TLive-Omni: Omni-Modal Model for E-Commerce Live Streaming Understanding

TaoLiveAIGC · hf · 2026-08-25

TLive-Omni is an omni-modal model designed for e-commerce live streaming. It unifies image, video, audio, and text processing via timestamped token grouping. The model utilizes staged supervised training and reinforcement fine-tuning with verifiable feedback to achieve accurate real-time understanding of live content.

Original post →

More from Multimodal

Multimodal channel →