This is home to a set of miscellaneous of Bayesian modeling projects developed using tensorflow probability and/or pymc3. My motivation for developing this is primary as a learning too. The plan is to collect a number of public domain examples and work them using a common framework and codebase using tensorflow probability or pymc3. I will do my best to document source material and any changes/modifications.
I maintain dependencies using pipenv. Most work will be done in a notebook dev environment which can be installed and activated using:
pipenv install --dev
pipenv shellI also created a docker environment that encapsulates (with too many layers) the entire development framework.
A jupyterlab notebook can be activated by running the shell script runjupyter.sh in the notebooks directory or using docker:
pipenv run notebooks/runjupyter.shdocker-compose up