AI models
Discriminator Guidance
Training compute
2.2×10²⁰ FLOP
Published
Nov 28, 2022
Discriminator Guidance is an AI model developed by Korea Advanced Institute of Science and Technology (KAIST) and NAVER (South Korea), first published in November 2022. It works in the image generation domain, on tasks such as image generation.
Training it took an estimated 2.2×10²⁰ FLOP of compute (estimation method: hardware). It was trained on roughly 328M datapoints. Training ran on NVIDIA A100 PCIe for about 481 hours. The compute alone is estimated at $338 in 2023 dollars.
Access: Open weights (non-commercial). Its weights are openly available. The reference paper has 110 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Korea Advanced Institute of Science and Technology (KAIST), NAVER
- Country of organization
- South Korea
- Domain
- Image generation
- Task
- Image generation
- Training compute
- 2.2×10²⁰ FLOP
- Compute estimation method
- Hardware
- Dataset size
- 328M
- Training hardware
- NVIDIA A100 PCIe
- Training time
- 481 h
- Training cost (2023 USD)
- $338
- Numerical format
- FP16
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Citations
- 110
- Epoch confidence
- Confident
More from Korea Advanced Institute of Science and Technology (KAIST),NAVER
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