Skip to content

modl-org/modl-registry

Repository files navigation

modl-registry

Model registry for modl — the CLI model manager for AI image generation.

Website · Browse Models · CLI Docs · CLI Repo

This repository contains YAML manifest files that describe models, LoRAs, VAEs, text encoders, recipes, and other assets. The modl CLI fetches a compiled index from this registry to know what's available and how to download it.

Current Registry

Type Count Examples
Checkpoints 8 FLUX.1 Dev, FLUX.1 Schnell, SDXL Base, SD 1.5
Diffusion Models 17 FLUX Kontext, Z-Image Turbo, Qwen Image, WAN 2.2
LoRAs 4 LCM LoRA (SDXL/SD1.5), Z-Image Distill, Qwen Edit Lightning
VAEs 9 Flux VAE, SDXL VAE, SD VAE, WAN VAE
Text Encoders 10 T5-XXL, CLIP-L, Flux2 Mistral/Qwen3, UMT5-XXL
ControlNets 6 FLUX Depth, SD1.5 Canny/Depth/OpenPose, SDXL Canny
IP-Adapters 3 FLUX Redux, SDXL IP-Adapter, FaceID
Upscalers 3 4x UltraSharp, BSRGANx2, RealESRGAN
Segmentation 1 BiRefNet DIS
Recipes 4 Flux Face, Flux Style, Flux Product, Z-Image Face

Cloud-Available Models

These models are available on modl cloud for training and inference:

Model Training GPU Est. Time Inference GPU
FLUX.1 Schnell A100 80GB ~10 min A10G
FLUX.1 Dev A100 80GB ~25 min A100 80GB
Z-Image Turbo A100 80GB ~10 min A10G
Qwen Image A100 80GB ~15 min A100 80GB

Recipes

Recipes are training presets — shareable, version-controlled best practices for LoRA training. Instead of guessing parameters, pick a recipe:

modl train --recipe flux-face-lora --dataset ./photos
modl train --recipe flux-product-lora --dataset ./products

Each recipe specifies:

  • Base model and architecture
  • Training parameters — steps, learning rate, rank, optimizer, resolution
  • Generation defaults — steps, CFG, sampler for using the trained LoRA
  • Dataset guidance — min/max images, caption strategy
  • Tips — practical advice from real usage

Recipes live in manifests/recipes/ and are first-class registry items.

See CONTRIBUTING.md for how to submit your own recipes.

For Users

You don't interact with this repo directly. Just use the modl CLI:

modl update              # Fetch latest index from this registry
modl search flux         # Search available models
modl install flux-dev    # Install a model
modl info flux-dev       # View model details, cloud availability

For Contributors

Want to add a model or recipe? See CONTRIBUTING.md.

# Quick start
git clone https://github.com/modl-org/modl-registry
cd modl-registry
pip install pyyaml

# Add your manifest
cp manifests/checkpoints/flux-dev.yaml manifests/checkpoints/my-model.yaml
# Edit the file...

# Validate
python scripts/validate.py manifests/checkpoints/my-model.yaml
python scripts/build_index.py

Schema

All manifests are validated against manifest.schema.json.

License

MIT

About

Model registry for mods — YAML manifests for AI image generation models, LoRAs, VAEs, and more

Resources

License

Contributing

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages