DiffSBDD (CrossDocked)
DiffSBDD (CrossDocked) is an AI model developed by Ecole Polytechnique Federale de Lausanne (EPFL), University of Cambridge, Cornell University, Chinese Academy of Mathematics and System Science, University of Rome, Microsoft Research and 2 more (Switzerland, United Kingdom, United States, China and 2 more), first published in October 2022. It works in the biology domain, on tasks such as drug discovery.
Training it took an estimated 2.7×10²⁰ FLOP of compute (estimation method: hardware). It was trained on roughly 2.9M datapoints. Training ran on 1 NVIDIA A100 for about 600 hours.
Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 404 citations. Epoch AI rates the confidence of this record as confident.
- Organization
- Ecole Polytechnique Federale de Lausanne (EPFL), University of Cambridge, Cornell University, Chinese Academy of Mathematics and System Science, University of Rome, Microsoft Research, University of Oxford, AITHYRA Institute
- Country of organization
- Switzerland, United Kingdom, United States, China, Italy, Austria
- Domain
- Biology
- Task
- Drug discovery
- Training compute
- 2.7×10²⁰ FLOP
- Compute estimation method
- Hardware
- Dataset size
- 2.9M
- Training hardware
- NVIDIA A100
- Chips used
- 1
- Training time
- 600 h
- Training power draw
- 440 W
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Citations
- 404
- Epoch confidence
- Confident