Get a weekly rundown of the latest AI models and research... subscribe!

Pix2pix Zero


AI model preview image
The pix2pix-zero model is an image-to-image translation model that allows for zero-shot translation. It can generate outputs for unseen input-output pairs without retraining the model. The model achieves this through latent space interpolation and reconstruction. It is trained using a variational autoencoder (VAE) architecture and a GAN objective function. This enables it to learn a shared latent space representation for the input and output domains. The model can be used to perform various image translation tasks, such as converting sketches to color images or converting aerial images to maps, without needing to train on specific input-output pairs.

Use cases

The pix2pix-zero model has several potential use cases for technical audiences. One use case could be in the field of computer graphics, where the model could be used to generate realistic images based on rough sketches or low-resolution input. This could be particularly useful in game development, where artists could quickly iterate on designs by sketching out rough ideas and using the model to generate high-quality images. Another use case could be in the field of urban planning and architecture, where the model could be used to generate realistic maps based on aerial images. This could aid in the visualization of proposed architectural designs or in the creation of detailed urban planning models. In addition, the model could be used in the field of image editing and enhancement, allowing users to easily transform images into a desired style or apply artistic filters without needing to train the model on specific styles or filters. Overall, the pix2pix-zero model has the potential to simplify and streamline various image translation tasks, opening up new possibilities in computer graphics, urban planning, and image editing.



Cost per run
Avg run time
Nvidia A100 (40GB) GPU

Creator Models

Dreambooth Pikachu$0.08195513
Night Enhancement$0.0104538,658

Similar Models

Try it!

You can use this area to play around with demo applications that incorporate the Pix2pix Zero model. These demos are maintained and hosted externally by third-party creators. If you see an error, message me on Twitter.

Currently, there are no demos available for this model.


Summary of this model and related resources.

Model NamePix2pix Zero
Zero-shot Image-to-Image Translation
Model LinkView on Replicate
API SpecView on Replicate
Github LinkView on Github
Paper LinkNo paper link provided


How popular is this model, by number of runs? How popular is the creator, by the sum of all their runs?

Model Rank
Creator Rank


How much does it cost to run this model? How long, on average, does it take to complete a run?

Cost per Run$-
Prediction HardwareNvidia A100 (40GB) GPU
Average Completion Time-