Reproduce “HyperCLIP++: Fine-tuning CLIP forOpen-vocabulary Semantic Segmentation in Hyperbolic Space” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Zelin Peng, Zhengqin Xu, Changsong Wen, Yu Huang, Yaoming Wang, Xiaokang Yang, Wei Shen

HyperCLIP++ adapts CLIP for open-vocabulary semantic segmentation by learning small hyperbolic radius adjustments in its frozen image and text pathways, coordinated by Dual Cross-Relation Communication. Smaller radii represent finer-grained information in the authors’ interpretation. On ADE20K A-847 with ViT-B/16, it reaches 13.1 mIoU versus CAT-Seg’s 12.0, with fewer trainable parameters but slightly higher latency.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.