exa_amd v0.1.0

exa_amd 0.1.0: first public release

exa_amd is a Parsl-powered workflow for accelerated materials discovery. It orchestrates ML (CGCNN), first-principles (VASP), and post-processing (Ehull + convex hull) in a scalable, HPC-friendly pipeline.

Highlights

• Packaged CLI: exa_amd
• Built-in ML assets
• Parsl configs for NERSC Perlmutter and LANL Chicoma (extensible via registry)
• End-to-end workflow: structure generation → CGCNN prediction → structure selection → dft optimization → post-processing
• Support for ternary and quaternary structures

Install

pip install "https://github.com/ML-AMD/exa-amd/releases/download/v0.1.0/exa_amd-0.1.0-py3-none-any.whl"