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 sha256:534c42076a246c6da5a00ffd44149115e4d50f42ee2ee4186468f5798dbb1ccf
|
| 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:a663829a3b941a4048ffcc2de6e0512c94c579c7a489c3e009cf30a2a53e694d
|
| 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:edb36318fb7a6485a66f873289f77615fb974210ae47a75c352e9d4d2d4426d8
|
| 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 sha256:6802368c5f12cbf6130a30d93da9380768ce9f37bbb6bd21b02b9e602182fbcd
|
| 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:b7a755e47d66d671add11d66f6099b1dd83a6c13121c2ef15fdfdde9a3177177
|
| 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:cb3edd1ccbdea3e3f2d56cd3a4646f38afe7ca93815da1414f65fe03b9b673a2
|
| 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 sha256:4599060b53a4939dd8f840249e269e7830878980b2cf9fafb1b39f1203aaa960
|
| 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:e02bfd751b50b769ac97d99b9a8385f2091de1188f94cbd07e5f93afeae257da
|
| 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 sha256:51118da612b8171b6675abc2602bed7ce97edefe29a8f466ea28ed45a226a206
|
| 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 sha256:62710a98a3c7bb382ced086930b4b07cc7dc4c19e47a9f58b3464ec46167033a
|
| 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 sha256:3e4c5366934a63a65595c9de33e3c7b09bdf1751d64db6f76892cbdd781442b1
|
| 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:f747b50387c790da0638d4436a970217188c80f6b7d2d6cc099b8ddf28c5197c
|
| 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:a81a95c7d38a4c117734641266299d17605df7b45470c7b744f36bacf620813f
|
| 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": 5.861,
|
| 780 |
"eval_steps_per_second": 0.201,
|
| 781 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 782 |
}
|
| 783 |
],
|
| 784 |
"logging_steps": 1,
|
|
@@ -798,7 +1558,7 @@
|
|
| 798 |
"attributes": {}
|
| 799 |
}
|
| 800 |
},
|
| 801 |
-
"total_flos":
|
| 802 |
"train_batch_size": 8,
|
| 803 |
"trial_name": null,
|
| 804 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.1694915254237288,
|
| 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": 5.861,
|
| 780 |
"eval_steps_per_second": 0.201,
|
| 781 |
"step": 100
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.08559322033898305,
|
| 785 |
+
"grad_norm": 3.773057222366333,
|
| 786 |
+
"learning_rate": 1.711864406779661e-05,
|
| 787 |
+
"loss": 0.1186,
|
| 788 |
+
"step": 101
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.08644067796610169,
|
| 792 |
+
"grad_norm": 1.8200677633285522,
|
| 793 |
+
"learning_rate": 1.728813559322034e-05,
|
| 794 |
+
"loss": 0.0917,
|
| 795 |
+
"step": 102
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.08728813559322034,
|
| 799 |
+
"grad_norm": 5.035698890686035,
|
| 800 |
+
"learning_rate": 1.745762711864407e-05,
|
| 801 |
+
"loss": 0.1199,
|
| 802 |
+
"step": 103
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.08813559322033898,
|
| 806 |
+
"grad_norm": 4.666184425354004,
|
| 807 |
+
"learning_rate": 1.76271186440678e-05,
|
| 808 |
+
"loss": 0.1286,
|
| 809 |
+
"step": 104
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.08898305084745763,
|
| 813 |
+
"grad_norm": 2.4662208557128906,
|
| 814 |
+
"learning_rate": 1.7796610169491526e-05,
|
| 815 |
+
"loss": 0.0845,
|
| 816 |
+
"step": 105
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.08983050847457627,
|
| 820 |
+
"grad_norm": 2.8939261436462402,
|
| 821 |
+
"learning_rate": 1.7966101694915256e-05,
|
| 822 |
+
"loss": 0.0989,
|
| 823 |
+
"step": 106
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.09067796610169492,
|
| 827 |
+
"grad_norm": 4.854353904724121,
|
| 828 |
+
"learning_rate": 1.8135593220338986e-05,
|
| 829 |
+
"loss": 0.1092,
|
| 830 |
+
"step": 107
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.09152542372881356,
|
| 834 |
+
"grad_norm": 3.3702521324157715,
|
| 835 |
+
"learning_rate": 1.8305084745762713e-05,
|
| 836 |
+
"loss": 0.1564,
|
| 837 |
+
"step": 108
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.0923728813559322,
|
| 841 |
+
"grad_norm": 4.387780666351318,
|
| 842 |
+
"learning_rate": 1.8474576271186443e-05,
|
| 843 |
+
"loss": 0.1399,
|
| 844 |
+
"step": 109
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.09322033898305085,
|
| 848 |
+
"grad_norm": 4.5151495933532715,
|
| 849 |
+
"learning_rate": 1.864406779661017e-05,
|
| 850 |
+
"loss": 0.0979,
|
| 851 |
+
"step": 110
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.0940677966101695,
|
| 855 |
+
"grad_norm": 3.65556001663208,
|
| 856 |
+
"learning_rate": 1.88135593220339e-05,
|
| 857 |
+
"loss": 0.1338,
|
| 858 |
+
"step": 111
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.09491525423728814,
|
| 862 |
+
"grad_norm": 2.1862547397613525,
|
| 863 |
+
"learning_rate": 1.898305084745763e-05,
|
| 864 |
+
"loss": 0.0791,
|
| 865 |
+
"step": 112
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.09576271186440678,
|
| 869 |
+
"grad_norm": 5.004955291748047,
|
| 870 |
+
"learning_rate": 1.9152542372881357e-05,
|
| 871 |
+
"loss": 0.0972,
|
| 872 |
+
"step": 113
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.09661016949152543,
|
| 876 |
+
"grad_norm": 3.057615041732788,
|
| 877 |
+
"learning_rate": 1.9322033898305087e-05,
|
| 878 |
+
"loss": 0.0611,
|
| 879 |
+
"step": 114
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.09745762711864407,
|
| 883 |
+
"grad_norm": 5.858815670013428,
|
| 884 |
+
"learning_rate": 1.9491525423728814e-05,
|
| 885 |
+
"loss": 0.1005,
|
| 886 |
+
"step": 115
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.09830508474576272,
|
| 890 |
+
"grad_norm": 3.9090657234191895,
|
| 891 |
+
"learning_rate": 1.9661016949152545e-05,
|
| 892 |
+
"loss": 0.1294,
|
| 893 |
+
"step": 116
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.09915254237288136,
|
| 897 |
+
"grad_norm": 2.838143825531006,
|
| 898 |
+
"learning_rate": 1.9830508474576275e-05,
|
| 899 |
+
"loss": 0.089,
|
| 900 |
+
"step": 117
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.1,
|
| 904 |
+
"grad_norm": 6.167634963989258,
|
| 905 |
+
"learning_rate": 2e-05,
|
| 906 |
+
"loss": 0.1497,
|
| 907 |
+
"step": 118
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.10084745762711865,
|
| 911 |
+
"grad_norm": 2.7352752685546875,
|
| 912 |
+
"learning_rate": 1.9999956245734325e-05,
|
| 913 |
+
"loss": 0.0708,
|
| 914 |
+
"step": 119
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.1016949152542373,
|
| 918 |
+
"grad_norm": 5.203207969665527,
|
| 919 |
+
"learning_rate": 1.9999824983320176e-05,
|
| 920 |
+
"loss": 0.0971,
|
| 921 |
+
"step": 120
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.1016949152542373,
|
| 925 |
+
"eval_accuracy": 0.9914893617021276,
|
| 926 |
+
"eval_f1": 0.9846938775510204,
|
| 927 |
+
"eval_loss": 0.04216673597693443,
|
| 928 |
+
"eval_precision": 0.9747474747474747,
|
| 929 |
+
"eval_recall": 0.9948453608247423,
|
| 930 |
+
"eval_runtime": 50.369,
|
| 931 |
+
"eval_samples_per_second": 5.797,
|
| 932 |
+
"eval_steps_per_second": 0.199,
|
| 933 |
+
"step": 120
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.10254237288135593,
|
| 937 |
+
"grad_norm": 3.716186046600342,
|
| 938 |
+
"learning_rate": 1.999960621390622e-05,
|
| 939 |
+
"loss": 0.0875,
|
| 940 |
+
"step": 121
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.10338983050847457,
|
| 944 |
+
"grad_norm": 3.5088372230529785,
|
| 945 |
+
"learning_rate": 1.9999299939406875e-05,
|
| 946 |
+
"loss": 0.0642,
|
| 947 |
+
"step": 122
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.10423728813559321,
|
| 951 |
+
"grad_norm": 3.1328420639038086,
|
| 952 |
+
"learning_rate": 1.9998906162502298e-05,
|
| 953 |
+
"loss": 0.0559,
|
| 954 |
+
"step": 123
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.10508474576271186,
|
| 958 |
+
"grad_norm": 1.2235671281814575,
|
| 959 |
+
"learning_rate": 1.9998424886638382e-05,
|
| 960 |
+
"loss": 0.0267,
|
| 961 |
+
"step": 124
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.1059322033898305,
|
| 965 |
+
"grad_norm": 2.749112129211426,
|
| 966 |
+
"learning_rate": 1.9997856116026692e-05,
|
| 967 |
+
"loss": 0.0547,
|
| 968 |
+
"step": 125
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.10677966101694915,
|
| 972 |
+
"grad_norm": 1.985463261604309,
|
| 973 |
+
"learning_rate": 1.999719985564446e-05,
|
| 974 |
+
"loss": 0.0422,
|
| 975 |
+
"step": 126
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.10762711864406779,
|
| 979 |
+
"grad_norm": 1.806302785873413,
|
| 980 |
+
"learning_rate": 1.999645611123453e-05,
|
| 981 |
+
"loss": 0.0311,
|
| 982 |
+
"step": 127
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.10847457627118644,
|
| 986 |
+
"grad_norm": 1.514439582824707,
|
| 987 |
+
"learning_rate": 1.9995624889305286e-05,
|
| 988 |
+
"loss": 0.0303,
|
| 989 |
+
"step": 128
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.10932203389830508,
|
| 993 |
+
"grad_norm": 4.151468276977539,
|
| 994 |
+
"learning_rate": 1.9994706197130645e-05,
|
| 995 |
+
"loss": 0.0615,
|
| 996 |
+
"step": 129
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.11016949152542373,
|
| 1000 |
+
"grad_norm": 4.021677494049072,
|
| 1001 |
+
"learning_rate": 1.9993700042749937e-05,
|
| 1002 |
+
"loss": 0.0374,
|
| 1003 |
+
"step": 130
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.11101694915254237,
|
| 1007 |
+
"grad_norm": 4.167133331298828,
|
| 1008 |
+
"learning_rate": 1.9992606434967877e-05,
|
| 1009 |
+
"loss": 0.0969,
|
| 1010 |
+
"step": 131
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.11186440677966102,
|
| 1014 |
+
"grad_norm": 2.3483352661132812,
|
| 1015 |
+
"learning_rate": 1.9991425383354462e-05,
|
| 1016 |
+
"loss": 0.0587,
|
| 1017 |
+
"step": 132
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.11271186440677966,
|
| 1021 |
+
"grad_norm": 3.196880340576172,
|
| 1022 |
+
"learning_rate": 1.99901568982449e-05,
|
| 1023 |
+
"loss": 0.0665,
|
| 1024 |
+
"step": 133
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.1135593220338983,
|
| 1028 |
+
"grad_norm": 1.4003562927246094,
|
| 1029 |
+
"learning_rate": 1.998880099073952e-05,
|
| 1030 |
+
"loss": 0.0287,
|
| 1031 |
+
"step": 134
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.11440677966101695,
|
| 1035 |
+
"grad_norm": 6.164405822753906,
|
| 1036 |
+
"learning_rate": 1.9987357672703674e-05,
|
| 1037 |
+
"loss": 0.0557,
|
| 1038 |
+
"step": 135
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.1152542372881356,
|
| 1042 |
+
"grad_norm": 4.226889133453369,
|
| 1043 |
+
"learning_rate": 1.998582695676762e-05,
|
| 1044 |
+
"loss": 0.0575,
|
| 1045 |
+
"step": 136
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.11610169491525424,
|
| 1049 |
+
"grad_norm": 3.4632186889648438,
|
| 1050 |
+
"learning_rate": 1.9984208856326433e-05,
|
| 1051 |
+
"loss": 0.0405,
|
| 1052 |
+
"step": 137
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.11694915254237288,
|
| 1056 |
+
"grad_norm": 1.7408661842346191,
|
| 1057 |
+
"learning_rate": 1.9982503385539865e-05,
|
| 1058 |
+
"loss": 0.0208,
|
| 1059 |
+
"step": 138
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.11779661016949153,
|
| 1063 |
+
"grad_norm": 4.273019790649414,
|
| 1064 |
+
"learning_rate": 1.9980710559332248e-05,
|
| 1065 |
+
"loss": 0.0439,
|
| 1066 |
+
"step": 139
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.11864406779661017,
|
| 1070 |
+
"grad_norm": 2.678131341934204,
|
| 1071 |
+
"learning_rate": 1.9978830393392338e-05,
|
| 1072 |
+
"loss": 0.0478,
|
| 1073 |
+
"step": 140
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.11864406779661017,
|
| 1077 |
+
"eval_accuracy": 0.9985815602836879,
|
| 1078 |
+
"eval_f1": 0.9974160206718347,
|
| 1079 |
+
"eval_loss": 0.0119753647595644,
|
| 1080 |
+
"eval_precision": 1.0,
|
| 1081 |
+
"eval_recall": 0.9948453608247423,
|
| 1082 |
+
"eval_runtime": 49.8796,
|
| 1083 |
+
"eval_samples_per_second": 5.854,
|
| 1084 |
+
"eval_steps_per_second": 0.2,
|
| 1085 |
+
"step": 140
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.11949152542372882,
|
| 1089 |
+
"grad_norm": 3.534904718399048,
|
| 1090 |
+
"learning_rate": 1.997686290417319e-05,
|
| 1091 |
+
"loss": 0.0676,
|
| 1092 |
+
"step": 141
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 0.12033898305084746,
|
| 1096 |
+
"grad_norm": 2.0324580669403076,
|
| 1097 |
+
"learning_rate": 1.9974808108892017e-05,
|
| 1098 |
+
"loss": 0.0522,
|
| 1099 |
+
"step": 142
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.1211864406779661,
|
| 1103 |
+
"grad_norm": 1.198857069015503,
|
| 1104 |
+
"learning_rate": 1.9972666025530027e-05,
|
| 1105 |
+
"loss": 0.0296,
|
| 1106 |
+
"step": 143
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.12203389830508475,
|
| 1110 |
+
"grad_norm": 3.2384731769561768,
|
| 1111 |
+
"learning_rate": 1.9970436672832276e-05,
|
| 1112 |
+
"loss": 0.0295,
|
| 1113 |
+
"step": 144
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.1228813559322034,
|
| 1117 |
+
"grad_norm": 2.0026895999908447,
|
| 1118 |
+
"learning_rate": 1.9968120070307503e-05,
|
| 1119 |
+
"loss": 0.0426,
|
| 1120 |
+
"step": 145
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.12372881355932204,
|
| 1124 |
+
"grad_norm": 0.9530765414237976,
|
| 1125 |
+
"learning_rate": 1.996571623822796e-05,
|
| 1126 |
+
"loss": 0.0192,
|
| 1127 |
+
"step": 146
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 0.12457627118644068,
|
| 1131 |
+
"grad_norm": 4.019771575927734,
|
| 1132 |
+
"learning_rate": 1.9963225197629223e-05,
|
| 1133 |
+
"loss": 0.0693,
|
| 1134 |
+
"step": 147
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.12542372881355932,
|
| 1138 |
+
"grad_norm": 2.6940274238586426,
|
| 1139 |
+
"learning_rate": 1.9960646970310027e-05,
|
| 1140 |
+
"loss": 0.032,
|
| 1141 |
+
"step": 148
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 0.12627118644067797,
|
| 1145 |
+
"grad_norm": 1.5307412147521973,
|
| 1146 |
+
"learning_rate": 1.995798157883206e-05,
|
| 1147 |
+
"loss": 0.0238,
|
| 1148 |
+
"step": 149
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.1271186440677966,
|
| 1152 |
+
"grad_norm": 2.437249183654785,
|
| 1153 |
+
"learning_rate": 1.995522904651977e-05,
|
| 1154 |
+
"loss": 0.0348,
|
| 1155 |
+
"step": 150
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.12796610169491526,
|
| 1159 |
+
"grad_norm": 1.985588550567627,
|
| 1160 |
+
"learning_rate": 1.995238939746016e-05,
|
| 1161 |
+
"loss": 0.0441,
|
| 1162 |
+
"step": 151
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.1288135593220339,
|
| 1166 |
+
"grad_norm": 1.3132153749465942,
|
| 1167 |
+
"learning_rate": 1.9949462656502588e-05,
|
| 1168 |
+
"loss": 0.0146,
|
| 1169 |
+
"step": 152
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 0.12966101694915255,
|
| 1173 |
+
"grad_norm": 2.7921226024627686,
|
| 1174 |
+
"learning_rate": 1.994644884925853e-05,
|
| 1175 |
+
"loss": 0.0445,
|
| 1176 |
+
"step": 153
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.13050847457627118,
|
| 1180 |
+
"grad_norm": 4.007246017456055,
|
| 1181 |
+
"learning_rate": 1.9943348002101374e-05,
|
| 1182 |
+
"loss": 0.0593,
|
| 1183 |
+
"step": 154
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 0.13135593220338984,
|
| 1187 |
+
"grad_norm": 5.683608531951904,
|
| 1188 |
+
"learning_rate": 1.9940160142166172e-05,
|
| 1189 |
+
"loss": 0.0705,
|
| 1190 |
+
"step": 155
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.13220338983050847,
|
| 1194 |
+
"grad_norm": 3.090878963470459,
|
| 1195 |
+
"learning_rate": 1.9936885297349426e-05,
|
| 1196 |
+
"loss": 0.0516,
|
| 1197 |
+
"step": 156
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.13305084745762713,
|
| 1201 |
+
"grad_norm": 1.8990111351013184,
|
| 1202 |
+
"learning_rate": 1.993352349630882e-05,
|
| 1203 |
+
"loss": 0.0218,
|
| 1204 |
+
"step": 157
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 0.13389830508474576,
|
| 1208 |
+
"grad_norm": 2.009877920150757,
|
| 1209 |
+
"learning_rate": 1.9930074768462974e-05,
|
| 1210 |
+
"loss": 0.0294,
|
| 1211 |
+
"step": 158
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 0.13474576271186442,
|
| 1215 |
+
"grad_norm": 5.6568193435668945,
|
| 1216 |
+
"learning_rate": 1.992653914399121e-05,
|
| 1217 |
+
"loss": 0.0532,
|
| 1218 |
+
"step": 159
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 0.13559322033898305,
|
| 1222 |
+
"grad_norm": 1.7055613994598389,
|
| 1223 |
+
"learning_rate": 1.992291665383325e-05,
|
| 1224 |
+
"loss": 0.0373,
|
| 1225 |
+
"step": 160
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 0.13559322033898305,
|
| 1229 |
+
"eval_accuracy": 0.9985815602836879,
|
| 1230 |
+
"eval_f1": 0.9974160206718347,
|
| 1231 |
+
"eval_loss": 0.009858837351202965,
|
| 1232 |
+
"eval_precision": 1.0,
|
| 1233 |
+
"eval_recall": 0.9948453608247423,
|
| 1234 |
+
"eval_runtime": 49.8997,
|
| 1235 |
+
"eval_samples_per_second": 5.852,
|
| 1236 |
+
"eval_steps_per_second": 0.2,
|
| 1237 |
+
"step": 160
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.13644067796610168,
|
| 1241 |
+
"grad_norm": 4.821517467498779,
|
| 1242 |
+
"learning_rate": 1.9919207329688974e-05,
|
| 1243 |
+
"loss": 0.0473,
|
| 1244 |
+
"step": 161
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.13728813559322034,
|
| 1248 |
+
"grad_norm": 3.094421863555908,
|
| 1249 |
+
"learning_rate": 1.9915411204018137e-05,
|
| 1250 |
+
"loss": 0.0347,
|
| 1251 |
+
"step": 162
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.13813559322033897,
|
| 1255 |
+
"grad_norm": 2.942777156829834,
|
| 1256 |
+
"learning_rate": 1.9911528310040073e-05,
|
| 1257 |
+
"loss": 0.0534,
|
| 1258 |
+
"step": 163
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 0.13898305084745763,
|
| 1262 |
+
"grad_norm": 1.5976642370224,
|
| 1263 |
+
"learning_rate": 1.990755868173342e-05,
|
| 1264 |
+
"loss": 0.0214,
|
| 1265 |
+
"step": 164
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 0.13983050847457626,
|
| 1269 |
+
"grad_norm": 1.8799856901168823,
|
| 1270 |
+
"learning_rate": 1.9903502353835812e-05,
|
| 1271 |
+
"loss": 0.0335,
|
| 1272 |
+
"step": 165
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 0.14067796610169492,
|
| 1276 |
+
"grad_norm": 1.2853425741195679,
|
| 1277 |
+
"learning_rate": 1.989935936184358e-05,
|
| 1278 |
+
"loss": 0.0168,
|
| 1279 |
+
"step": 166
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 0.14152542372881355,
|
| 1283 |
+
"grad_norm": 0.8122027516365051,
|
| 1284 |
+
"learning_rate": 1.9895129742011434e-05,
|
| 1285 |
+
"loss": 0.0137,
|
| 1286 |
+
"step": 167
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.1423728813559322,
|
| 1290 |
+
"grad_norm": 3.085028886795044,
|
| 1291 |
+
"learning_rate": 1.989081353135216e-05,
|
| 1292 |
+
"loss": 0.0671,
|
| 1293 |
+
"step": 168
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 0.14322033898305084,
|
| 1297 |
+
"grad_norm": 2.514724016189575,
|
| 1298 |
+
"learning_rate": 1.9886410767636284e-05,
|
| 1299 |
+
"loss": 0.0641,
|
| 1300 |
+
"step": 169
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 0.1440677966101695,
|
| 1304 |
+
"grad_norm": 1.2149631977081299,
|
| 1305 |
+
"learning_rate": 1.9881921489391738e-05,
|
| 1306 |
+
"loss": 0.024,
|
| 1307 |
+
"step": 170
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 0.14491525423728813,
|
| 1311 |
+
"grad_norm": 4.677899360656738,
|
| 1312 |
+
"learning_rate": 1.9877345735903546e-05,
|
| 1313 |
+
"loss": 0.0553,
|
| 1314 |
+
"step": 171
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.14576271186440679,
|
| 1318 |
+
"grad_norm": 0.9889124631881714,
|
| 1319 |
+
"learning_rate": 1.9872683547213446e-05,
|
| 1320 |
+
"loss": 0.025,
|
| 1321 |
+
"step": 172
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.14661016949152542,
|
| 1325 |
+
"grad_norm": 1.9920752048492432,
|
| 1326 |
+
"learning_rate": 1.9867934964119575e-05,
|
| 1327 |
+
"loss": 0.0254,
|
| 1328 |
+
"step": 173
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.14745762711864407,
|
| 1332 |
+
"grad_norm": 1.7707115411758423,
|
| 1333 |
+
"learning_rate": 1.986310002817608e-05,
|
| 1334 |
+
"loss": 0.0258,
|
| 1335 |
+
"step": 174
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 0.1483050847457627,
|
| 1339 |
+
"grad_norm": 1.0819811820983887,
|
| 1340 |
+
"learning_rate": 1.9858178781692777e-05,
|
| 1341 |
+
"loss": 0.021,
|
| 1342 |
+
"step": 175
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 0.14915254237288136,
|
| 1346 |
+
"grad_norm": 1.0780267715454102,
|
| 1347 |
+
"learning_rate": 1.985317126773477e-05,
|
| 1348 |
+
"loss": 0.0194,
|
| 1349 |
+
"step": 176
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 0.15,
|
| 1353 |
+
"grad_norm": 2.7838480472564697,
|
| 1354 |
+
"learning_rate": 1.9848077530122083e-05,
|
| 1355 |
+
"loss": 0.0543,
|
| 1356 |
+
"step": 177
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 0.15084745762711865,
|
| 1360 |
+
"grad_norm": 1.6005308628082275,
|
| 1361 |
+
"learning_rate": 1.984289761342926e-05,
|
| 1362 |
+
"loss": 0.0216,
|
| 1363 |
+
"step": 178
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 0.15169491525423728,
|
| 1367 |
+
"grad_norm": 1.7227445840835571,
|
| 1368 |
+
"learning_rate": 1.9837631562984995e-05,
|
| 1369 |
+
"loss": 0.0232,
|
| 1370 |
+
"step": 179
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.15254237288135594,
|
| 1374 |
+
"grad_norm": 2.9609763622283936,
|
| 1375 |
+
"learning_rate": 1.983227942487172e-05,
|
| 1376 |
+
"loss": 0.0357,
|
| 1377 |
+
"step": 180
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 0.15254237288135594,
|
| 1381 |
+
"eval_accuracy": 0.9971631205673759,
|
| 1382 |
+
"eval_f1": 0.9948453608247423,
|
| 1383 |
+
"eval_loss": 0.007280215620994568,
|
| 1384 |
+
"eval_precision": 0.9948453608247423,
|
| 1385 |
+
"eval_recall": 0.9948453608247423,
|
| 1386 |
+
"eval_runtime": 50.033,
|
| 1387 |
+
"eval_samples_per_second": 5.836,
|
| 1388 |
+
"eval_steps_per_second": 0.2,
|
| 1389 |
+
"step": 180
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.15338983050847457,
|
| 1393 |
+
"grad_norm": 1.3586597442626953,
|
| 1394 |
+
"learning_rate": 1.982684124592521e-05,
|
| 1395 |
+
"loss": 0.0248,
|
| 1396 |
+
"step": 181
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.15423728813559323,
|
| 1400 |
+
"grad_norm": 2.4621241092681885,
|
| 1401 |
+
"learning_rate": 1.9821317073734173e-05,
|
| 1402 |
+
"loss": 0.0244,
|
| 1403 |
+
"step": 182
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.15508474576271186,
|
| 1407 |
+
"grad_norm": 1.4555177688598633,
|
| 1408 |
+
"learning_rate": 1.9815706956639824e-05,
|
| 1409 |
+
"loss": 0.0252,
|
| 1410 |
+
"step": 183
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.15593220338983052,
|
| 1414 |
+
"grad_norm": 3.31247878074646,
|
| 1415 |
+
"learning_rate": 1.981001094373548e-05,
|
| 1416 |
+
"loss": 0.029,
|
| 1417 |
+
"step": 184
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.15677966101694915,
|
| 1421 |
+
"grad_norm": 1.3305749893188477,
|
| 1422 |
+
"learning_rate": 1.9804229084866103e-05,
|
| 1423 |
+
"loss": 0.0071,
|
| 1424 |
+
"step": 185
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.1576271186440678,
|
| 1428 |
+
"grad_norm": 2.1134912967681885,
|
| 1429 |
+
"learning_rate": 1.9798361430627898e-05,
|
| 1430 |
+
"loss": 0.0223,
|
| 1431 |
+
"step": 186
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.15847457627118644,
|
| 1435 |
+
"grad_norm": 2.255300521850586,
|
| 1436 |
+
"learning_rate": 1.979240803236785e-05,
|
| 1437 |
+
"loss": 0.0165,
|
| 1438 |
+
"step": 187
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.15932203389830507,
|
| 1442 |
+
"grad_norm": 1.72796630859375,
|
| 1443 |
+
"learning_rate": 1.9786368942183262e-05,
|
| 1444 |
+
"loss": 0.0216,
|
| 1445 |
+
"step": 188
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.16016949152542373,
|
| 1449 |
+
"grad_norm": 3.4896645545959473,
|
| 1450 |
+
"learning_rate": 1.9780244212921333e-05,
|
| 1451 |
+
"loss": 0.0188,
|
| 1452 |
+
"step": 189
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.16101694915254236,
|
| 1456 |
+
"grad_norm": 8.25186538696289,
|
| 1457 |
+
"learning_rate": 1.9774033898178668e-05,
|
| 1458 |
+
"loss": 0.0585,
|
| 1459 |
+
"step": 190
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.16186440677966102,
|
| 1463 |
+
"grad_norm": 2.215669870376587,
|
| 1464 |
+
"learning_rate": 1.9767738052300816e-05,
|
| 1465 |
+
"loss": 0.015,
|
| 1466 |
+
"step": 191
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.16271186440677965,
|
| 1470 |
+
"grad_norm": 1.952848196029663,
|
| 1471 |
+
"learning_rate": 1.9761356730381806e-05,
|
| 1472 |
+
"loss": 0.023,
|
| 1473 |
+
"step": 192
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.1635593220338983,
|
| 1477 |
+
"grad_norm": 2.2259209156036377,
|
| 1478 |
+
"learning_rate": 1.975488998826364e-05,
|
| 1479 |
+
"loss": 0.0186,
|
| 1480 |
+
"step": 193
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.16440677966101694,
|
| 1484 |
+
"grad_norm": 2.9435532093048096,
|
| 1485 |
+
"learning_rate": 1.974833788253584e-05,
|
| 1486 |
+
"loss": 0.0237,
|
| 1487 |
+
"step": 194
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.1652542372881356,
|
| 1491 |
+
"grad_norm": 3.2941129207611084,
|
| 1492 |
+
"learning_rate": 1.9741700470534904e-05,
|
| 1493 |
+
"loss": 0.0168,
|
| 1494 |
+
"step": 195
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.16610169491525423,
|
| 1498 |
+
"grad_norm": 2.563314437866211,
|
| 1499 |
+
"learning_rate": 1.9734977810343868e-05,
|
| 1500 |
+
"loss": 0.0245,
|
| 1501 |
+
"step": 196
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.1669491525423729,
|
| 1505 |
+
"grad_norm": 5.807168006896973,
|
| 1506 |
+
"learning_rate": 1.9728169960791736e-05,
|
| 1507 |
+
"loss": 0.0526,
|
| 1508 |
+
"step": 197
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.16779661016949152,
|
| 1512 |
+
"grad_norm": 2.7468652725219727,
|
| 1513 |
+
"learning_rate": 1.9721276981452995e-05,
|
| 1514 |
+
"loss": 0.03,
|
| 1515 |
+
"step": 198
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.16864406779661018,
|
| 1519 |
+
"grad_norm": 7.025511741638184,
|
| 1520 |
+
"learning_rate": 1.97142989326471e-05,
|
| 1521 |
+
"loss": 0.0794,
|
| 1522 |
+
"step": 199
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.1694915254237288,
|
| 1526 |
+
"grad_norm": 1.384521722793579,
|
| 1527 |
+
"learning_rate": 1.9707235875437932e-05,
|
| 1528 |
+
"loss": 0.0147,
|
| 1529 |
+
"step": 200
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.1694915254237288,
|
| 1533 |
+
"eval_accuracy": 0.9985815602836879,
|
| 1534 |
+
"eval_f1": 0.9974160206718347,
|
| 1535 |
+
"eval_loss": 0.010543613694608212,
|
| 1536 |
+
"eval_precision": 1.0,
|
| 1537 |
+
"eval_recall": 0.9948453608247423,
|
| 1538 |
+
"eval_runtime": 50.7835,
|
| 1539 |
+
"eval_samples_per_second": 5.75,
|
| 1540 |
+
"eval_steps_per_second": 0.197,
|
| 1541 |
+
"step": 200
|
| 1542 |
}
|
| 1543 |
],
|
| 1544 |
"logging_steps": 1,
|
|
|
|
| 1558 |
"attributes": {}
|
| 1559 |
}
|
| 1560 |
},
|
| 1561 |
+
"total_flos": 6.207084150810214e+16,
|
| 1562 |
"train_batch_size": 8,
|
| 1563 |
"trial_name": null,
|
| 1564 |
"trial_params": null
|