Training in progress, step 900, 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 sha256:8dcfd94f62ff19b8a9933a9541a5733207408d391abbc19039ab84f7f6b53b2b
|
| 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 sha256:4dbf500184a745e32656bd72aed96444d450b8822d32f095305fab829cad5dcd
|
| 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 sha256:19d8b41074874752835461e4b0b43cfae611a1b52b65b06efded073a8540f0a3
|
| 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 sha256:420c4af07d8d11503cf38a19956fa08bce83ddcb83c306ed5f4f325d2289c54f
|
| 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 sha256:5b5d70ab7c84844a31deff224e2783035d5a27b4e451d959e41e69e6ba1bfba0
|
| 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:acfc04e204792faef03bd393bc190684855a1199069618e09617dabdc06f61f4
|
| 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 sha256:6560ad83e9bf537962a4628c52535beab1489d42b33161c72bbe9a7052130abb
|
| 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 sha256:9290d11cadcf051134f7e3a6303ac31bd31867fa4eca0a0ba3fad88d33d73e1e
|
| 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 sha256:0f014cc69554457011024b3f78c613f5d6d21afe59987b579c8148d7a14d4266
|
| 3 |
size 14960
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:211fe4cc875fc5d117275383fb9805fb9073808b1bf8c46687ccdf31963af654
|
| 3 |
size 14960
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c181295f448353cc0ad867ce15d877dc5e4902c02d374aebe08a6faa6eaf73cb
|
| 3 |
size 14960
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ff94a6828461c3753dda9d571a0a287a6967bc241a3276d73771206e60da3988
|
| 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:da38d22b5a3dd941093c6e3a0d9705292603a8a72ab0757d88c7c3b61a44e0be
|
| 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,
|
|
@@ -6099,6 +6099,766 @@
|
|
| 6099 |
"eval_samples_per_second": 5.622,
|
| 6100 |
"eval_steps_per_second": 0.189,
|
| 6101 |
"step": 800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6102 |
}
|
| 6103 |
],
|
| 6104 |
"logging_steps": 1,
|
|
@@ -6118,7 +6878,7 @@
|
|
| 6118 |
"attributes": {}
|
| 6119 |
}
|
| 6120 |
},
|
| 6121 |
-
"total_flos": 2.
|
| 6122 |
"train_batch_size": 8,
|
| 6123 |
"trial_name": null,
|
| 6124 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.6089309878213802,
|
| 5 |
"eval_steps": 20,
|
| 6 |
+
"global_step": 900,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 6099 |
"eval_samples_per_second": 5.622,
|
| 6100 |
"eval_steps_per_second": 0.189,
|
| 6101 |
"step": 800
|
| 6102 |
+
},
|
| 6103 |
+
{
|
| 6104 |
+
"epoch": 0.5419485791610285,
|
| 6105 |
+
"grad_norm": 4.048157691955566,
|
| 6106 |
+
"learning_rate": 1.0283414013680233e-05,
|
| 6107 |
+
"loss": 0.1629,
|
| 6108 |
+
"step": 801
|
| 6109 |
+
},
|
| 6110 |
+
{
|
| 6111 |
+
"epoch": 0.5426251691474966,
|
| 6112 |
+
"grad_norm": 8.180608749389648,
|
| 6113 |
+
"learning_rate": 1.0259801736097634e-05,
|
| 6114 |
+
"loss": 0.2779,
|
| 6115 |
+
"step": 802
|
| 6116 |
+
},
|
| 6117 |
+
{
|
| 6118 |
+
"epoch": 0.5433017591339648,
|
| 6119 |
+
"grad_norm": 2.2637126445770264,
|
| 6120 |
+
"learning_rate": 1.023618800894798e-05,
|
| 6121 |
+
"loss": 0.0888,
|
| 6122 |
+
"step": 803
|
| 6123 |
+
},
|
| 6124 |
+
{
|
| 6125 |
+
"epoch": 0.543978349120433,
|
| 6126 |
+
"grad_norm": 3.4222660064697266,
|
| 6127 |
+
"learning_rate": 1.0212572963984358e-05,
|
| 6128 |
+
"loss": 0.1072,
|
| 6129 |
+
"step": 804
|
| 6130 |
+
},
|
| 6131 |
+
{
|
| 6132 |
+
"epoch": 0.5446549391069012,
|
| 6133 |
+
"grad_norm": 5.149094104766846,
|
| 6134 |
+
"learning_rate": 1.0188956732967208e-05,
|
| 6135 |
+
"loss": 0.1834,
|
| 6136 |
+
"step": 805
|
| 6137 |
+
},
|
| 6138 |
+
{
|
| 6139 |
+
"epoch": 0.5453315290933695,
|
| 6140 |
+
"grad_norm": 4.928592681884766,
|
| 6141 |
+
"learning_rate": 1.0165339447663586e-05,
|
| 6142 |
+
"loss": 0.1065,
|
| 6143 |
+
"step": 806
|
| 6144 |
+
},
|
| 6145 |
+
{
|
| 6146 |
+
"epoch": 0.5460081190798376,
|
| 6147 |
+
"grad_norm": 3.3745458126068115,
|
| 6148 |
+
"learning_rate": 1.0141721239846436e-05,
|
| 6149 |
+
"loss": 0.1105,
|
| 6150 |
+
"step": 807
|
| 6151 |
+
},
|
| 6152 |
+
{
|
| 6153 |
+
"epoch": 0.5466847090663058,
|
| 6154 |
+
"grad_norm": 3.140056848526001,
|
| 6155 |
+
"learning_rate": 1.0118102241293848e-05,
|
| 6156 |
+
"loss": 0.1057,
|
| 6157 |
+
"step": 808
|
| 6158 |
+
},
|
| 6159 |
+
{
|
| 6160 |
+
"epoch": 0.547361299052774,
|
| 6161 |
+
"grad_norm": 2.9225945472717285,
|
| 6162 |
+
"learning_rate": 1.0094482583788311e-05,
|
| 6163 |
+
"loss": 0.1409,
|
| 6164 |
+
"step": 809
|
| 6165 |
+
},
|
| 6166 |
+
{
|
| 6167 |
+
"epoch": 0.5480378890392422,
|
| 6168 |
+
"grad_norm": 5.8072333335876465,
|
| 6169 |
+
"learning_rate": 1.0070862399116016e-05,
|
| 6170 |
+
"loss": 0.1697,
|
| 6171 |
+
"step": 810
|
| 6172 |
+
},
|
| 6173 |
+
{
|
| 6174 |
+
"epoch": 0.5487144790257105,
|
| 6175 |
+
"grad_norm": 2.9896950721740723,
|
| 6176 |
+
"learning_rate": 1.0047241819066069e-05,
|
| 6177 |
+
"loss": 0.0893,
|
| 6178 |
+
"step": 811
|
| 6179 |
+
},
|
| 6180 |
+
{
|
| 6181 |
+
"epoch": 0.5493910690121786,
|
| 6182 |
+
"grad_norm": 9.982451438903809,
|
| 6183 |
+
"learning_rate": 1.0023620975429803e-05,
|
| 6184 |
+
"loss": 0.2119,
|
| 6185 |
+
"step": 812
|
| 6186 |
+
},
|
| 6187 |
+
{
|
| 6188 |
+
"epoch": 0.5500676589986468,
|
| 6189 |
+
"grad_norm": 5.057304382324219,
|
| 6190 |
+
"learning_rate": 1e-05,
|
| 6191 |
+
"loss": 0.1417,
|
| 6192 |
+
"step": 813
|
| 6193 |
+
},
|
| 6194 |
+
{
|
| 6195 |
+
"epoch": 0.550744248985115,
|
| 6196 |
+
"grad_norm": 8.598350524902344,
|
| 6197 |
+
"learning_rate": 9.976379024570202e-06,
|
| 6198 |
+
"loss": 0.2122,
|
| 6199 |
+
"step": 814
|
| 6200 |
+
},
|
| 6201 |
+
{
|
| 6202 |
+
"epoch": 0.5514208389715832,
|
| 6203 |
+
"grad_norm": 2.8650100231170654,
|
| 6204 |
+
"learning_rate": 9.952758180933933e-06,
|
| 6205 |
+
"loss": 0.0901,
|
| 6206 |
+
"step": 815
|
| 6207 |
+
},
|
| 6208 |
+
{
|
| 6209 |
+
"epoch": 0.5520974289580515,
|
| 6210 |
+
"grad_norm": 5.409826278686523,
|
| 6211 |
+
"learning_rate": 9.929137600883986e-06,
|
| 6212 |
+
"loss": 0.1613,
|
| 6213 |
+
"step": 816
|
| 6214 |
+
},
|
| 6215 |
+
{
|
| 6216 |
+
"epoch": 0.5527740189445196,
|
| 6217 |
+
"grad_norm": 2.837448835372925,
|
| 6218 |
+
"learning_rate": 9.90551741621169e-06,
|
| 6219 |
+
"loss": 0.0946,
|
| 6220 |
+
"step": 817
|
| 6221 |
+
},
|
| 6222 |
+
{
|
| 6223 |
+
"epoch": 0.5534506089309879,
|
| 6224 |
+
"grad_norm": 3.9413063526153564,
|
| 6225 |
+
"learning_rate": 9.881897758706155e-06,
|
| 6226 |
+
"loss": 0.1099,
|
| 6227 |
+
"step": 818
|
| 6228 |
+
},
|
| 6229 |
+
{
|
| 6230 |
+
"epoch": 0.554127198917456,
|
| 6231 |
+
"grad_norm": 2.284583330154419,
|
| 6232 |
+
"learning_rate": 9.858278760153567e-06,
|
| 6233 |
+
"loss": 0.0653,
|
| 6234 |
+
"step": 819
|
| 6235 |
+
},
|
| 6236 |
+
{
|
| 6237 |
+
"epoch": 0.5548037889039242,
|
| 6238 |
+
"grad_norm": 6.636195182800293,
|
| 6239 |
+
"learning_rate": 9.834660552336415e-06,
|
| 6240 |
+
"loss": 0.1583,
|
| 6241 |
+
"step": 820
|
| 6242 |
+
},
|
| 6243 |
+
{
|
| 6244 |
+
"epoch": 0.5548037889039242,
|
| 6245 |
+
"eval_accuracy": 0.7788990825688074,
|
| 6246 |
+
"eval_f1": 0.3674540682414698,
|
| 6247 |
+
"eval_loss": 0.47594916820526123,
|
| 6248 |
+
"eval_precision": 0.8433734939759037,
|
| 6249 |
+
"eval_recall": 0.2348993288590604,
|
| 6250 |
+
"eval_runtime": 53.6382,
|
| 6251 |
+
"eval_samples_per_second": 5.556,
|
| 6252 |
+
"eval_steps_per_second": 0.186,
|
| 6253 |
+
"step": 820
|
| 6254 |
+
},
|
| 6255 |
+
{
|
| 6256 |
+
"epoch": 0.5554803788903924,
|
| 6257 |
+
"grad_norm": 3.6005120277404785,
|
| 6258 |
+
"learning_rate": 9.811043267032797e-06,
|
| 6259 |
+
"loss": 0.0887,
|
| 6260 |
+
"step": 821
|
| 6261 |
+
},
|
| 6262 |
+
{
|
| 6263 |
+
"epoch": 0.5561569688768606,
|
| 6264 |
+
"grad_norm": 3.6898558139801025,
|
| 6265 |
+
"learning_rate": 9.787427036015647e-06,
|
| 6266 |
+
"loss": 0.147,
|
| 6267 |
+
"step": 822
|
| 6268 |
+
},
|
| 6269 |
+
{
|
| 6270 |
+
"epoch": 0.5568335588633289,
|
| 6271 |
+
"grad_norm": 6.481770992279053,
|
| 6272 |
+
"learning_rate": 9.763811991052021e-06,
|
| 6273 |
+
"loss": 0.1877,
|
| 6274 |
+
"step": 823
|
| 6275 |
+
},
|
| 6276 |
+
{
|
| 6277 |
+
"epoch": 0.557510148849797,
|
| 6278 |
+
"grad_norm": 2.4457807540893555,
|
| 6279 |
+
"learning_rate": 9.74019826390237e-06,
|
| 6280 |
+
"loss": 0.1004,
|
| 6281 |
+
"step": 824
|
| 6282 |
+
},
|
| 6283 |
+
{
|
| 6284 |
+
"epoch": 0.5581867388362652,
|
| 6285 |
+
"grad_norm": 2.3346476554870605,
|
| 6286 |
+
"learning_rate": 9.716585986319769e-06,
|
| 6287 |
+
"loss": 0.1199,
|
| 6288 |
+
"step": 825
|
| 6289 |
+
},
|
| 6290 |
+
{
|
| 6291 |
+
"epoch": 0.5588633288227334,
|
| 6292 |
+
"grad_norm": 5.753274917602539,
|
| 6293 |
+
"learning_rate": 9.692975290049228e-06,
|
| 6294 |
+
"loss": 0.1646,
|
| 6295 |
+
"step": 826
|
| 6296 |
+
},
|
| 6297 |
+
{
|
| 6298 |
+
"epoch": 0.5595399188092016,
|
| 6299 |
+
"grad_norm": 4.540411949157715,
|
| 6300 |
+
"learning_rate": 9.669366306826919e-06,
|
| 6301 |
+
"loss": 0.1275,
|
| 6302 |
+
"step": 827
|
| 6303 |
+
},
|
| 6304 |
+
{
|
| 6305 |
+
"epoch": 0.5602165087956699,
|
| 6306 |
+
"grad_norm": 4.377498149871826,
|
| 6307 |
+
"learning_rate": 9.645759168379463e-06,
|
| 6308 |
+
"loss": 0.1508,
|
| 6309 |
+
"step": 828
|
| 6310 |
+
},
|
| 6311 |
+
{
|
| 6312 |
+
"epoch": 0.560893098782138,
|
| 6313 |
+
"grad_norm": 4.894872188568115,
|
| 6314 |
+
"learning_rate": 9.622154006423185e-06,
|
| 6315 |
+
"loss": 0.1608,
|
| 6316 |
+
"step": 829
|
| 6317 |
+
},
|
| 6318 |
+
{
|
| 6319 |
+
"epoch": 0.5615696887686062,
|
| 6320 |
+
"grad_norm": 3.9579596519470215,
|
| 6321 |
+
"learning_rate": 9.598550952663383e-06,
|
| 6322 |
+
"loss": 0.0748,
|
| 6323 |
+
"step": 830
|
| 6324 |
+
},
|
| 6325 |
+
{
|
| 6326 |
+
"epoch": 0.5622462787550744,
|
| 6327 |
+
"grad_norm": 3.1920807361602783,
|
| 6328 |
+
"learning_rate": 9.574950138793593e-06,
|
| 6329 |
+
"loss": 0.0958,
|
| 6330 |
+
"step": 831
|
| 6331 |
+
},
|
| 6332 |
+
{
|
| 6333 |
+
"epoch": 0.5629228687415426,
|
| 6334 |
+
"grad_norm": 2.8056745529174805,
|
| 6335 |
+
"learning_rate": 9.551351696494854e-06,
|
| 6336 |
+
"loss": 0.1528,
|
| 6337 |
+
"step": 832
|
| 6338 |
+
},
|
| 6339 |
+
{
|
| 6340 |
+
"epoch": 0.5635994587280109,
|
| 6341 |
+
"grad_norm": 2.493863105773926,
|
| 6342 |
+
"learning_rate": 9.527755757434968e-06,
|
| 6343 |
+
"loss": 0.0693,
|
| 6344 |
+
"step": 833
|
| 6345 |
+
},
|
| 6346 |
+
{
|
| 6347 |
+
"epoch": 0.564276048714479,
|
| 6348 |
+
"grad_norm": 3.3314931392669678,
|
| 6349 |
+
"learning_rate": 9.504162453267776e-06,
|
| 6350 |
+
"loss": 0.1227,
|
| 6351 |
+
"step": 834
|
| 6352 |
+
},
|
| 6353 |
+
{
|
| 6354 |
+
"epoch": 0.5649526387009473,
|
| 6355 |
+
"grad_norm": 3.9033989906311035,
|
| 6356 |
+
"learning_rate": 9.480571915632422e-06,
|
| 6357 |
+
"loss": 0.1199,
|
| 6358 |
+
"step": 835
|
| 6359 |
+
},
|
| 6360 |
+
{
|
| 6361 |
+
"epoch": 0.5656292286874154,
|
| 6362 |
+
"grad_norm": 3.6395678520202637,
|
| 6363 |
+
"learning_rate": 9.456984276152598e-06,
|
| 6364 |
+
"loss": 0.1057,
|
| 6365 |
+
"step": 836
|
| 6366 |
+
},
|
| 6367 |
+
{
|
| 6368 |
+
"epoch": 0.5663058186738836,
|
| 6369 |
+
"grad_norm": 6.916732311248779,
|
| 6370 |
+
"learning_rate": 9.43339966643584e-06,
|
| 6371 |
+
"loss": 0.1741,
|
| 6372 |
+
"step": 837
|
| 6373 |
+
},
|
| 6374 |
+
{
|
| 6375 |
+
"epoch": 0.5669824086603519,
|
| 6376 |
+
"grad_norm": 3.8561432361602783,
|
| 6377 |
+
"learning_rate": 9.409818218072774e-06,
|
| 6378 |
+
"loss": 0.1654,
|
| 6379 |
+
"step": 838
|
| 6380 |
+
},
|
| 6381 |
+
{
|
| 6382 |
+
"epoch": 0.56765899864682,
|
| 6383 |
+
"grad_norm": 3.962113380432129,
|
| 6384 |
+
"learning_rate": 9.386240062636388e-06,
|
| 6385 |
+
"loss": 0.1459,
|
| 6386 |
+
"step": 839
|
| 6387 |
+
},
|
| 6388 |
+
{
|
| 6389 |
+
"epoch": 0.5683355886332883,
|
| 6390 |
+
"grad_norm": 2.5661449432373047,
|
| 6391 |
+
"learning_rate": 9.362665331681294e-06,
|
| 6392 |
+
"loss": 0.1363,
|
| 6393 |
+
"step": 840
|
| 6394 |
+
},
|
| 6395 |
+
{
|
| 6396 |
+
"epoch": 0.5683355886332883,
|
| 6397 |
+
"eval_accuracy": 0.8009174311926606,
|
| 6398 |
+
"eval_f1": 0.4668304668304668,
|
| 6399 |
+
"eval_loss": 0.43011632561683655,
|
| 6400 |
+
"eval_precision": 0.8715596330275229,
|
| 6401 |
+
"eval_recall": 0.3187919463087248,
|
| 6402 |
+
"eval_runtime": 54.0976,
|
| 6403 |
+
"eval_samples_per_second": 5.509,
|
| 6404 |
+
"eval_steps_per_second": 0.185,
|
| 6405 |
+
"step": 840
|
| 6406 |
+
},
|
| 6407 |
+
{
|
| 6408 |
+
"epoch": 0.5690121786197564,
|
| 6409 |
+
"grad_norm": 3.792685031890869,
|
| 6410 |
+
"learning_rate": 9.339094156743007e-06,
|
| 6411 |
+
"loss": 0.1408,
|
| 6412 |
+
"step": 841
|
| 6413 |
+
},
|
| 6414 |
+
{
|
| 6415 |
+
"epoch": 0.5696887686062246,
|
| 6416 |
+
"grad_norm": 4.538841247558594,
|
| 6417 |
+
"learning_rate": 9.315526669337189e-06,
|
| 6418 |
+
"loss": 0.1399,
|
| 6419 |
+
"step": 842
|
| 6420 |
+
},
|
| 6421 |
+
{
|
| 6422 |
+
"epoch": 0.5703653585926928,
|
| 6423 |
+
"grad_norm": 5.182969570159912,
|
| 6424 |
+
"learning_rate": 9.291963000958932e-06,
|
| 6425 |
+
"loss": 0.1753,
|
| 6426 |
+
"step": 843
|
| 6427 |
+
},
|
| 6428 |
+
{
|
| 6429 |
+
"epoch": 0.571041948579161,
|
| 6430 |
+
"grad_norm": 3.905219554901123,
|
| 6431 |
+
"learning_rate": 9.268403283082025e-06,
|
| 6432 |
+
"loss": 0.143,
|
| 6433 |
+
"step": 844
|
| 6434 |
+
},
|
| 6435 |
+
{
|
| 6436 |
+
"epoch": 0.5717185385656293,
|
| 6437 |
+
"grad_norm": 3.7634634971618652,
|
| 6438 |
+
"learning_rate": 9.244847647158203e-06,
|
| 6439 |
+
"loss": 0.1469,
|
| 6440 |
+
"step": 845
|
| 6441 |
+
},
|
| 6442 |
+
{
|
| 6443 |
+
"epoch": 0.5723951285520974,
|
| 6444 |
+
"grad_norm": 3.5530450344085693,
|
| 6445 |
+
"learning_rate": 9.221296224616443e-06,
|
| 6446 |
+
"loss": 0.1334,
|
| 6447 |
+
"step": 846
|
| 6448 |
+
},
|
| 6449 |
+
{
|
| 6450 |
+
"epoch": 0.5730717185385656,
|
| 6451 |
+
"grad_norm": 6.1246161460876465,
|
| 6452 |
+
"learning_rate": 9.197749146862193e-06,
|
| 6453 |
+
"loss": 0.1216,
|
| 6454 |
+
"step": 847
|
| 6455 |
+
},
|
| 6456 |
+
{
|
| 6457 |
+
"epoch": 0.5737483085250338,
|
| 6458 |
+
"grad_norm": 7.231658458709717,
|
| 6459 |
+
"learning_rate": 9.174206545276678e-06,
|
| 6460 |
+
"loss": 0.2128,
|
| 6461 |
+
"step": 848
|
| 6462 |
+
},
|
| 6463 |
+
{
|
| 6464 |
+
"epoch": 0.574424898511502,
|
| 6465 |
+
"grad_norm": 6.129051685333252,
|
| 6466 |
+
"learning_rate": 9.150668551216134e-06,
|
| 6467 |
+
"loss": 0.2178,
|
| 6468 |
+
"step": 849
|
| 6469 |
+
},
|
| 6470 |
+
{
|
| 6471 |
+
"epoch": 0.5751014884979703,
|
| 6472 |
+
"grad_norm": 4.892454624176025,
|
| 6473 |
+
"learning_rate": 9.127135296011102e-06,
|
| 6474 |
+
"loss": 0.1496,
|
| 6475 |
+
"step": 850
|
| 6476 |
+
},
|
| 6477 |
+
{
|
| 6478 |
+
"epoch": 0.5757780784844384,
|
| 6479 |
+
"grad_norm": 6.8777337074279785,
|
| 6480 |
+
"learning_rate": 9.103606910965666e-06,
|
| 6481 |
+
"loss": 0.2008,
|
| 6482 |
+
"step": 851
|
| 6483 |
+
},
|
| 6484 |
+
{
|
| 6485 |
+
"epoch": 0.5764546684709067,
|
| 6486 |
+
"grad_norm": 3.538118600845337,
|
| 6487 |
+
"learning_rate": 9.080083527356755e-06,
|
| 6488 |
+
"loss": 0.1232,
|
| 6489 |
+
"step": 852
|
| 6490 |
+
},
|
| 6491 |
+
{
|
| 6492 |
+
"epoch": 0.5771312584573748,
|
| 6493 |
+
"grad_norm": 5.2440080642700195,
|
| 6494 |
+
"learning_rate": 9.056565276433378e-06,
|
| 6495 |
+
"loss": 0.1973,
|
| 6496 |
+
"step": 853
|
| 6497 |
+
},
|
| 6498 |
+
{
|
| 6499 |
+
"epoch": 0.577807848443843,
|
| 6500 |
+
"grad_norm": 2.282479763031006,
|
| 6501 |
+
"learning_rate": 9.033052289415914e-06,
|
| 6502 |
+
"loss": 0.0696,
|
| 6503 |
+
"step": 854
|
| 6504 |
+
},
|
| 6505 |
+
{
|
| 6506 |
+
"epoch": 0.5784844384303113,
|
| 6507 |
+
"grad_norm": 3.643191337585449,
|
| 6508 |
+
"learning_rate": 9.009544697495373e-06,
|
| 6509 |
+
"loss": 0.1378,
|
| 6510 |
+
"step": 855
|
| 6511 |
+
},
|
| 6512 |
+
{
|
| 6513 |
+
"epoch": 0.5791610284167794,
|
| 6514 |
+
"grad_norm": 3.0240986347198486,
|
| 6515 |
+
"learning_rate": 8.986042631832656e-06,
|
| 6516 |
+
"loss": 0.1579,
|
| 6517 |
+
"step": 856
|
| 6518 |
+
},
|
| 6519 |
+
{
|
| 6520 |
+
"epoch": 0.5798376184032477,
|
| 6521 |
+
"grad_norm": 3.921018362045288,
|
| 6522 |
+
"learning_rate": 8.962546223557838e-06,
|
| 6523 |
+
"loss": 0.1194,
|
| 6524 |
+
"step": 857
|
| 6525 |
+
},
|
| 6526 |
+
{
|
| 6527 |
+
"epoch": 0.5805142083897158,
|
| 6528 |
+
"grad_norm": 3.4192543029785156,
|
| 6529 |
+
"learning_rate": 8.93905560376942e-06,
|
| 6530 |
+
"loss": 0.1817,
|
| 6531 |
+
"step": 858
|
| 6532 |
+
},
|
| 6533 |
+
{
|
| 6534 |
+
"epoch": 0.581190798376184,
|
| 6535 |
+
"grad_norm": 4.1514363288879395,
|
| 6536 |
+
"learning_rate": 8.915570903533615e-06,
|
| 6537 |
+
"loss": 0.1489,
|
| 6538 |
+
"step": 859
|
| 6539 |
+
},
|
| 6540 |
+
{
|
| 6541 |
+
"epoch": 0.5818673883626523,
|
| 6542 |
+
"grad_norm": 3.885377883911133,
|
| 6543 |
+
"learning_rate": 8.892092253883602e-06,
|
| 6544 |
+
"loss": 0.1456,
|
| 6545 |
+
"step": 860
|
| 6546 |
+
},
|
| 6547 |
+
{
|
| 6548 |
+
"epoch": 0.5818673883626523,
|
| 6549 |
+
"eval_accuracy": 0.7926605504587156,
|
| 6550 |
+
"eval_f1": 0.42346938775510207,
|
| 6551 |
+
"eval_loss": 0.4394099712371826,
|
| 6552 |
+
"eval_precision": 0.8829787234042553,
|
| 6553 |
+
"eval_recall": 0.2785234899328859,
|
| 6554 |
+
"eval_runtime": 53.3745,
|
| 6555 |
+
"eval_samples_per_second": 5.583,
|
| 6556 |
+
"eval_steps_per_second": 0.187,
|
| 6557 |
+
"step": 860
|
| 6558 |
+
},
|
| 6559 |
+
{
|
| 6560 |
+
"epoch": 0.5825439783491204,
|
| 6561 |
+
"grad_norm": 3.6462066173553467,
|
| 6562 |
+
"learning_rate": 8.8686197858188e-06,
|
| 6563 |
+
"loss": 0.1782,
|
| 6564 |
+
"step": 861
|
| 6565 |
+
},
|
| 6566 |
+
{
|
| 6567 |
+
"epoch": 0.5832205683355887,
|
| 6568 |
+
"grad_norm": 3.2633800506591797,
|
| 6569 |
+
"learning_rate": 8.84515363030414e-06,
|
| 6570 |
+
"loss": 0.1656,
|
| 6571 |
+
"step": 862
|
| 6572 |
+
},
|
| 6573 |
+
{
|
| 6574 |
+
"epoch": 0.5838971583220568,
|
| 6575 |
+
"grad_norm": 5.255461692810059,
|
| 6576 |
+
"learning_rate": 8.821693918269334e-06,
|
| 6577 |
+
"loss": 0.1306,
|
| 6578 |
+
"step": 863
|
| 6579 |
+
},
|
| 6580 |
+
{
|
| 6581 |
+
"epoch": 0.584573748308525,
|
| 6582 |
+
"grad_norm": 4.525811672210693,
|
| 6583 |
+
"learning_rate": 8.798240780608143e-06,
|
| 6584 |
+
"loss": 0.1684,
|
| 6585 |
+
"step": 864
|
| 6586 |
+
},
|
| 6587 |
+
{
|
| 6588 |
+
"epoch": 0.5852503382949933,
|
| 6589 |
+
"grad_norm": 2.788585901260376,
|
| 6590 |
+
"learning_rate": 8.774794348177641e-06,
|
| 6591 |
+
"loss": 0.1456,
|
| 6592 |
+
"step": 865
|
| 6593 |
+
},
|
| 6594 |
+
{
|
| 6595 |
+
"epoch": 0.5859269282814614,
|
| 6596 |
+
"grad_norm": 3.1500301361083984,
|
| 6597 |
+
"learning_rate": 8.751354751797492e-06,
|
| 6598 |
+
"loss": 0.1347,
|
| 6599 |
+
"step": 866
|
| 6600 |
+
},
|
| 6601 |
+
{
|
| 6602 |
+
"epoch": 0.5866035182679297,
|
| 6603 |
+
"grad_norm": 3.487180471420288,
|
| 6604 |
+
"learning_rate": 8.727922122249221e-06,
|
| 6605 |
+
"loss": 0.1393,
|
| 6606 |
+
"step": 867
|
| 6607 |
+
},
|
| 6608 |
+
{
|
| 6609 |
+
"epoch": 0.5872801082543978,
|
| 6610 |
+
"grad_norm": 2.1133573055267334,
|
| 6611 |
+
"learning_rate": 8.704496590275479e-06,
|
| 6612 |
+
"loss": 0.0814,
|
| 6613 |
+
"step": 868
|
| 6614 |
+
},
|
| 6615 |
+
{
|
| 6616 |
+
"epoch": 0.587956698240866,
|
| 6617 |
+
"grad_norm": 3.227505922317505,
|
| 6618 |
+
"learning_rate": 8.68107828657931e-06,
|
| 6619 |
+
"loss": 0.1104,
|
| 6620 |
+
"step": 869
|
| 6621 |
+
},
|
| 6622 |
+
{
|
| 6623 |
+
"epoch": 0.5886332882273342,
|
| 6624 |
+
"grad_norm": 2.8195204734802246,
|
| 6625 |
+
"learning_rate": 8.657667341823449e-06,
|
| 6626 |
+
"loss": 0.1073,
|
| 6627 |
+
"step": 870
|
| 6628 |
+
},
|
| 6629 |
+
{
|
| 6630 |
+
"epoch": 0.5893098782138024,
|
| 6631 |
+
"grad_norm": 6.85077428817749,
|
| 6632 |
+
"learning_rate": 8.63426388662954e-06,
|
| 6633 |
+
"loss": 0.2117,
|
| 6634 |
+
"step": 871
|
| 6635 |
+
},
|
| 6636 |
+
{
|
| 6637 |
+
"epoch": 0.5899864682002707,
|
| 6638 |
+
"grad_norm": 3.027806043624878,
|
| 6639 |
+
"learning_rate": 8.61086805157747e-06,
|
| 6640 |
+
"loss": 0.13,
|
| 6641 |
+
"step": 872
|
| 6642 |
+
},
|
| 6643 |
+
{
|
| 6644 |
+
"epoch": 0.5906630581867388,
|
| 6645 |
+
"grad_norm": 3.608955144882202,
|
| 6646 |
+
"learning_rate": 8.587479967204584e-06,
|
| 6647 |
+
"loss": 0.1323,
|
| 6648 |
+
"step": 873
|
| 6649 |
+
},
|
| 6650 |
+
{
|
| 6651 |
+
"epoch": 0.591339648173207,
|
| 6652 |
+
"grad_norm": 3.6784167289733887,
|
| 6653 |
+
"learning_rate": 8.564099764004998e-06,
|
| 6654 |
+
"loss": 0.1205,
|
| 6655 |
+
"step": 874
|
| 6656 |
+
},
|
| 6657 |
+
{
|
| 6658 |
+
"epoch": 0.5920162381596752,
|
| 6659 |
+
"grad_norm": 3.6753430366516113,
|
| 6660 |
+
"learning_rate": 8.540727572428854e-06,
|
| 6661 |
+
"loss": 0.1728,
|
| 6662 |
+
"step": 875
|
| 6663 |
+
},
|
| 6664 |
+
{
|
| 6665 |
+
"epoch": 0.5926928281461434,
|
| 6666 |
+
"grad_norm": 3.4869165420532227,
|
| 6667 |
+
"learning_rate": 8.51736352288158e-06,
|
| 6668 |
+
"loss": 0.1363,
|
| 6669 |
+
"step": 876
|
| 6670 |
+
},
|
| 6671 |
+
{
|
| 6672 |
+
"epoch": 0.5933694181326117,
|
| 6673 |
+
"grad_norm": 6.327773571014404,
|
| 6674 |
+
"learning_rate": 8.494007745723197e-06,
|
| 6675 |
+
"loss": 0.1723,
|
| 6676 |
+
"step": 877
|
| 6677 |
+
},
|
| 6678 |
+
{
|
| 6679 |
+
"epoch": 0.5940460081190798,
|
| 6680 |
+
"grad_norm": 4.366674423217773,
|
| 6681 |
+
"learning_rate": 8.47066037126754e-06,
|
| 6682 |
+
"loss": 0.1557,
|
| 6683 |
+
"step": 878
|
| 6684 |
+
},
|
| 6685 |
+
{
|
| 6686 |
+
"epoch": 0.5947225981055481,
|
| 6687 |
+
"grad_norm": 3.28305721282959,
|
| 6688 |
+
"learning_rate": 8.447321529781597e-06,
|
| 6689 |
+
"loss": 0.1253,
|
| 6690 |
+
"step": 879
|
| 6691 |
+
},
|
| 6692 |
+
{
|
| 6693 |
+
"epoch": 0.5953991880920162,
|
| 6694 |
+
"grad_norm": 3.0015041828155518,
|
| 6695 |
+
"learning_rate": 8.423991351484715e-06,
|
| 6696 |
+
"loss": 0.1318,
|
| 6697 |
+
"step": 880
|
| 6698 |
+
},
|
| 6699 |
+
{
|
| 6700 |
+
"epoch": 0.5953991880920162,
|
| 6701 |
+
"eval_accuracy": 0.8192660550458716,
|
| 6702 |
+
"eval_f1": 0.5553047404063205,
|
| 6703 |
+
"eval_loss": 0.3900049328804016,
|
| 6704 |
+
"eval_precision": 0.8482758620689655,
|
| 6705 |
+
"eval_recall": 0.412751677852349,
|
| 6706 |
+
"eval_runtime": 53.5058,
|
| 6707 |
+
"eval_samples_per_second": 5.569,
|
| 6708 |
+
"eval_steps_per_second": 0.187,
|
| 6709 |
+
"step": 880
|
| 6710 |
+
},
|
| 6711 |
+
{
|
| 6712 |
+
"epoch": 0.5960757780784844,
|
| 6713 |
+
"grad_norm": 2.9270071983337402,
|
| 6714 |
+
"learning_rate": 8.400669966547925e-06,
|
| 6715 |
+
"loss": 0.1256,
|
| 6716 |
+
"step": 881
|
| 6717 |
+
},
|
| 6718 |
+
{
|
| 6719 |
+
"epoch": 0.5967523680649527,
|
| 6720 |
+
"grad_norm": 2.6574175357818604,
|
| 6721 |
+
"learning_rate": 8.377357505093183e-06,
|
| 6722 |
+
"loss": 0.0761,
|
| 6723 |
+
"step": 882
|
| 6724 |
+
},
|
| 6725 |
+
{
|
| 6726 |
+
"epoch": 0.5974289580514208,
|
| 6727 |
+
"grad_norm": 3.648263692855835,
|
| 6728 |
+
"learning_rate": 8.35405409719266e-06,
|
| 6729 |
+
"loss": 0.124,
|
| 6730 |
+
"step": 883
|
| 6731 |
+
},
|
| 6732 |
+
{
|
| 6733 |
+
"epoch": 0.5981055480378891,
|
| 6734 |
+
"grad_norm": 4.690035820007324,
|
| 6735 |
+
"learning_rate": 8.330759872868022e-06,
|
| 6736 |
+
"loss": 0.182,
|
| 6737 |
+
"step": 884
|
| 6738 |
+
},
|
| 6739 |
+
{
|
| 6740 |
+
"epoch": 0.5987821380243572,
|
| 6741 |
+
"grad_norm": 3.0360960960388184,
|
| 6742 |
+
"learning_rate": 8.307474962089676e-06,
|
| 6743 |
+
"loss": 0.1437,
|
| 6744 |
+
"step": 885
|
| 6745 |
+
},
|
| 6746 |
+
{
|
| 6747 |
+
"epoch": 0.5994587280108254,
|
| 6748 |
+
"grad_norm": 3.8773977756500244,
|
| 6749 |
+
"learning_rate": 8.284199494776083e-06,
|
| 6750 |
+
"loss": 0.0975,
|
| 6751 |
+
"step": 886
|
| 6752 |
+
},
|
| 6753 |
+
{
|
| 6754 |
+
"epoch": 0.6001353179972937,
|
| 6755 |
+
"grad_norm": 3.7407238483428955,
|
| 6756 |
+
"learning_rate": 8.260933600793003e-06,
|
| 6757 |
+
"loss": 0.1422,
|
| 6758 |
+
"step": 887
|
| 6759 |
+
},
|
| 6760 |
+
{
|
| 6761 |
+
"epoch": 0.6008119079837618,
|
| 6762 |
+
"grad_norm": 4.789558410644531,
|
| 6763 |
+
"learning_rate": 8.237677409952784e-06,
|
| 6764 |
+
"loss": 0.1737,
|
| 6765 |
+
"step": 888
|
| 6766 |
+
},
|
| 6767 |
+
{
|
| 6768 |
+
"epoch": 0.6014884979702301,
|
| 6769 |
+
"grad_norm": 5.8444929122924805,
|
| 6770 |
+
"learning_rate": 8.214431052013636e-06,
|
| 6771 |
+
"loss": 0.1658,
|
| 6772 |
+
"step": 889
|
| 6773 |
+
},
|
| 6774 |
+
{
|
| 6775 |
+
"epoch": 0.6021650879566982,
|
| 6776 |
+
"grad_norm": 2.8387224674224854,
|
| 6777 |
+
"learning_rate": 8.191194656678905e-06,
|
| 6778 |
+
"loss": 0.1317,
|
| 6779 |
+
"step": 890
|
| 6780 |
+
},
|
| 6781 |
+
{
|
| 6782 |
+
"epoch": 0.6028416779431665,
|
| 6783 |
+
"grad_norm": 3.5490684509277344,
|
| 6784 |
+
"learning_rate": 8.16796835359635e-06,
|
| 6785 |
+
"loss": 0.0796,
|
| 6786 |
+
"step": 891
|
| 6787 |
+
},
|
| 6788 |
+
{
|
| 6789 |
+
"epoch": 0.6035182679296346,
|
| 6790 |
+
"grad_norm": 3.600038766860962,
|
| 6791 |
+
"learning_rate": 8.144752272357424e-06,
|
| 6792 |
+
"loss": 0.1059,
|
| 6793 |
+
"step": 892
|
| 6794 |
+
},
|
| 6795 |
+
{
|
| 6796 |
+
"epoch": 0.6041948579161028,
|
| 6797 |
+
"grad_norm": 6.19486665725708,
|
| 6798 |
+
"learning_rate": 8.12154654249654e-06,
|
| 6799 |
+
"loss": 0.2211,
|
| 6800 |
+
"step": 893
|
| 6801 |
+
},
|
| 6802 |
+
{
|
| 6803 |
+
"epoch": 0.6048714479025711,
|
| 6804 |
+
"grad_norm": 3.217571973800659,
|
| 6805 |
+
"learning_rate": 8.098351293490365e-06,
|
| 6806 |
+
"loss": 0.0893,
|
| 6807 |
+
"step": 894
|
| 6808 |
+
},
|
| 6809 |
+
{
|
| 6810 |
+
"epoch": 0.6055480378890392,
|
| 6811 |
+
"grad_norm": 3.447753667831421,
|
| 6812 |
+
"learning_rate": 8.07516665475708e-06,
|
| 6813 |
+
"loss": 0.1373,
|
| 6814 |
+
"step": 895
|
| 6815 |
+
},
|
| 6816 |
+
{
|
| 6817 |
+
"epoch": 0.6062246278755075,
|
| 6818 |
+
"grad_norm": 4.001631259918213,
|
| 6819 |
+
"learning_rate": 8.051992755655672e-06,
|
| 6820 |
+
"loss": 0.1635,
|
| 6821 |
+
"step": 896
|
| 6822 |
+
},
|
| 6823 |
+
{
|
| 6824 |
+
"epoch": 0.6069012178619756,
|
| 6825 |
+
"grad_norm": 3.870436191558838,
|
| 6826 |
+
"learning_rate": 8.0288297254852e-06,
|
| 6827 |
+
"loss": 0.1659,
|
| 6828 |
+
"step": 897
|
| 6829 |
+
},
|
| 6830 |
+
{
|
| 6831 |
+
"epoch": 0.6075778078484438,
|
| 6832 |
+
"grad_norm": 7.8299479484558105,
|
| 6833 |
+
"learning_rate": 8.005677693484077e-06,
|
| 6834 |
+
"loss": 0.2432,
|
| 6835 |
+
"step": 898
|
| 6836 |
+
},
|
| 6837 |
+
{
|
| 6838 |
+
"epoch": 0.6082543978349121,
|
| 6839 |
+
"grad_norm": 2.8993029594421387,
|
| 6840 |
+
"learning_rate": 7.98253678882937e-06,
|
| 6841 |
+
"loss": 0.0963,
|
| 6842 |
+
"step": 899
|
| 6843 |
+
},
|
| 6844 |
+
{
|
| 6845 |
+
"epoch": 0.6089309878213802,
|
| 6846 |
+
"grad_norm": 3.3452529907226562,
|
| 6847 |
+
"learning_rate": 7.959407140636034e-06,
|
| 6848 |
+
"loss": 0.1336,
|
| 6849 |
+
"step": 900
|
| 6850 |
+
},
|
| 6851 |
+
{
|
| 6852 |
+
"epoch": 0.6089309878213802,
|
| 6853 |
+
"eval_accuracy": 0.8,
|
| 6854 |
+
"eval_f1": 0.47342995169082125,
|
| 6855 |
+
"eval_loss": 0.42485949397087097,
|
| 6856 |
+
"eval_precision": 0.8448275862068966,
|
| 6857 |
+
"eval_recall": 0.3288590604026846,
|
| 6858 |
+
"eval_runtime": 53.6449,
|
| 6859 |
+
"eval_samples_per_second": 5.555,
|
| 6860 |
+
"eval_steps_per_second": 0.186,
|
| 6861 |
+
"step": 900
|
| 6862 |
}
|
| 6863 |
],
|
| 6864 |
"logging_steps": 1,
|
|
|
|
| 6878 |
"attributes": {}
|
| 6879 |
}
|
| 6880 |
},
|
| 6881 |
+
"total_flos": 2.732359159822418e+17,
|
| 6882 |
"train_batch_size": 8,
|
| 6883 |
"trial_name": null,
|
| 6884 |
"trial_params": null
|