AI models
DDPM-IP (CelebA)
Training compute
3.5×10²⁰ FLOP
Parameters
295M
Published
Jan 27, 2023
DDPM-IP (CelebA) is an AI model developed by Utrecht University (Netherlands), first published in January 2023. It works in the image generation domain, on tasks such as image generation and text-to-image.
Training it took an estimated 3.5×10²⁰ FLOP of compute (estimation method: hardware). The model has 295,000,000 parameters. It was trained on roughly 831.5M datapoints. Training ran on NVIDIA V100 for about 120 hours. The compute alone is estimated at $390 in 2023 dollars.
Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 102 citations. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- Utrecht University
- Country of organization
- Netherlands
- Domain
- Image generation
- Task
- Image generation, Text-to-image
- Training compute
- 3.5×10²⁰ FLOP
- Compute estimation method
- Hardware
- Parameters
- 295,000,000
- Dataset size
- 831.5M
- Training hardware
- NVIDIA V100
- Training time
- 120 h
- Training cost (2023 USD)
- $390
- Numerical format
- FP16
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Citations
- 102
- 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