LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via Hybrid Architecture Paper • 2409.02889 • Published Sep 4 • 54
