# nvidia/pix2pixhd

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/nvidia-pix2pixhd).**

6,920 stars · 1,411 forks · Python · NOASSERTION

## Links

- GitHub: https://github.com/NVIDIA/pix2pixHD
- Homepage: https://tcwang0509.github.io/pix2pixHD/
- awesome-repositories: https://awesome-repositories.com/repository/nvidia-pix2pixhd.md

## Description

pix2pixHD is a conditional generative adversarial network designed to transform semantic label maps into high-resolution photorealistic images. It functions as a high-resolution image synthesizer and an image-to-image translation model capable of producing synthetic images at 2048x1024 resolution.

The system includes a semantic image editor that allows for the modification of high-resolution visuals by updating the underlying semantic label maps. This enables interactive image editing and the generation of photorealistic images based on source images or discrete label maps.

The framework provides tools for image translation model training using custom datasets. It incorporates training acceleration through automatic mixed precision and multi-GPU data parallelism to manage high-resolution tensors.

## Tags

### Artificial Intelligence & ML

- [Conditional Image Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation-models/conditional-image-generation.md) — Functions as a conditional generative adversarial network that synthesizes images based on semantic label maps.
- [Image-to-Image Translation](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-ai-resources/diffusion-visual-models/generative-ai-pipelines/text-to-image-generators/image-inpainting/image-to-image-translation.md) — Enables image-to-image translation by mapping semantic label maps to photorealistic images.
- [Conditional Training](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-model-training-tools/conditional-training.md) — Provides a framework for training conditional GANs optimized for high-fidelity image generation.
- [High-Resolution Synthesis](https://awesome-repositories.com/f/artificial-intelligence-ml/image-super-resolution-models/high-resolution-synthesis.md) — Synthesizes high-resolution photorealistic images up to 2048x1024 resolution.
- [Semantic Synthesis Frameworks](https://awesome-repositories.com/f/artificial-intelligence-ml/image-super-resolution-models/semantic-synthesis-frameworks.md) — Transforms semantic label maps into photorealistic, high-resolution imagery via conditional GANs. ([source](https://github.com/nvidia/pix2pixhd#readme))
- [Image Translation Training](https://awesome-repositories.com/f/artificial-intelligence-ml/image-translation-training.md) — Provides a framework for building image translation models using custom datasets and label maps. ([source](https://github.com/nvidia/pix2pixhd#readme))
- [Data-Parallel Training](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-training-frameworks/data-parallel-training.md) — Distributes training workloads across multiple GPUs using data parallelism to handle high-resolution tensors.
- [Hierarchical Image Generators](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-ai-resources/diffusion-visual-models/generative-ai-pipelines/text-to-image-generators/hierarchical-image-generators.md) — Uses a two-stage generator hierarchy to synthesize images from low to high resolution.
- [Semantic Editing](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation/image-editing/semantic-editing.md) — Allows for the modification of synthetic visuals by updating the underlying semantic label maps.
- [GPU Training Accelerators](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/distributed-and-accelerated-compute/training-acceleration-tools/gpu-training-accelerators.md) — Accelerates training using multi-GPU data parallelism and automatic mixed precision. ([source](https://github.com/nvidia/pix2pixhd#readme))
- [Mixed Precision Training](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/distributed-and-accelerated-compute/training-acceleration-tools/mixed-precision-training.md) — Implements mixed precision training using float16 and float32 formats to optimize memory and speed.

### Part of an Awesome List

- [Semantic Image Synthesis](https://awesome-repositories.com/f/awesome-lists/ai/image-generation-and-synthesis/semantic-image-synthesis.md) — Generates photorealistic images from semantic segmentation label maps. ([source](https://github.com/nvidia/pix2pixhd#readme))
- [Foundational Generative Models](https://awesome-repositories.com/f/awesome-lists/ai/foundational-generative-models.md) — High-resolution image synthesis and manipulation framework.

### Data & Databases

- [Semantic-to-Visual Mappings](https://awesome-repositories.com/f/data-databases/enum-definitions/enum-label-mappings/category-identifier-mappings/semantic-to-visual-mappings.md) — Translates discrete semantic category IDs into photorealistic visual textures and shapes.
- [Multi-Scale Discriminators](https://awesome-repositories.com/f/data-databases/model-as-a-table-integrations/discriminator-networks/multi-scale-discriminators.md) — Employs an array of discriminators at different scales to ensure both high-frequency detail and global consistency.