uncha-hyperbolic-compositional-alignment

【效率与协作】Swap fixed entailment loss with uncertainty-guided piecewise-continuous formulation to improve part-to-whole compositional alignment in hyperbolic vision-language models by +3.0-3.1% on zero-shot classification. Use when training hyperbolic VLMs on part-object relationships and need better multi-object compositional understanding.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准