[CV] MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems? https://t.co/MzqTdehotT - Existing benchmarks for evaluating MLLMs on visual math problems may contain too much redundant text that duplicates diagram information. This allows MLLMs… https://t.co/Y85htjZm07

Xin Eric Wang@xwang_lk

3 years ago

CLIP, the backbone of SOTA Multimodal LLMs such as LLaVA (and GPT-4V?), still suffers from compositional generalization. Fine-grained visual understanding remains a bottleneck for Multimodal LLMs. To this end, our recent #NAACL2024 paper introduces a training-free method,… https://t.co/cmychcrgud

Xin Eric Wang@xwang_lk

3 years ago

CLIP is the backbone of existing SOTA Multimodal LLMs such as LLaVA (and GPT-4V?). But it still suffers from compositional generalization, and fine-grained visual understanding remains a bottleneck for Multimodal LLMs. Our recent #NAACL2024 paper introduces a training-free… https://t.co/cmychcrgud