Home > Categories > Recognition and synthesis of things > Speech Recognition (ASR) > Whisper

Whisper

Rating:

4.9 / 5.0

Tags

Speech Recognition, ASR, NLP, Open Source, OpenAI, Transcription, Speech Translation

Categories:
Machine learning and neural networksNatural language processingRecognition and synthesis of things
Creator OpenAI
Date 2022-09-21
Platforms Various, API
Status Active (Development)
Website openai.com
Price Model Free (model)
Sections:
Chatbots and Conversational AIDL FrameworksInformation ExtractionSpeech Recognition (ASR)Translation

Pricing Details

Free (open source). Costs may apply when using third-party services or cloud resources to run the model (e.g., via OpenAI API or cloud platforms).

Features

Automatic speech recognition (ASR), multilingual speech recognition, speech translation (to English), robust to noise and accents, open-source, various model sizes.

Integrations

Used as a library in Python and other language projects. Available via OpenAI API and can be integrated into various applications.

Preview

Whisper is a general-purpose automatic speech recognition (ASR) model developed by OpenAI. It is trained on a large dataset of diverse audio from the internet, allowing it to perform well with various accents, background noise, and technical language. Whisper is capable of not only transcribing speech in English but also automatically recognizing and translating speech from other languages into English. The model is open-source, allowing developers to use it in their applications. Various model sizes are available, offering a trade-off between speed and accuracy.