How can we improve the documentation?
In the quacc manuscript, I have a section like the following:
To streamline the MLIP evaluation process, quacc provides predefined DFT recipes designed to reproduce the computational settings used for several widely adopted MLIP datasets, including OC20,50 OMat24,51 OMC25,52 OMol25,53 ODAC25,54,55 MPtrj,56,57 MatPES,58 and MP-ALOE.59 The same recipes can be used both to construct internally consistent benchmark sets and to generate additional training configurations that extend an existing dataset without introducing potential methodological incompatibilities.
I would like to have documentation in quacc to demonstrate how people can use quacc to generate training set-compatible DFT calculations. Perhaps we would have a table of the various MLIP training datasets and the corresponding recipes in quacc. I have added most of them.
See the /docs folder for how the web documentation is built and written.
How can we improve the documentation?
In the quacc manuscript, I have a section like the following:
I would like to have documentation in quacc to demonstrate how people can use quacc to generate training set-compatible DFT calculations. Perhaps we would have a table of the various MLIP training datasets and the corresponding recipes in quacc. I have added most of them.
See the /docs folder for how the web documentation is built and written.