Saving weights and logs of step 5000

events.out.tfevents.1634745538.t1v-n-f6f5b6cc-w-0.1277181.0.v2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90963df3c664485bf73c3dec95d7fb5562d10a204cef07190a1cad02697c526f
-size 23092745
+oid sha256:b62997695b8c5a06d2ec0d0cc9ed4f46bfed941a3d49fd348aceb7577bce4fd2
+size 23391025

events.out.tfevents.1635022617.t1v-n-f6f5b6cc-w-0.56283.0.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10313d235914ada691df4001460a21e4cac30520a984648f072b00ea7a854473
+size 735759

flax_model.msgpack

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5fc4dab859542930ffe9922a09c7e916d50f96ac83a28d9e14c4dd1964297b7b
+oid sha256:dc16a04ccd995ca3adccd9eec1d93449a80b7ed9095f8a2860f08d4d6ba0116d
 size 497764120

run.sh

@@ -11,11 +11,12 @@ python run_clm_flax.py \
     --block_size="512" \
     --per_device_train_batch_size="64" \
     --per_device_eval_batch_size="64" \
-    --learning_rate="5e-3" --warmup_steps="1000" \
+    --learning_rate="3e-3" \
+    --warmup_steps="0" \
     --adam_beta1="0.9" --adam_beta2="0.98" --weight_decay="0.01" \
     --overwrite_output_dir \
-    --num_train_epochs="20" \
+    --num_train_epochs="30" \
     --logging_steps="500" \
-    --save_steps="2500" \
-    --eval_steps="2500" \
+    --save_steps="5000" \
+    --eval_steps="5000" \
     --push_to_hub