Environments

This repo provides the environments suitable for running our ice codes on Della-gpu, group workstation, and your local machine

cudatoolkit is requrired for IAS's cluster (according to Yongji and Ray). Ray also tested and confirmed the cudatoolkit is also compatible with our workstation (though not required)

cudann is the toolbox needed to activate GPU for tensorflow

Princeton Cluster

environment-cluster-tf24.yml: For running our code using Tensorflow 2.4 on Della-gpu

Group Workstation

environment-wokstation-tf24.yml: For running our code using Tensorflow 2.4 on our workstation

environment-wokstation-tf115.yml: For running our code using Tensorflow 1.15 on our workstation

environment-wokstation-tf114.yml: For running the origional PINN code in Raissi (2019). It uses tf.contrib and thus too old for Tensorflow 1.15

Local Machines (e.g. personal computer)

environment-win.yml: For running our code using Tensorflow 2.5 on a windows computer (Note that Tensorflow 2.4 wouldn't work on Windows)

environment-osx.yml: For running our code using Tensorflow 2.4 on a mac computer