| python run_speech_recognition_ctc.py \ | |
| --dataset_name="mozilla-foundation/common_voice_7_0" \ | |
| --model_name_or_path="facebook/wav2vec2-xls-r-300m" \ | |
| --dataset_config_name="hi" \ | |
| --output_dir="./" \ | |
| --overwrite_output_dir \ | |
| --max_steps="8000" \ | |
| --per_device_train_batch_size="16" \ | |
| --learning_rate="3e-4" \ | |
| --warmup_steps="500" \ | |
| --save_steps="200" \ | |
| --eval_steps="400" \ | |
| --save_total_limit="3" \ | |
| --evaluation_strategy="steps" \ | |
| --text_column_name="sentence" \ | |
| --length_column_name="input_length" \ | |
| --layerdrop="0.0" \ | |
| --freeze_feature_encoder \ | |
| --gradient_checkpointing \ | |
| --fp16 \ | |
| --group_by_length \ | |
| --push_to_hub \ | |
| --use_auth_token \ | |
| --do_train \ | |
| --do_eval \ | |
| --chars_to_ignore , ? . ! - \; \: \" “ % ‘ ” � \ | |