Core functions: ggml_ot

The core functions of the ggml-ot package to train, tune and test on AnnData or array-like data.

Training & hyperparameter tuning

train	Perform supervised optimal transport by ground metric learning.
train_emd2	Train GGML with exact OT (EMD2).
train_sinkhorn	Train GGML with Sinkhorn-regularized OT.
tune	Tune hyperparameters by performing a Grid Search and Cross-Validation.

Cross-validation

test	Tests ground metric on a given dataset.
train_test	Trains and cross-validates ground metrics on train-test splits.

Setup dataset

from_anndata	Dataset to train GGML based on AnnData.
from_numpy	Dataset to train GGML based on array data.