AI models
FoldFlow2
Training compute
7.6×10²¹ FLOP
Published
May 30, 2024
FoldFlow2 is an AI model developed by Dreamfold, University of Montreal / Université de Montréal, McGill University and University of Oxford (Canada and United Kingdom), first published in May 2024. It works in the biology domain, on tasks such as protein generation and protein design.
Training it took an estimated 7.6×10²¹ FLOP of compute (estimation method: hardware). It was trained on roughly 48M datapoints. Training ran on 2 NVIDIA A100 for about 96 hours.
Access: Open weights (non-commercial). Its weights are openly available. It is built on top of ESM2-650M. The reference paper has 49 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Dreamfold, University of Montreal / Université de Montréal, McGill University, University of Oxford
- Country of organization
- Canada, United Kingdom
- Domain
- Biology
- Task
- Protein generation, Protein design
- Training compute
- 7.6×10²¹ FLOP
- Compute estimation method
- Hardware
- Dataset size
- 48M
- Training hardware
- NVIDIA A100
- Chips used
- 2
- Training time
- 96 h
- Training power draw
- 1.6 kW
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Base model
- ESM2-650M
- Citations
- 49
- Epoch confidence
- Confident
SourceEpoch AI, 'AI Models'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/ai-models. Licensed under CC BY 4.0.
← All ai models