VILA1.5-40B

Reported on 2 benchmarks across 2 tasks · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing1 result

Question AnsweringonMSVD-QA
Accuracy· 2023-12-12
80.1
best: 80.3 (Tarsier (34B))
SOTA
VILA: On Pre-training for Visual Language Models arXiv:2312.07533

Reasoning1 result

Video Question AnsweringonMSVD-QA
Accuracy· 2023-12-12
80.1
best: 80.3 (Tarsier (34B))
SOTA
VILA: On Pre-training for Visual Language Models arXiv:2312.07533