Testing PyTorch implementation of absimal
# make a new conda environment
conda create -n abismal_torch python=3.12
conda activate abismal_torch
# install torch
pip install torch
# ===== optional: test GPU support by torch =====
# python -c "import torch; print(torch.cuda.is_available())"
# install abismal-torch
pip install abismal-torch
# ===== alternatively, if you want the latest version from source or make changes to the code =====
# git clone [email protected]:DorisMai/abismal_torch.git
# cd abismal_torch
# pip install -e .Basic usage:
abismal-yaml fitExample of overriding default values by command line flags:
abismal-yaml fit --data.dmin 1.5 --mtz_output.save_every_n_epoch=10Example of overriding default values by yaml files:
abismal-yaml fit -c abismal_torch/command_line/configs/custom_config.yaml