a1111 — Stable Diffusion WebUI
The popular AUTOMATIC1111 / stable-diffusion-webui GUI for image generation. Browser-based control panel for txt2img, img2img, inpainting, extras, and a deep extension ecosystem.
Image tag
docker.io/manvarharsh/a1111:cuda12
What’s in this image
- Base:
nvidia/cuda:12.4.1-cudnn-devel-ubuntu22.04 - Python 3.10 (conda)
- PyTorch with CUDA 12 support
- AUTOMATIC1111/stable-diffusion-webui (cloned at build time)
- xformers for memory-efficient attention
- OpenSSH server
Default ports
| Port | Service |
|---|---|
| 22 | SSH |
| 7860 | A1111 WebUI |
Use cases
- Interactive Stable Diffusion image generation
- Trying SD 1.5 / SDXL / SD3 checkpoints with one-click UI
- Running A1111 extensions (ControlNet, ADetailer, Regional Prompter, etc.)
- Quick prompt iteration with built-in batch / X/Y/Z plots
Environment variables
| Variable | Description |
|---|---|
ENABLE_SSH | Enable SSH server (default false) |
ENABLE_A1111 | Start the WebUI on port 7860 |
A1111_EXTRA_ARGS | Pass extra CLI args (--xformers, --medvram, --api) |
SSH_PUBLIC_KEY | Public key to authorize for SSH |
Persistence
Mount your storage at /data. Place model checkpoints under /data/models/Stable-diffusion/, LoRAs under /data/models/Lora/, and outputs under /data/output/.