Papers › ChemBERTa-2: Towards Chemical Foundation Models

ChemBERTa-2: Towards Chemical Foundation Models

5 Sep 2022arXiv:2209.01712archive 2025-07-28

Walid Ahmad, Elana Simon, Seyone Chithrananda, Gabriel Grand, Bharath Ramsundar

Large pretrained models such as GPT-3 have had tremendous impact on modern natural language processing by leveraging self-supervised learning to learn salient representations that can be used to readily finetune on a wide variety of downstream tasks. We investigate the possibility of transferring such advances to molecular machine learning by building a chemical foundation model, ChemBERTa-2, using the language of SMILES. While labeled data for molecular prediction tasks is typically scarce, libraries of SMILES strings are readily available. In this work, we build upon ChemBERTa by optimizing the pretraining process. We compare multi-task and self-supervised pretraining by varying hyperparameters and pretraining dataset size, up to 77M compounds from PubChem. To our knowledge, the 77M set constitutes one of the largest datasets used for molecular pretraining to date. We find that with these pretraining improvements, we are competitive with existing state-of-the-art architectures on the MoleculeNet benchmark suite. We analyze the degree to which improvements in pretraining translate to improvement on downstream tasks.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

seyonechithrananda/bert-loves-chemistry mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Molecular Property PredictionSelf-Supervised Learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Molecular Property Prediction BACE ChemBERTa-2 (MTR-77M) RMSE 1.363 #12 of 20 Archive leaderboard report
Molecular Property Prediction BACE ChemBERTa-2 (MTR-77M) ROC-AUC 79.9 #12 of 20 Archive leaderboard report
Molecular Property Prediction BBBP ChemBERTa-2 (MTR-77M) ROC-AUC 72.8 #17 of 29 Archive leaderboard report
Molecular Property Prediction Clearance ChemBERTa-2 (MTR-77M) RMSE 48.515 #2 of 2 Archive leaderboard report
Molecular Property Prediction ESOL ChemBERTa-2 (MTR-77M) RMSE 0.889 #18 of 20 Archive leaderboard report
Molecular Property Prediction Lipophilicity ChemBERTa-2 (MTR-77M) RMSE 0.798 #9 of 13 Archive leaderboard report
Molecular Property Prediction clintox ChemBERTa-2 (MTR-77M) Molecules (M) 77 #19 of 20 Archive leaderboard report
Molecular Property Prediction clintox ChemBERTa-2 (MTR-77M) ROC-AUC 56.3 #19 of 20 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AdamAttentionAttention DropoutBPECosine AnnealingDense ConnectionsDropoutGPT-3Layer NormalizationLinear LayerLinear Warmup With Cosine AnnealingMulti-Head AttentionResidual ConnectionSoftmaxWeight Decay

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections