Event language
UI language
Whisper is an automatic speech recognition (ASR) system trained on 680,000 hours of multilingual and multitask supervised data collected from the web and was released in opensource around October, 2022.<br>In this talk, I will be covering what is the capabilities of Whisper model, and the various features of whisper model. This will serve as an introduction to whisper model.<br><br>One of the key things, I want to highlight with this talk, the fine tuning capabilities of Whisper model in any language as whisper support close to 90+ languages with the base model. We have got great results on fine-tuning a Malayalam base whisper model with reported accuracy of a WER of 108% we were able to fine tune to reach a WER of close to 10%(approx 90 percentage accuracy).<br>
