Frequently asked questions

1. What should I do when my ‘posterior samples are outside of the prior support’ in SNPE?
2. Can the algorithms deal with invalid data, e.g., NaN or inf?
3. When using multiple workers, I get a pickling error. Can I still use multiprocessing?
4. Can I use the GPU for training the density estimator?
5. How should I save and load objects in sbi?
6. Can I stop neural network training and resume it later?
7. How can I use a prior that is not defined in PyTorch?

See also discussion page and issue tracker on the sbi GitHub repository for recent questions and problems.