This repository is the official implementation of 'Gaussian Process Dynamic Mode Decomposition' an extension of the Master Thesis "Probabilistic Dynamic Mode Primitives" under the supervision of Hany Abdulsamad and Svenja Stark.
This project is compatible with Python 3.8 and conda. Further JAX is required.
You can also change the environment name if necessary.
$ conda env create -f base_env.yml
There can be some issues with installing jax.
- Kay Hansel
- All old DMD related methods are not tested yet if they still work.
- At the moment, only ARD-Kernels are supported.
- Implement ProMPs and other methods to get good benchmarks.
- The current yml file is not up to date. It contains also packages which are not used anymore.