ViTPose is an interactive Hugging Face demo showcasing a vision transformer model specialized for human pose estimation. Users upload images, select detection and pose models, and view visualized keypoints. It serves both as a demonstration and a reference implementation for researchers working with pose estimation models.
ViTPose is a Data science & ML workbench project. It focuses on detecting and visualizing human body poses in images using state-of-the-art vision models. It is built as an open-source project for computer vision researchers and developers. The project is open source (Open Source). ViTPose is available on the web, the command line, embeddable surfaces, and API.
Gradio-Blocks builds and maintains ViTPose, and it first shipped in 2022. Among its 3 catalogued features are Pose Estimation, Human Keypoints, and Model Visualization.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do