Transformer-CNN: Fast and Reliable Tool for QSAR

21 October 2019, Version 1
This content is a preprint and has not undergone peer review at the time of posting.


We present SMILES-embeddings derived from internal encoder state of a Transformer model trained to canonize SMILES as a Seq2Seq problem. Using CharNN architecture upon the embeddings results in a higher quality QSAR/QSPR models on diverse benchmark datasets including regression and classification tasks. The proposed Transformer-CNN method uses SMILES augmentation for training and inference, and thus the prognosis grounds on an internal consensus. Both the augmentation and transfer learning based on embedding allows the method to provide good results for small datasets. We discuss the reasons for such effectiveness and draft future directions for the development of the method. The source code and the embeddings are available on, whereas the OCHEM environment ( hosts its on-line implementation.


Transformer model
Convolutional neural networks
Character-based models


Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.