Top Tweets for #beoptimism

Fleming-VL: Towards Universal Medical Visual Reasoning with Multimodal LLMs I'm not sure why a Beijing-based hedge fund (Ubiquant) is building state-of-the-art medical multimodal models but awesome! "we propose Fleming-VL, a unified end-to-end framework for comprehensive medical visual understanding across heterogeneous modalities. "First, we scale up the pretraining process by integrating long-context data from both natural images and medical-specific domains to establish robust foundational capabilities. Second, during fine-tuning, we strategically complement the training with rare medical data, including holistic video analysis and underrepresented 2D modalities such as ultrasound and dermoscopy images. Third, we extend existing medical multimodal evaluation frameworks to incorporate 3D volumetric and video understanding benchmarks."