Kitten TTS model

• August 06, 2025 at 09:36 IST
• tool

I came across this post from Divam Gupta who has developed a tiny text-to-speech model which is only 25 MB in size and runs without a GPU. It's 15M parameter model and is very fast even on low-end devices.

You can find the model on GitHub here and also on HuggingFace here.

I think, it would be great to use this on a Raspberry Pi when using a physical device.

• ← Previous
  OpenAI launches open-source AI models
• Next →
  Perplexity ranking patterns