KI-Forschungspapiere Täglich

LLaMA-Omni: Seamless Speech Interaction with Large Language Models

GroUSE: A Benchmark to Evaluate Evaluators in Grounded Question Answering

INTRA: Interaction Relationship-aware Weakly Supervised Affordance Grounding

SongCreator: Lyrics-based Universal Song Generation

Draw an Audio: Leveraging Multi-Instruction for Video-to-Audio Synthesis

SaRA: High-Efficient Diffusion Model Fine-tuning with Progressive Sparse Low-Rank Adaptation

LEIA: Latent View-invariant Embeddings for Implicit 3D Articulation