Upload and evaluate video models
Leaderboard and arena of Video Generation models
Create masks and inpaint video
Generate a cartoon video from two images
https://huggingface.co/papers/2501.03006
Train a custom video model
Generate Talking avatars from Text-to-Speech
MagicTime: Time-lapse Video Generation Models as Metamorphic
Remove/Change background of video.
Generate and apply matching music background to video shot
Generate animated videos from text prompts
Generate videos from text prompts
HQ human motion video gen with pose-guided control
VBench Leaderboard is a comprehensive evaluation platform designed for video model benchmarking. It allows users to upload and evaluate video models, providing detailed performance metrics and comparative analysis. This tool is particularly useful for researchers and developers in the video generation domain, enabling them to track advancements and improvements in their models over time. VBench Leaderboard serves as a centralized hub for model comparison, fostering innovation and transparency in the field.
What types of video models can I evaluate on VBench Leaderboard?
VBench Leaderboard supports a wide range of video models, including but not limited to generative models, video-to-video translation models, and video enhancement models.
Is the VBench Leaderboard public?
Yes, the Leaderboard is publicly accessible, but you can choose to keep your model's results private if you prefer not to share them with the community.
What performance metrics does VBench Leaderboard use?
The platform evaluates models based on accuracy, inference speed, and memory usage, among other customizable metrics.