Evaluating different ASR systems

Loading evaluation data from HF collection

python asr_evaluation.py --load_from_disk "/external1/ksingla/earnings22_local" --num_samples 2000 --services Whissle Deepgram AssemblyAI OpenAI

Run evaluation using manifest

python asr_evaluation.py --manifest "/external1/datasets/svarah/svarah_manifest.json" --audio_folder "/external1/datasets/svarah/audio" --num_samples 2000 --services Whissle Deepgram AssemblyAI OpenAI