Wenquan Lu

HandRefiner. [ACM MM 2024] Offical Code for "HandRefiner: Refining Malformed Hands in Generated Images by Diffusion-based Conditional Inpainting"

huginn-latent-cot. [COLM 2025: 1st Workshop on the Application of LLM Explainability to Reasoning and Planning] Latent Chain-of-Thought? Decoding the Depth-Recurrent Transformer

PrAg-PO. PrAg-PO improves RL-based mathematical reasoning by training models across diverse prompt templates with template-specific format rewards, yielding more diverse reasoning trajectories, stronger benchmark accuracy, and more stable training.

noisy_dinov2. [NeurIPS 2025] "Ditch the Denoiser: Emergence of Noise Robustness in Self-Supervised Learning from Data Curriculum": Improve Noise-robustness of Joint-embedding SSL Models, e.g., DINOv2

rl-reasoning-optimizer. Fully open reproduction of DeepSeek-R1

mocov3. Python