superstar1225/DensePose_from_WiFi — reverse-engineered prompt

Build me a Python research prototype for estimating DensePose from WiFi signals, like the paper in this repo describes.

I want it to take WiFi phase and amplitude data as input and use a deep learning model to predict dense human pose, including UV coordinates and the 24 human body regions. Please make it easy to run training and inference, with a clean notebook for experimenting and a normal Python script for repeatable runs. If the real dataset is not included, add a small fake sample dataset so the full pipeline still runs, and make it very clear where real WiFi data should go.

Use PyTorch, support GPU if available, and include simple visual outputs so I can see the predicted pose results. Please also write clear setup steps in the README, including the conda environment, dependencies, and example commands. Look up current docs online if you need to.