pix2pixHD and pix2pix
About pix2pixHD
NVIDIA/pix2pixHD
Synthesizing and manipulating 2048x1024 images with conditional GANs
Uses a coarse-to-fine generator architecture with multi-scale discriminators to enable photorealistic semantic label-to-image translation, supporting interactive real-time editing via instance-aware conditioning. Built in PyTorch with support for Automatic Mixed Precision training and multi-GPU distributed learning, scaling from 512x256 to full 2048x1024 resolution depending on available VRAM.
About pix2pix
phillipi/pix2pix
Image-to-image translation with conditional adversarial nets
Implements a U-Net generator paired with a PatchGAN discriminator to learn bidirectional mappings between image domains, enabling tasks like semantic label-to-photo synthesis and edge-to-object generation. Built in Torch with CUDA support, it trains efficiently on modest datasets (e.g., 400 images in ~2 hours) and includes pre-trained models for facades, Cityscapes, shoes, and handbags. Provides dataset utilities for edge detection via HED and Lab colorspace preprocessing, plus evaluation scripts for semantic segmentation metrics on Cityscapes.
Scores updated daily from GitHub, PyPI, and npm data. How scores work