In these notebooks we will see how we can use physics laws and small amounts of data to solve partial differential equation such as Poisson, Burger, and others using Neural Network. The first part pytorch was used but this built on Tensorflow.
Examples for the HAL Physics Informed AI Training Session.
On HAL, the PINNs examples should be run using the openCE/1.3.1 conda environment.
I had difficulties installing Jax on HAL, so I will run the DeepONetPI example which requires Jax on Bridges2.
In these notebooks we will see how we can use physics laws and small amounts of data to solve partial differential equation such as Poisson, Burger, and others using Neural Network. The first part pytorch was used but this built on Tensorflow.