Generative AI Meta AI Open-Sourced Perception Encoder Audiovisual (PE-AV): Audiovisual encoder powering SAM audio and large-scale multimodal retrieval by December 22, 2025 December 22, 2025 Meta Researchers Introduce Perception Encoder Audiovisual, PEAVAs a new family of encoders for joint audio and video understanding. The model learns aligned audio, video and text representations in a single … 0 FacebookTwitterPinterestEmail
