Tianyi Wei

HairCLIP. [CVPR 2022] HairCLIP: Design Your Hair by Text and Reference Image

HairCLIPv2. [ICCV 2023] HairCLIPv2: Unifying Hair Editing via Proxy Feature Blending

e2style. [TIP 2022] E2Style: Improve the Efficiency and Effectiveness of StyleGAN Inversion

FreeFlux. [ICCV 2025] FreeFlux: Understanding and Exploiting Layer-Specific Roles in RoPE-Based MMDiT for Versatile Image Editing

EnMMDiT. [TPAMI 2026] Enhancing MMDiT-Based Text-to-Image Models for Similar Subject Generation

Deep-Flow-Guided-Video-Inpainting. pytorch implementation for "Deep Flow-Guided Video Inpainting"(CVPR'19)

inversion. JavaScript

glide-text2im. GLIDE: a diffusion-based text-conditional image synthesis model