unsloth/gemma-4-26B-A4B-it-qat-GGUF · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
Welcome! Type "help" for available commands.
$
Machine Learning Model
Summary
VOID is an open-source AI model developed by Netflix designed for advanced video object removal that accounts for physical interactions, such as items falling after an object is deleted. It utilizes a CogVideoX 3B Transformer architecture with specialized quadmask conditioning to maintain scene consistency.
Highlights
- Removes objects along with their induced physical effects like shadows, reflections, and displaced items.
- Uses a 'quadmask' system (remove, overlap, affected, background) for precise spatial conditioning.
- Features a two-pass inference process where the second pass uses optical flow-warped latent initialization to improve temporal consistency.
- Requires high-end hardware with at least 40GB of VRAM (e.g., NVIDIA A100).
auto-generated
Context
Audience
AI Researchers and Machine Learning Engineers
DomainComputer Vision
Formatopen source model repository
Accessopen source
Topics
Related
CogVideoXVideo InpaintingSAM2GeminiTemporal Consistency
