CASA Gallery
🏠
4
Video Gallery for CASA: Cross-Attention over Self-Attention
CASA
CASA: Cross-Attention over Self-Attention for Efficient Vision-Language Fusion on long-context streaming inputs
-
-
CASA: Cross-Attention via Self-Attention for Efficient Vision-Language Fusion
Paper • 2512.19535 • Published • 14 -
kyutai/CASA-Qwen2_5-VL-3B
Image-Text-to-Text • 4B • Updated • 155 • 3 -
kyutai/CASA-Qwen2_5-VL-3B-Shared
Image-Text-to-Text • 4B • Updated • 33 -
kyutai/CASA-Qwen2_5-VL-3B-LiveCC
Video-Text-to-Text • 4B • Updated • 165 • 6 -
kyutai/CASA-Helium1-VL-2B
Image-Text-to-Text • 3B • Updated • 204 • 10 -
kyutai/CASA-Helium1-VL-2B-Shared
Image-Text-to-Text • 3B • Updated • 32
kyutai/Helium1-VL-2B
Image-Text-to-Text • 3B • Updated • 153 • 2Note Token-insertion VLM baseline trained on Helium1-2B