update docs

Author: liugangcode

docs/source/index.rst

@@ -8,7 +8,7 @@ torch-molecule documentation
 
 Welcome to the Torch-Molecule documentation. Torch-Molecule is an actively developed package designed to facilitate molecular discovery through deep learning. It offers a user-friendly interface similar to `sklearn`, and includes model checkpoints for efficient deployment and benchmarking across various molecular tasks. The package currently focuses on three main components:
 
-- Predictors: Molecular property prediction models that make predictions based on molecular graphs.
+- Predictors: Molecular property prediction models that make predictions (regression or classification) based on molecular graphs.
 - Generators: Molecular graph generators that generate new molecules.
 - Encoders: Molecular graph encoders that convert molecular graphs into a vector representation.