#!/usr/bin/env bash
export TOKENIZERS_PARALLELISM=0
./run_mlm_flax_stream.py \
    --output_dir="./" \
    --model_type="roberta" \
    --config_name="./" \
    --tokenizer_name="./" \
    --dataset_name="oscar" \
    --dataset_config_name="unshuffled_deduplicated_en" \
    --max_seq_length="128" \
    --weight_decay="0.01" \
    --per_device_train_batch_size="2" \
    --per_device_eval_batch_size="2" \
    --learning_rate="3e-4" \
    --warmup_steps="1000" \
    --overwrite_output_dir \
    --adam_beta1="0.9" \
    --adam_beta2="0.98" \
		--num_train_steps="1000" \
    --num_eval_samples="200" \
		--logging_steps="10" \
		--eval_steps="100" \
    --push_to_hub