Ammmob/PixelSmile
A diffusion model system for fine-grained facial expression editing with continuous control and identity preservation.
Not currently ranked — collecting fresh signals.
star history
PixelSmile is a diffusion-based image editing system for manipulating facial expressions with continuous control over expression attributes. It aims to reduce semantic entanglement between different expression features while preserving subject identity. The project includes model weights hosted on HuggingFace, a live demo, training and inference code, and a benchmark dataset (FFE-Bench) for evaluating fine-grained facial expression editing performance.
Frequently asked
- What is Ammmob/PixelSmile?
- A diffusion model system for fine-grained facial expression editing with continuous control and identity preservation.
- Is PixelSmile open source?
- Yes — Ammmob/PixelSmile is open source, released under the Apache-2.0 license.
- What language is PixelSmile written in?
- Ammmob/PixelSmile is primarily written in Python.
- How popular is PixelSmile?
- Ammmob/PixelSmile has 478 stars on GitHub.
- Where can I find PixelSmile?
- Ammmob/PixelSmile is on GitHub at https://github.com/Ammmob/PixelSmile.