VERSION 0.1 · LOCAL CPU SIMULATION

Load a robot. Add a pair of shoes.

Six original rigid shoe proxies across trainers, ankle boots and slippers, with left/right identities, three supported robot models and reusable scene files. Inspect the native 3D geometry, choose a pair, then download the scene and run it locally.

These are synthetic geometry and contact-development assets. Dimensions, mass and friction are declared assumptions. Deformation, shoe grasping and hardware transfer have not been validated. The browser replays recorded settling; native physics runs locally.

Run your own scene

git clone https://github.com/Dvidia-Inference/dvidia-training.git
cd dvidia-training
git checkout d291ee11f28b0d3c7a0b9ed062174e46fe764d15
python3.12 -m venv .venv
.venv/bin/python -m pip install '.[arm]'
.venv/bin/dvidia-assets fetch piper --output runs/assets
.venv/bin/dvidia-assets scene --robot piper --assets runs/assets --pair trainer --output runs/my-shoe-scene
.venv/bin/python -m http.server 8360 --bind 127.0.0.1 --directory runs/my-shoe-scene

Open localhost:8360 after the command finishes. On Windows use .venv\Scripts\ executables. Provisioning needs a network connection; scene simulation uses the pinned local models. A discrete GPU is not required. Minimum memory and hard real-time performance remain unmeasured.