2 papers across 2 sessions
a new learning scheme for multi-modal LLM, LLAVA
optimizing 3D point cloud transformer model for large-scale processing