Product launches

AGIBOT Releases WITA-Omni Embodied Omni-Modal Interaction Large Model

AGIBOT officially released the WITA-Omni omni-modal large model, claimed to be the industry's first embodied omni-modal end-to-end interaction large model. It supports tactile and proprioceptive state inputs and action and expression outputs, with interaction response as fast as about 400 milliseconds and wake-word-free natural dialogue.

Published · Updated · Source: IT之家 · 机器人筛选 · AI translation
Illustration
IT Home, November 10: AGIBOT's official WeChat account announced in an evening post the official release of WITA-Omni (Silicon Light Motion Language Omni-Modal Large Model), which the company calls the industry's first embodied omni-modal end-to-end interaction large model. The model is deeply trained on robot proprioceptive perception data and, in addition to audio-visual input and output, adds inputs such as touch and proprioceptive state, as well as outputs such as robot actions and expressions. Based on a streaming full-duplex architecture, WITA-Omni can listen, think, respond, move, and make expressions simultaneously, with interaction response as fast as about 400 milliseconds and an average of about 600 milliseconds, nearly 3 times faster than the industry average. WITA-Omni supports omni-modal rejection, enabling natural dialogue without wake words. It combines dialogue orientation, video, audio, voiceprint, lip movement, posture, semantics, and context to decide whether to respond, reducing false triggers amid interference such as multi-person conversations. The model supports free communication in over a hundred languages and dialects without switching, and enables streaming simultaneous interpretation and bidirectional translation. Its emotional expression covers dozens of emotions at different intensities, can visualize complex content in real time on a facial screen, and can proactively initiate interactions. In July this year, WITA-Omni Preview topped the DailyOmni omni-modal understanding leaderboard with a composite score of 85.21. The above information comes from the original source and was organized by AI.
Read the original source

Related news