SeamlessM4T v2
Meta's open model that translates speech straight into speech.
FreeOpen sourceRuns on your PC
- Best for
- Research on offline voice translation
- On your PC
- Technical
- Works on
- Linux, Windows, Mac
- License
- CC-BY-NC-4.0
What’s good
- Speech in about 100 languages, spoken output in 35
- Speech to speech, speech to text and text translation
Downsides
- Non-commercial license, no business use
- From 2023, behind today's chat AIs in quality
- Technical setup with Python
The details.
- Version
- seamless-m4t-v2-largeReleased 2023-12
- Open source
- Yes, CC-BY-NC-4.0
- Made by
- Meta
- Website
- huggingface.co
Should you use it?
It is #9 of 10. Good for research on offline voice translation; for most people, Gemini is the better start.
Can I run it myself?
Yes. Technical: command line or python.
Use Transformers, Seamless Communication code.
- You need
- NVIDIA GPU with 8 GB recommended
- Setup
- TechnicalCommand line or Python
- Laptop, 8 GBToo small
- PC, 16 GB+Too small
- Mac, 16 GBToo small
- Mac, 32 GBToo small
- Mac, 64 GB+Too small
- PC, 8 GB GPURuns
- PC, 12 GB GPURuns
- PC, 16 GB GPURuns
- PC, 24 GB GPURuns
- WorkstationRuns