Closed filljonas closed 2 years ago
Hi, and thank you for contributing to BLiTZ. I'm gonna run the tests and review the PR.
Tks, -Pi
The tests did not pass as you changed the return value of a important function of @variational_estimator
.
I've presented some approaches you can go to to solve this issue, and you can always create a new function return the values as you wish without changing this function that is a central part of BLiTZ's API.
Tks!
Hi, I restored the original function within @variational_estimator and added my own function where I changed the return type. Please let me know if there is anything else that I need to adjust.
Best, Jonas
Gonna run the tests and let you know. Thanks again :D
Now it works. I'm merging it, thou it would be nice if you could add the unit tests for this function on other PR.
For my recent work about Bayesian Neural Networks for Time Series Forecasting in Hydrology I needed a slight modification of the BLiTZ-library where the function _sampleelbo in _variationalestimator.py does not only output the total loss, but the likelihood part and performance part individually. In addition, my version outputs the model predictions.
As I think that this could be sensible addition for public use of the library, I would like to apply for a change here.
Best regards, Jonas Fill