aka "Bayesian Methods for Hackers": An introduction to Bayesian methods + probabilistic programming with a computation/understanding-first, mathematics-second point of view. All in pure Python ;)
After having installs of Theano kill my Windows 10 machine repeatedly, I decided to use the Docker image (pymc/pymc3) to go through the book. I had no issues (after a couple of false starts) sharing my Windows scripts with the running image), but when I try to run first (!) Jupyter script in Chapter 1 that uses pymc, I get message after message about missing components - fastprogress, h5py, typing_extensions, dill,, arviz. Even after using conda to add all those packages, I get errors about "no model on the context stack"
I've used dozens of packages, including Tensorflow / Keras, and never seen issues like this. I'm pretty much ready to give up - I'd like to learn this, but I have only so many days to burn up on this stuff. Is there some simple change to what I'm doing that would make this stuff work better? Are there any instructions somewhere about using the book with a Docker image?
After having installs of Theano kill my Windows 10 machine repeatedly, I decided to use the Docker image (pymc/pymc3) to go through the book. I had no issues (after a couple of false starts) sharing my Windows scripts with the running image), but when I try to run first (!) Jupyter script in Chapter 1 that uses pymc, I get message after message about missing components - fastprogress, h5py, typing_extensions, dill,, arviz. Even after using conda to add all those packages, I get errors about "no model on the context stack"
I've used dozens of packages, including Tensorflow / Keras, and never seen issues like this. I'm pretty much ready to give up - I'd like to learn this, but I have only so many days to burn up on this stuff. Is there some simple change to what I'm doing that would make this stuff work better? Are there any instructions somewhere about using the book with a Docker image?