Jan Kautz

SpatialRGPT enhances object arrangement understanding in Vision Language Models.

New adaptable models can meet diverse needs without retraining.

MambaVision combines Mamba and Transformers for better image recognition.

This study explores methods to create smaller language models effectively and affordably.

This article analyzes model performance across various tasks and datasets.

A new method improves data quality for visual language models using augmentation techniques.

A method to shrink language models without sacrificing effectiveness through pruning and distillation.

A new method enhances LLM performance while reducing complexity.

NaVILA helps robots navigate using language and vision.

A look at Gated DeltaNet and its impact on language models.

Discover emerging techniques revolutionizing how machines see and understand images.

StreamChat transforms how we engage with streaming video in real-time.