Multimodal Benchmarks, Visual Task Transfer, and 3D Object Generation

AI Papers Podcast

MMIU: Multimodal Multi-image Understanding for Evaluating Large Vision-Language Models LLaVA-OneVision: Easy Visual Task Transfer An Object is Worth 64x64 Pixels: Generating 3D Object via Image Diffusion MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine IPAdapter-Instruct: Resolving Ambiguity in Image-based Conditioning using Instruct Prompts Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters Diffusion Models as Data Mining Tools