Alpachino

Seeing the Forest and the Trees: Query-Aware Tokenizer for Long-Video Multimodal Language Models

$μ^2$Tokenizer: Differentiable Multi-Scale Multi-Modal Tokenizer for Radiology Report Generation