Triple Transformer Loop: Retrosynthesis Models Trained with OpenNMT

Models of the Triple Transformer Loop for retrosynthesis trained using the OpenNMT framework. The work was authored by David Kreutter from the University of Bern, with full details published in the journal Chemical Science. The dataset is associated with the paperswithcode platform and is licensed for Open Access.

Use Cases

• Training retrosynthesis prediction models based on the described Triple Transformer Loop architecture.
• Benchmarking transformer-based models for chemical reaction prediction tasks.
• Reproducing or extending the methodology detailed in the associated Chemical Science publication.

Strengths

• The models are associated with a peer-reviewed publication in Chemical Science, suggesting academic validation.
• The use of the OpenNMT framework indicates a standardized and reproducible training approach.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

University of Bern

Collection Method

Models trained using OpenNMT, as described in the associated publication.