We wrote a Tesseract-JAX demo on using differentiable programming to optimize fluid simulations. The goal is to find initial conditions that evolve into a target pattern (the Pasteur Labs logo) using JAX-CFD and gradient-based optimization.
Check it out here!
The tutorial covers:
- Building differentiable fluid simulators using JAX-CFD wrapped in a Tesseract container
- Leveraging gradient-based optimization with scipy for complex physical simulations
