python3 run_speech_recognition_seq2seq_streaming.py \ --model_name_or_path="arun100/whisper-base-bn" \ --dataset_name="mozilla-foundation/common_voice_16_0" \ --dataset_config_name="bn" \ --language="bengali" \ --train_split_name="train+validation" \ --eval_split_name="test" \ --model_index_name="Whisper Base Bengali" \ --max_steps="1000" \ --output_dir="./" \ --per_device_train_batch_size="32" \ --per_device_eval_batch_size="32" \ --gradient_accumulation_steps=2 \ --logging_steps="25" \ --learning_rate="1e-6" \ --warmup_steps="50" \ --evaluation_strategy="steps" \ --eval_steps="500" \ --save_strategy="steps" \ --save_steps="500" \ --generation_max_length="225" \ --length_column_name="input_length" \ --max_duration_in_seconds="30" \ --text_column_name="sentence" \ --freeze_feature_encoder="False" \ --report_to="tensorboard" \ --metric_for_best_model="wer" \ --greater_is_better="False" \ --load_best_model_at_end \ --gradient_checkpointing \ --fp16 \ --overwrite_output_dir \ --do_train \ --do_eval \ --predict_with_generate \ --do_normalize_eval \ --streaming="False" \ --use_auth_token \ --push_to_hub