MOSS-VL-Realtime Open-Sourced for Real-Time Vision
huggingface · x · 2026-07-14
MOSS-VL-Realtime has been open-sourced as a VLM designed for real-time visual understanding:
- 11B parameters, 256k context
- Released under the Apache-2.0 license
- Design goals include:
- Staying silent when information is insufficient
- Allowing questions at any point during a video stream
- Generating answers while watching
- Correcting or interrupting its own answers when the scene changes
- Supporting both Chinese and English comprehension
The creators are inviting users to try it out and provide feedback.
Related event: MOSS-VL-Realtime Open-Sourced for Real-time Vision(2 posts)→
More from Multimodal
- Fake BTS of a dam disaster made with Dreamina Seedance for 29 cents — Standard_Spell5873 · 2026-10-27
- Redditor struggles with outfit swap: Qwen Image 2.1 breaks body proportions — Existing-Box-7866 · 2026-10-27
- Consistent characters in ComfyUI: RealVisXL V5 + img2img face reference workflow — MiaLarsonJL · 2026-10-27
- 29 cents gets you a Michael Bay rooftop shot with Seedance — PersonalSpace_4225 · 2026-10-27
- Moss Image Turbo 3-Step LoRA Tested: Fast Generation at 6-8 Steps — JadedScorpion · 2026-10-27
- Magnific teases October 28 release: 'We're making it move' — cuenca · 2026-10-27