DOCSEARCHbyAlgolia
Submit Docs

DiffTaichi Documentation

repository·master·Indexed 25 days ago

https://github.com/taichi-dev/difftaichi

A differentiable programming framework for physical simulation that enables the optimization of neural network controllers via gradient descent. It includes differentiable simulators for elastic objects, fluids, height field water, rigid bodies, mass-spring systems, and billiards.

Tokens
382
Snippets
2
Records
2
Agent score
33%

What's inside DiffTaichi

  1. Run DiffTaichi simulation examples

    master

    You can run various differentiable physical simulation examples using the provided Python scripts. Below is a list of available simulators and their corresponding execution commands:

    # Differentiable Elastic Object Simulator
python3 diffmpm.py

# Differentiable 3D Elastic Object Simulator
python3 diffmpm3d.py

# Differentiable 3D Fluid Simulator
python3 liquid.py

# Differentiable Height Field Water Simulator
python3 wave.py

# Differentiable (Adversarial) Water Renderer
python3 water_renderer.py

# Differentiable Rigid Body Simulator
python3 rigid_body.py [1/2] train

# Differentiable Mass-Spring Simulator
python3 mass_spring.py [1/2/3] train

# Differentiable Billiard Simulator
python3 billiards.py

  2. Install Taichi to use DiffTaichi examples

    master

    DiffTaichi examples require the taichi library. Most examples do not require a GPU to run. Ensure you are using at least Taichi v0.8.3 for compatibility with these examples.

    python3 -m pip install taichi