Search Results for author:
Found 2 papers, 2 papers with code
OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems
Notably, the best-performing model, GPT-4V, attains an average score of 17. 23% on OlympiadBench, with a mere 11. 28% in physics, highlighting the benchmark rigor and the intricacy of physical reasoning.
Tripartite Information Mining and Integration for Image Matting
Regarding the nature of image matting, most researches have focused on solutions for transition regions.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
