#!/usr/bin/env bash

export TOKENIZERS_PARALLELISM=0

python ./run_mlm_flax.py \
    --push_to_hub \
    --output_dir="./" \
    --model_type="big_bird" \
    --config_name="./" \
    --tokenizer_name="./" \
    --max_seq_length="4096" \
    --weight_decay="0.0095" \
    --warmup_steps="5000" \
    --overwrite_output_dir \
    --adam_beta1="0.9" \
    --adam_beta2="0.98" \
    --logging_steps="500" \
    --eval_steps="92768" \
    --num_train_epochs="5" \
    --preprocessing_num_workers="64" \
    --save_steps="20000" \
    --adafactor \
    --learning_rate="5e-5" \
    --per_device_train_batch_size="2" \
    --per_device_eval_batch_size="2" \
    --save_total_limit="5"\
    --dtype="bfloat16" \
    #--resume_from_checkpoint="./"\
    #--gradient_accumulation_steps="4" \