Table of contents

sd-forge — SD WebUI Forge

A1111-compatible fork by lllyasviel focused on faster inference and lower VRAM use, with native support for SVD, Flux, and modern attention kernels.

Image tag

docker.io/manvarharsh/sd-forge: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
  • lllyasviel/stable-diffusion-webui-forge
  • xformers
  • OpenSSH server

Default ports

PortService
22SSH
7860Forge WebUI

Use cases

  • A1111 workflows on lower-VRAM GPUs (e.g. RTX 3060, T4)
  • Faster image generation than vanilla A1111
  • Flux.1 schnell / dev image generation
  • Stable Video Diffusion (SVD) experiments

Environment variables

VariableDescription
ENABLE_SSHEnable SSH server
ENABLE_FORGEStart the WebUI on port 7860
FORGE_EXTRA_ARGSExtra CLI args (--xformers, --cuda-malloc)
SSH_PUBLIC_KEYPublic key for SSH

Persistence

Mount at /data. Forge follows A1111’s directory conventions — checkpoints under /data/models/Stable-diffusion/, LoRAs under /data/models/Lora/, outputs under /data/output/.

See also