DiaBayes Community#

Frequently Asked Questions#

The maximum-likelihood inversion crashes before converging#

To ensure that everything is wired-up correctly, check that you are able to run a forward simulation with your specified forward model. If that works well, make sure that the initial guess of the parameters produces a friction curve that is sufficiently close to the observed friction. For parameter estimates that are too far off, the Gauss-Newton step of the Levenberg-Marquardt algorithm may fail.

I am getting a mysterious error#

Under the hood, most of the computations are handled by JAX, which compiles and ports the code to a GPU or multi-core CPU. One disadvantage of this automatic optimisation and acceleration, is that the error messages that are being produced tend to be rather cryptic, and often the final Exception is not at all informative of the code segment that causes trouble. To facilitate debugging, you can turn off the Just-in-Time (JIT) compiler by including the following at the top of your script:

import jax
jax.config.update("jax_disable_jit", True)

If NaNs are produced, it may also help to turn on NaN debugging:

jax.config.update("jax_debug_nans", True)

Note that jax.config should be set only at the start of your script before any JAX routines are called. The configuration will stay enabled until the script terminates.

How to cite?#

A publication describing this software package is underway. Until then, feel free to refer to the GitHub repository: https://github.com/martijnende/diabayes

How to contribute?#

If you have ideas for new features to include, you can submit an issue on GitHub to discuss it. See the User Guide for general guidelines on how to extend DiaBayes with new physics. Contributed example Jupyter Notebooks and documentation improvements are also welcome.