Projected GAN
No estimate
No hardware requirements for this model
This model's weights are open, but no parameter count has been published for it. Every memory and speed figure starts from that number, so we would rather show nothing than a fabricated estimate.
On record
Full specification
Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.
Origin
Who built this model, where, and when it was published.
- Organisation
- Heidelberg University
- Organisation type
- Academia
- Country
- Germany
- Published
- 1 November 2021
- Authors
- Axel Sauer, Kashyap Chitta, Jens Müller, Andreas Geiger
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Image generation
- Task
- Image generation
Size
How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.
- Training data
- 3,000,000 tokens
They experiment with 22 image datasets. Largest appears to be LSUN-Bedroom at 3M images.
Training compute
The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.
- Training compute
- 1.1 × 10¹⁹ FLOP
- How it was established
- Hardware
"With this setting, each experiment takes roughly 100-200 GPU hours on a NVIDIA V100, for more details we refer to the appendix." "We conduct our experiments on an internal cluster with several nodes, each with up to 8 Quadro RTX 6000 or NVIDIA V100 using PyTorch 1.7.1 and CUDA 11.0." In appendix table 7, takes 10.1 seconds per 1k images on 8 Quadro RTX 6000s. Longest training run for Projected GAN appears to be in Figure 4 (left), at 14M images, though this is overtrained and the largest chec…
The training run
What it physically took to train: which chips, how many, for how long, and what that drew from the wall.
- Training hardware
- NVIDIA V100,NVIDIA Quadro RTX 6000
Availability
Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.
- Weights
- Open — downloadable
- Model access
- Open weights (unrestricted)
- Training code
- Open source
MIT license https://github.com/autonomousvision/projected-gan
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Why it is tracked
- SOTA improvement
- Record confidence
- Confident
- Citations
- 280
Table 3 "It is further compatible with resolutions of up to one Megapixel and advances the state-of-the-art Fréchet Inception Distance (FID) on twenty-two benchmark datasets"
Sources
Where this record came from and when it was last checked.
- Reference
- Projected GANs Converge Faster
- Last updated
- 1 January 2026
What the numbers mean
Background
Projected GAN was published by Heidelberg University, in Germany, in November 2021. The organisation is categorised as academia.
It works in Image generation, and is recorded as doing image generation.
The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.
Training and provenance
Producing it required around 1.1 × 10¹⁹ FLOP of arithmetic, on NVIDIA V100,NVIDIA Quadro RTX 6000, which is a statement about the training budget rather than about inference.
The training set ran to roughly 3,000,000 tokens.
Its inclusion criterion is sOTA improvement.
Answers
Projected GAN — common questions
Who created Projected GAN?
Projected GAN was published by Heidelberg University, based in Germany, categorised as academia.
When was Projected GAN released?
Projected GAN was published in November 2021. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is Projected GAN used for?
Projected GAN works in Image generation, and is recorded as handling image generation. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.
Where can I download Projected GAN?
The weights for Projected GAN are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.
How much compute was used to train Projected GAN?
Around 1.1 × 10¹⁹ FLOP, on NVIDIA V100,NVIDIA Quadro RTX 6000. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.
What GPU do I need to run Projected GAN?
We cannot say. Projected GAN has open weights, but no parameter count has been published for it, and every memory and speed calculation starts from that number. We would rather show nothing than a fabricated estimate.
Is Projected GAN open source?
Its weights are published, so Projected GAN can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.
How many parameters does Projected GAN have?
No parameter count has been published for Projected GAN, which is why no memory or speed figure appears on this page.
The other direction
Looking at it from the other side?
This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.