| language: | |
| - en | |
| tags: | |
| - esc | |
| datasets: | |
| - gigaspeech | |
| To reproduce this run, execute: | |
| ```python | |
| #!/usr/bin/env bash | |
| CUDA_VISIBLE_DEVICES=0 python run_speech_recognition_whisper.py \ | |
| --model_name_or_path="medium.en" \ | |
| --dataset_name="esc-benchmark/esc-datasets" \ | |
| --dataset_config_name="gigaspeech" \ | |
| --max_steps="5000" \ | |
| --output_dir="./" \ | |
| --run_name="whisper-gigaspeech" \ | |
| --wandb_project="whisper" \ | |
| --per_device_train_batch_size="64" \ | |
| --per_device_eval_batch_size="16" \ | |
| --logging_steps="25" \ | |
| --learning_rate="1e-4" \ | |
| --warmup_steps="500" \ | |
| --report_to="wandb" \ | |
| --preprocessing_num_workers="16" \ | |
| --evaluation_strategy="steps" \ | |
| --eval_steps="1000" \ | |
| --save_strategy="steps" \ | |
| --save_steps="1000" \ | |
| --generation_max_length="224" \ | |
| --length_column_name="input_lengths" \ | |
| --gradient_checkpointing \ | |
| --group_by_length \ | |
| --freeze_encoder \ | |
| --fp16 \ | |
| --overwrite_output_dir \ | |
| --do_train \ | |
| --do_eval \ | |
| --do_predict \ | |
| --predict_with_generate \ | |
| --use_auth_token | |
| ``` | |