AI models
DARTS
Training compute
3.2×10¹⁷ FLOP
Parameters
33M
Published
Jun 24, 2018
DARTS is an AI model developed by DeepMind and Carnegie Mellon University (CMU) (United Kingdom and United States), first published in June 2018. It works in the language domain, on tasks such as language modeling and neural architecture search - nas.
Training it took an estimated 3.2×10¹⁷ FLOP of compute (estimation method: operation counting,hardware). The model has 33,000,000 parameters. It was trained on roughly 2M datapoints. Training ran on 1 NVIDIA GeForce GTX 1080 Ti for about 72 hours.
Access: Unreleased. Its weights are not openly released. The reference paper has 4,929 citations. Epoch AI rates the confidence of this record as speculative.
Full record
- Organization
- DeepMind, Carnegie Mellon University (CMU)
- Country of organization
- United Kingdom, United States
- Domain
- Language
- Task
- Language modeling, Neural Architecture Search - NAS
- Training compute
- 3.2×10¹⁷ FLOP
- Compute estimation method
- Operation counting, Hardware
- Parameters
- 33,000,000
- Dataset size
- 2M
- Training hardware
- NVIDIA GeForce GTX 1080 Ti
- Chips used
- 1
- Training time
- 72 h
- Training power draw
- 285 W
- Model accessibility
- Unreleased
- Open weights
- No
- Citations
- 4,929
- Epoch confidence
- Speculative
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