How does OpenAI Whisper's medium.en, large or whisper-large-v2 compare in terms of word error rate?
#79
by Franck-Dernoncourt - opened
Hi, they discuss the WER of each of the Whisper models in section D of their paper "Robust Speech Recognition via Large-Scale Weak Supervision" (pg. 22 - pg. 26). This paper is linked in the model card.