Using the databases I will create, we want to train the following models (note specifically, when we say "X-O", we mean a REST query to the Materials Project v2 API in which we make queries like ["X-O", "X-O-*", "X-O-*-*"]).
FEFF
[ ] Ti-O
[ ] V-O
[ ] Cr-O
[ ] Mn-O
[ ] Fe-O
[ ] Co-O
[ ] Ni-O
[ ] Cu-O
VASP
[ ] Ti-O
[ ] Cu-O
Transfer-learned FEFF-to-VASP models
[ ] Ti-O
[ ] Cu-O
Tasks
For each model, there will be two different train/validation/testing splits:
"random split": each dataset contains the site-wise features and XAS pairs. The random split simply splits the dataset randomly, as we are all used to.
"materials split": the random split can lead to atoms from the same material in the training and testing set. This is obviously undesirable because the testing set should represent the "real world deployment scenario" of the model, and we will be testing on previously unseen materials, not sites. The materials split splits the data roughly evenly across materials, then saves the site data.
For each of these, and each model, we will do the following:
Train and hyper parameter tune via Optuna (Crescendo has this functionality) to find the best model architecture and hyper-parameters.
Evaluate on the testing set and analyze.
Train a "production model" on all of the data (Crescendo also has this functionality).
Put this production model into the model zoo (crescendo/extern/m3gnet/zoo).
[Optional, not necessary for a paper] Train an ensemble model so we can accurately quantify uncertainty.
We also want to demonstrate the superiority of the transfer-learned FEFF-to-VASP models over the pure VASP models.
A quick update: we should also probably include model cards akin to Facebook's Llama model. Here's that example. At the very least, we should have the following fields:
Training details (where training data was source from especially)
FEFF/VASP transition metal models
Using the databases I will create, we want to train the following models (note specifically, when we say "X-O", we mean a REST query to the Materials Project v2 API in which we make queries like
["X-O", "X-O-*", "X-O-*-*"]
).FEFF
VASP
Transfer-learned FEFF-to-VASP models
Tasks
For each model, there will be two different train/validation/testing splits:
For each of these, and each model, we will do the following:
crescendo/extern/m3gnet/zoo
).We also want to demonstrate the superiority of the transfer-learned FEFF-to-VASP models over the pure VASP models.