open-mmlab/mmgeneration

Health v2 · maintenance only

Adoption not part of the score

2012 stars · 230 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

MMGeneration is a PyTorch-based toolkit from OpenMMLab for training and applying generative models, with a focus on GANs and support for diffusion-based text-to-image models. It has been merged into MMEditing 1.x, which now hosts its ongoing development and new features.

Use cases

• train a GAN like StyleGAN2 or BigGAN on my own dataset
• generate images with pretrained generative models
• run text-to-image generation with stable diffusion or GLIDE
• experiment with unconditional and conditional image generation in PyTorch
• benchmark generative adversarial networks with FID and IS metrics
• interpolate and edit latent codes of a pretrained GAN

When to choose

• you want a modular, config-driven PyTorch framework for GAN research
• you need pretrained model zoos for GANs and diffusion models
• you are already using the OpenMMLab ecosystem (MMCV, MMEditing)

When to avoid

• you need active development or the newest models - use MMEditing 1.x instead
• you want a simple one-off inference script without a framework
• you work outside PyTorch or need non-image generative tasks

Facets

library · maturity maintenance

machine-learning deep-learning image-processing deep-learning machine-learning computer-vision image-processing python gan generative-models diffusion-models pytorch mmcv openmmlab text-to-image model-zoo linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="open-mmlab/mmgeneration")