Compare Whisper vs. Whisper Notes

We have developed and are releasing an open-source neural network named Whisper, which achieves levels of accuracy and resilience in English speech recognition that are comparable to human performance. This automatic speech recognition (ASR) system is trained on an extensive dataset comprising 680,000 hours of multilingual and multitask supervised information gathered from online sources. Our research demonstrates that leveraging such a comprehensive and varied dataset significantly enhances the system's capability to handle different accents, ambient noise, and specialized terminology. Additionally, Whisper facilitates transcription across various languages and provides translation into English from those languages. We are making available both the models and the inference code to support the development of practical applications and to encourage further exploration in the field of robust speech processing. The architecture of Whisper follows a straightforward end-to-end design, utilizing an encoder-decoder Transformer framework. The process begins with dividing the input audio into 30-second segments, which are then transformed into log-Mel spectrograms before being input into the encoder. By making this technology accessible, we aim to foster innovation in speech recognition technologies.

Whisper Notes is a voice transcription application that operates offline, enabling users to convert spoken language into text with precision by utilizing the sophisticated Whisper model, compatible with both iOS and MacOS devices. This tool is ideal for capturing your everyday musings through voice input, as well as for transcribing audio recordings from meetings. By processing these tasks locally, Whisper Notes ensures that your personal information remains secure and private throughout the transcription process. Additionally, its user-friendly interface makes it accessible for anyone looking to streamline their note-taking experience.