Last released Aug 22, 2026
Multimodal embeddings for text, image, audio, and video in one shared vector space, trained on a single consumer GPU
Supported by
