AI models
DD-PPO
Training compute
7.8×10²⁰ FLOP
Published
Dec 19, 2019
DD-PPO is an AI model developed by Georgia Institute of Technology, Facebook AI Research, Oregon State University and Simon Fraser University (United States, France and Canada), first published in December 2019. It works in the robotics domain, on tasks such as object detection.
Training it took an estimated 7.8×10²⁰ FLOP of compute (estimation method: hardware). It was trained on roughly 2.5B datapoints. Training ran on 64 NVIDIA V100 for about 66 hours. The compute alone is estimated at $2K in 2023 dollars.
Access: Unreleased. Its weights are not openly released. The reference paper has 611 citations. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- Georgia Institute of Technology, Facebook AI Research, Oregon State University, Simon Fraser University
- Country of organization
- United States, France, Canada
- Domain
- Robotics
- Task
- Object detection
- Training compute
- 7.8×10²⁰ FLOP
- Compute estimation method
- Hardware
- Dataset size
- 2.5B
- Training hardware
- NVIDIA V100
- Chips used
- 64
- Training time
- 66 h
- Training power draw
- 39.3 kW
- Training cost (2023 USD)
- $2K
- Numerical format
- FP32
- Model accessibility
- Unreleased
- Open weights
- No
- Citations
- 611
- Epoch confidence
- Likely
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