Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 25 additions & 6 deletions docs/Documentation/Applications/mace.md
Original file line number Diff line number Diff line change
Expand Up @@ -6,17 +6,36 @@ The Multi Atomic Cluster Expansion (MACE) is a neural network-based approach for

## Installing MACE standalone package

The Kestrel-optimized MACE conda environment can be found at: `/nopt/nrel/apps/gpu_stack/MACE/environments`
The Kestrel-optimized MACE conda environment can be found at: `/nopt/nrel/apps/gpu_stack/software/mace/environments`

Copy the `mace_kestrel.yml` to one of your own directories, and build the environment as from that directory as follows:

`conda create -n <your_mace_env_name> -f mace_kestrel.yml`
`$ module load conda`

`conda env create --file mace_kestrel.yml`
`$ conda env create -n <your_mace_env_name> -f mace_kestrel.yml`

Then install MACE within the conda environment:
Then check that MACE is installed within the conda environment:

`python -m pip install -e mace`
`$ conda list mace`

Should show:
# Name Version Build Channel
mace-torch 0.3.5 pypi_0 pypi

And that it's working:

`$ python`
`>>> import torch`
`>>> x = torch.rand(5, 3)`
`>>> print(x)`
`>>> exit`

Should show the following after the print statement:
tensor([[0.1828, 0.4496, 0.8743],
[0.9411, 0.3940, 0.1104],
[0.9148, 0.7720, 0.7389],
[0.6979, 0.0421, 0.1763],
[0.6563, 0.7603, 0.8917]])

This environment is required for all commands in the MACE standalone package, including those for training (`mace_run_train`), evaluation (`mace_eval_configs`) and creating lammps models (`mace_create_lammps_model`).

Expand Down Expand Up @@ -93,4 +112,4 @@ mace_eval_configs \
--output="./your_output.xyz"
```

Alternatively, you can use the `.model` file within [ASE detailed here](https://mace-docs.readthedocs.io/en/latest/guide/ase.html#running-md-simulations), or the `.pt` file within [LAMMPS detailed here](https://mace-docs.readthedocs.io/en/latest/guide/lammps.html#id2).
Alternatively, you can use the `.model` file within [ASE detailed here](https://mace-docs.readthedocs.io/en/latest/guide/ase.html#running-md-simulations), or the `.pt` file within [LAMMPS detailed here](https://mace-docs.readthedocs.io/en/latest/guide/lammps.html#id2).