In these tutorials exercises, we review different AI approaches to predicting ground or excited state energy levels simply based on the atomic structure of materials. Here we focus on molecules and molecular orbitals, but the same principles apply to crystal structures and electronic bands.
The tutorials have been prepared by Milica Todorović (University of Turku) and Kunal Ghosh (Aalto University).
The notebooks are run using Mahti web interface, and they rely on CSC's PyTorch
module pytorch/2.6.
- Go to https://www.mahti.csc.fi
- Select Jupyter for courses
- Reservation: select if applicable
- Project: select suitable CSC project
- Course module: AI4Spec
- Partition: gpusmall (default resource settings)
- Launch!