886263375e-05,
|
| 1257 |
+
"loss": 0.2891,
|
| 1258 |
+
"step": 163
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 0.20585774058577405,
|
| 1262 |
+
"grad_norm": 5.457589149475098,
|
| 1263 |
+
"learning_rate": 1.932844678231977e-05,
|
| 1264 |
+
"loss": 0.3057,
|
| 1265 |
+
"step": 164
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 0.20711297071129708,
|
| 1269 |
+
"grad_norm": 4.439499378204346,
|
| 1270 |
+
"learning_rate": 1.9312549088296838e-05,
|
| 1271 |
+
"loss": 0.2107,
|
| 1272 |
+
"step": 165
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 0.2083682008368201,
|
| 1276 |
+
"grad_norm": 5.0200653076171875,
|
| 1277 |
+
"learning_rate": 1.929647211025497e-05,
|
| 1278 |
+
"loss": 0.2859,
|
| 1279 |
+
"step": 166
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 0.2096234309623431,
|
| 1283 |
+
"grad_norm": 3.7708117961883545,
|
| 1284 |
+
"learning_rate": 1.9280216157706113e-05,
|
| 1285 |
+
"loss": 0.2816,
|
| 1286 |
+
"step": 167
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.2108786610878661,
|
| 1290 |
+
"grad_norm": 3.947610855102539,
|
| 1291 |
+
"learning_rate": 1.9263781543607817e-05,
|
| 1292 |
+
"loss": 0.2431,
|
| 1293 |
+
"step": 168
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 0.21213389121338913,
|
| 1297 |
+
"grad_norm": 3.28195858001709,
|
| 1298 |
+
"learning_rate": 1.9247168584357195e-05,
|
| 1299 |
+
"loss": 0.296,
|
| 1300 |
+
"step": 169
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 0.21338912133891214,
|
| 1304 |
+
"grad_norm": 3.6983871459960938,
|
| 1305 |
+
"learning_rate": 1.923037759978484e-05,
|
| 1306 |
+
"loss": 0.3003,
|
| 1307 |
+
"step": 170
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 0.21464435146443514,
|
| 1311 |
+
"grad_norm": 4.456281661987305,
|
| 1312 |
+
"learning_rate": 1.921340891314867e-05,
|
| 1313 |
+
"loss": 0.2493,
|
| 1314 |
+
"step": 171
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.21589958158995817,
|
| 1318 |
+
"grad_norm": 3.2370941638946533,
|
| 1319 |
+
"learning_rate": 1.9196262851127695e-05,
|
| 1320 |
+
"loss": 0.2353,
|
| 1321 |
+
"step": 172
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.21715481171548118,
|
| 1325 |
+
"grad_norm": 2.977496862411499,
|
| 1326 |
+
"learning_rate": 1.9178939743815735e-05,
|
| 1327 |
+
"loss": 0.3062,
|
| 1328 |
+
"step": 173
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.21841004184100418,
|
| 1332 |
+
"grad_norm": 5.293909072875977,
|
| 1333 |
+
"learning_rate": 1.9161439924715063e-05,
|
| 1334 |
+
"loss": 0.2646,
|
| 1335 |
+
"step": 174
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 0.2196652719665272,
|
| 1339 |
+
"grad_norm": 3.4083428382873535,
|
| 1340 |
+
"learning_rate": 1.9143763730729987e-05,
|
| 1341 |
+
"loss": 0.2305,
|
| 1342 |
+
"step": 175
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 0.22092050209205022,
|
| 1346 |
+
"grad_norm": 2.7759830951690674,
|
| 1347 |
+
"learning_rate": 1.9125911502160365e-05,
|
| 1348 |
+
"loss": 0.2554,
|
| 1349 |
+
"step": 176
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 0.22217573221757322,
|
| 1353 |
+
"grad_norm": 3.9626009464263916,
|
| 1354 |
+
"learning_rate": 1.9107883582695043e-05,
|
| 1355 |
+
"loss": 0.2789,
|
| 1356 |
+
"step": 177
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 0.22343096234309623,
|
| 1360 |
+
"grad_norm": 3.340153932571411,
|
| 1361 |
+
"learning_rate": 1.9089680319405252e-05,
|
| 1362 |
+
"loss": 0.2874,
|
| 1363 |
+
"step": 178
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 0.22468619246861923,
|
| 1367 |
+
"grad_norm": 3.277308702468872,
|
| 1368 |
+
"learning_rate": 1.9071302062737915e-05,
|
| 1369 |
+
"loss": 0.1978,
|
| 1370 |
+
"step": 179
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.22594142259414227,
|
| 1374 |
+
"grad_norm": 5.420035362243652,
|
| 1375 |
+
"learning_rate": 1.905274916650891e-05,
|
| 1376 |
+
"loss": 0.2767,
|
| 1377 |
+
"step": 180
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 0.22594142259414227,
|
| 1381 |
+
"eval_accuracy": 0.8543046357615894,
|
| 1382 |
+
"eval_f1": 0.7066666666666667,
|
| 1383 |
+
"eval_loss": 0.3361983299255371,
|
| 1384 |
+
"eval_precision": 0.8112244897959183,
|
| 1385 |
+
"eval_recall": 0.6259842519685039,
|
| 1386 |
+
"eval_runtime": 51.9639,
|
| 1387 |
+
"eval_samples_per_second": 5.331,
|
| 1388 |
+
"eval_steps_per_second": 0.173,
|
| 1389 |
+
"step": 180
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.22719665271966527,
|
| 1393 |
+
"grad_norm": 6.328350067138672,
|
| 1394 |
+
"learning_rate": 1.903402198789625e-05,
|
| 1395 |
+
"loss": 0.3489,
|
| 1396 |
+
"step": 181
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.22845188284518828,
|
| 1400 |
+
"grad_norm": 3.141185760498047,
|
| 1401 |
+
"learning_rate": 1.9015120887433215e-05,
|
| 1402 |
+
"loss": 0.3043,
|
| 1403 |
+
"step": 182
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.2297071129707113,
|
| 1407 |
+
"grad_norm": 3.633781909942627,
|
| 1408 |
+
"learning_rate": 1.8996046229001407e-05,
|
| 1409 |
+
"loss": 0.3081,
|
| 1410 |
+
"step": 183
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.2309623430962343,
|
| 1414 |
+
"grad_norm": 3.6458773612976074,
|
| 1415 |
+
"learning_rate": 1.897679837982373e-05,
|
| 1416 |
+
"loss": 0.2259,
|
| 1417 |
+
"step": 184
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.23221757322175732,
|
| 1421 |
+
"grad_norm": 7.069568634033203,
|
| 1422 |
+
"learning_rate": 1.895737771045736e-05,
|
| 1423 |
+
"loss": 0.347,
|
| 1424 |
+
"step": 185
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.23347280334728032,
|
| 1428 |
+
"grad_norm": 5.888752460479736,
|
| 1429 |
+
"learning_rate": 1.8937784594786562e-05,
|
| 1430 |
+
"loss": 0.2361,
|
| 1431 |
+
"step": 186
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.23472803347280335,
|
| 1435 |
+
"grad_norm": 3.553389072418213,
|
| 1436 |
+
"learning_rate": 1.8918019410015527e-05,
|
| 1437 |
+
"loss": 0.2504,
|
| 1438 |
+
"step": 187
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.23598326359832636,
|
| 1442 |
+
"grad_norm": 3.6231913566589355,
|
| 1443 |
+
"learning_rate": 1.8898082536661097e-05,
|
| 1444 |
+
"loss": 0.2558,
|
| 1445 |
+
"step": 188
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.23723849372384936,
|
| 1449 |
+
"grad_norm": 3.3635237216949463,
|
| 1450 |
+
"learning_rate": 1.887797435854543e-05,
|
| 1451 |
+
"loss": 0.2867,
|
| 1452 |
+
"step": 189
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.2384937238493724,
|
| 1456 |
+
"grad_norm": 4.0168538093566895,
|
| 1457 |
+
"learning_rate": 1.885769526278865e-05,
|
| 1458 |
+
"loss": 0.3195,
|
| 1459 |
+
"step": 190
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.2397489539748954,
|
| 1463 |
+
"grad_norm": 4.260074615478516,
|
| 1464 |
+
"learning_rate": 1.8837245639801332e-05,
|
| 1465 |
+
"loss": 0.2861,
|
| 1466 |
+
"step": 191
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.2410041841004184,
|
| 1470 |
+
"grad_norm": 3.299710988998413,
|
| 1471 |
+
"learning_rate": 1.8816625883277044e-05,
|
| 1472 |
+
"loss": 0.2454,
|
| 1473 |
+
"step": 192
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.2422594142259414,
|
| 1477 |
+
"grad_norm": 4.68196439743042,
|
| 1478 |
+
"learning_rate": 1.8795836390184727e-05,
|
| 1479 |
+
"loss": 0.2976,
|
| 1480 |
+
"step": 193
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.24351464435146444,
|
| 1484 |
+
"grad_norm": 4.414516925811768,
|
| 1485 |
+
"learning_rate": 1.8774877560761082e-05,
|
| 1486 |
+
"loss": 0.2814,
|
| 1487 |
+
"step": 194
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.24476987447698745,
|
| 1491 |
+
"grad_norm": 4.108029365539551,
|
| 1492 |
+
"learning_rate": 1.8753749798502845e-05,
|
| 1493 |
+
"loss": 0.2478,
|
| 1494 |
+
"step": 195
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.24602510460251045,
|
| 1498 |
+
"grad_norm": 3.553065061569214,
|
| 1499 |
+
"learning_rate": 1.8732453510159025e-05,
|
| 1500 |
+
"loss": 0.2221,
|
| 1501 |
+
"step": 196
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.24728033472803349,
|
| 1505 |
+
"grad_norm": 3.1897339820861816,
|
| 1506 |
+
"learning_rate": 1.871098910572308e-05,
|
| 1507 |
+
"loss": 0.3001,
|
| 1508 |
+
"step": 197
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.2485355648535565,
|
| 1512 |
+
"grad_norm": 4.416936874389648,
|
| 1513 |
+
"learning_rate": 1.8689356998425007e-05,
|
| 1514 |
+
"loss": 0.3109,
|
| 1515 |
+
"step": 198
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.2497907949790795,
|
| 1519 |
+
"grad_norm": 3.162482976913452,
|
| 1520 |
+
"learning_rate": 1.8667557604723404e-05,
|
| 1521 |
+
"loss": 0.3104,
|
| 1522 |
+
"step": 199
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.2510460251046025,
|
| 1526 |
+
"grad_norm": 3.864384651184082,
|
| 1527 |
+
"learning_rate": 1.864559134429745e-05,
|
| 1528 |
+
"loss": 0.276,
|
| 1529 |
+
"step": 200
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.2510460251046025,
|
| 1533 |
+
"eval_accuracy": 0.8388520971302428,
|
| 1534 |
+
"eval_f1": 0.6439024390243903,
|
| 1535 |
+
"eval_loss": 0.3405630886554718,
|
| 1536 |
+
"eval_precision": 0.8461538461538461,
|
| 1537 |
+
"eval_recall": 0.5196850393700787,
|
| 1538 |
+
"eval_runtime": 52.2841,
|
| 1539 |
+
"eval_samples_per_second": 5.298,
|
| 1540 |
+
"eval_steps_per_second": 0.172,
|
| 1541 |
+
"step": 200
|
| 1542 |
}
|
| 1543 |
],
|
| 1544 |
"logging_steps": 1,
|
|
|
|
| 1558 |
"attributes": {}
|
| 1559 |
}
|
| 1560 |
},
|
| 1561 |
+
"total_flos": 7.32374373582766e+16,
|
| 1562 |
"train_batch_size": 6,
|
| 1563 |
"trial_name": null,
|
| 1564 |
"trial_params": null
|