Saving weights and logs of epoch 1

events.out.tfevents.1625926631.t1v-n-b95d739e-w-0.271163.3.v2 → events.out.tfevents.1626000474.t1v-n-b95d739e-w-0.337760.3.v2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6cbf4ddc7baf63d39606d3230140caf1a621f97f764558e46cd005983de1a734
-size 10097175
+oid sha256:e7082ff9b84abebdd289ad6e9f6cba82ba31b34241a5c5a58c4f72c068b11314
+size 3581817

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:07b38080d5fae88718b1c9823e4281be8e2a153b18335871ce636e4122466673
+oid sha256:152e424f233af1728a250e2bfa49f2f2f2a70ae9b3036f9818a755b981f2c6cb
 size 711588089

nohup.out

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7c39e451074f8a015bbdae06b1fe6e438c9d729fa2bf8b2d6348ffd170595fb6
-size 5967379
+oid sha256:0c793a02c4ec2e4459f0f5096470c7f1ae031352132db50ce76528187d0a7ef0
+size 2037184

run.sh

@@ -11,7 +11,7 @@ python3 run_mlm_flax.py \
     --preprocessing_num_workers="64" \
     --per_device_train_batch_size="128" \
     --per_device_eval_batch_size="128" \
-    --learning_rate="7e-5" \
+    --learning_rate="1e-4" \
     --warmup_steps="1000" \
     --overwrite_output_dir \
     --num_train_epochs="10" \