Once you rerun the model, you should be able to save it as a netcdf file after the az.from_pymc3 call: data.to_netcdf("my_data.nc"). This is built on hdf5, and is a nice way of storing pymc3 traces on disk.
Once you rerun the model, you should be able to save it as a netcdf file after the az.from_pymc3 call: data.to_netcdf("my_data.nc"). This is built on hdf5, and is a nice way of storing pymc3 traces on disk.