What is Whisper from OpenAI ? A speech to text, or automatic speech recognition model.
The team at OpenAI developed a speech recognition system to predict audio transcripts from the internet. They used 680,000 hours of audio data for system. The amazing part is the newly developed system can compete with a prior fully supervised model.
What can it do ?
- English Transcription
- Any-to-English Speech Transcription
- Non-English Transcription
- No Speech Prediction
Conclusion
In conclusion, Whisper shows that it can achieve massive results with speech recognition without the of complicated self-supervised and self-trained techniques. By simply training with a large and diverse dataset, they can make the speech recognition system more robust and accurate!
References:
- Open AI
- Whisper
- Whisper — Robust Speech Recognition via Large-Scale Weak Supervision
- Hugging Faces — Whisper
Connect
Are you looking to stay up-to-date with my writing and projects? Connect with me on Linkedin and follow the links below to learn more!
Check out my website — cover letter builder — and read my official blog post to unlock the power of automated background removal with AI.
Official Blog — https://aiapplicationsblog.com/unbelievable-ai-sites-you-never-knew-existed/