created backup

events.out.tfevents.1641299828.t1v-n-358ff5d1-w-0.355877.3.v2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c479c58a311b3c3bef4df1082214c3f83ba33424f1553351cf618d2d7d2c4ad1
+size 40

flax_model_backup_final.msgpack

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5255008970a84f6e4d55177ca133107efb4c186c01834c6dac31da4bc0e34e91
+size 498796983

run_step2.sh

@@ -0,0 +1,26 @@
+./run_mlm_flax.py \
+    --output_dir="./" \
+    --model_type="roberta" \
+    --model_name_or_path="./" \
+    --config_name="./" \
+    --tokenizer_name="./" \
+    --dataset_name="NbAiLab/NCC" \
+    --cache_dir="/mnt/disks/flaxdisk/cache/" \
+    --max_seq_length="128" \
+    --weight_decay="0.01" \
+    --per_device_train_batch_size="200" \
+    --per_device_eval_batch_size="200" \
+    --learning_rate="4e-4" \
+    --warmup_steps="10000" \
+    --overwrite_output_dir \
+    --num_train_epochs="10" \
+    --adam_beta1="0.9" \
+    --adam_beta2="0.98" \
+    --adam_epsilon="1e-6" \
+    --logging_steps="10000" \
+    --save_steps="10000" \
+    --eval_steps="10000" \
+    --preprocessing_num_workers="64" \
+    --auth_token="True" \
+    --dtype="bfloat16" \
+    --push_to_hub