This is an implementation of Neural Processes, in Python using TensorFlow. The original was ported from the implementation in R by Kaspar Märtens. Since then it has been updated for Tensorflow 2.
Python 3 is required, using a virtual environment is recommended.
If you want to run the accompanying notebook that explains Neural Processes and shows examples, then install all requirements:
pip install -r requirements.txt
If you don't have a global installation of Jupyter, then install that as well into your environment:
pip install jupyter
You can now run Jupyter and play around with the notebook:
jupyter notebook
If you want to use the neuralprocesses code standalone, you only
need to install tensorflow, tensorflow_probability and numpy as
listed in requirements.txt. At the moment there is no support to
install the package using pip/setuptools, so either copy it over or
put it on PYTHONPATH.