Gradio(@Gradio):<190 millisecond for image generation on a T4. Image generated on every keystroke. ML-Intern distilled a 260M text-to-image model from 100 network passes per image to just 4. It runs on a small GPU Space, or right in your browser with WebGPU. Links in reply⬇️

Gradio

@Gradio

Build and share machine learning apps in 3 lines of Python. Part of the @Huggingface family 🤗.

Joined April 2019

12 Following 57.8K Followers

Gradio@Gradio

2026.10.01 11:06

<190 millisecond for image generation on a T4. Image generated on every keystroke. ML-Intern distilled a 260M text-to-image model from 100 network passes per image to just 4. It runs on a small GPU Space, or right in your browser with WebGPU. Links in reply⬇️