Aurigene-AI/MoLFormer-XL-both-10pct
Feature Extraction • 46.8M • Updated • 35
Transformer encoders for SMILES/graph/image molecular representations — embeddings, ADMET & MoleculeNet property prediction.
Note IBM MoLFormer-XL: linear-attention SMILES encoder pretrained on 1.1B PubChem+ZINC molecules
Note DeepChem ChemBERTa-2 (77M) pretrained with multitask regression on 200 RDKit descriptors
Note IBM multi-view (SMILES + graph + image) small-molecule encoder for MoleculeNet-style property prediction