# From https://arxiv.org/pdf/1907.11692.pdf
python3 -c "import jax; print('TPUs', jax.device_count())"
python3 run_mlm_flax.py \
    --output_dir="./outputs" \
    --model_type="roberta" \
    --config_name="./configs/base" \
    --tokenizer_name="./" \
    --dataset_name="munggok/KoPI" \
    --cache_dir="/data/cache" \
    --dataset_config_name="full" \
    --max_seq_length="512" \
    --pad_to_max_length  \
    --per_device_train_batch_size="64" \
    --per_device_eval_batch_size="64" \
    --preprocessing_num_workers="96" \
    --adam_beta1="0.9" \
    --adam_beta2="0.98" \
    --adam_epsilon="1e-6" \
    --learning_rate="8e-5" \
    --num_train_epochs="15" \
    --weight_decay="0.01" \
    --save_strategy="steps" \
    --save_steps="10000" \
    --save_total_limit='5' \
    --warmup_steps="5000" \
    --overwrite_output_dir \
    --eval_steps="10000" \
    --logging_steps="500" \
    --dtype="bfloat16" 2>&1 | tee run.log