Two polished examples for band-gap prediction from Materials Project data.
01_tabular_baselines_bandgap.ipynb— matminer features → Ridge/XGB; IID + OOD, parity plots.02_ann_bandgap.ipynb— ANN (PyTorch) vs tree ensembles; leak-free CV; saved figures.
conda env create -f environment.yml
conda activate material-ml
export MP_API_KEY="your_api_key"