Run in docker

[DrMartiner]

No description provided.

[dstrop]

i will just leave this here

FROM nvidia/cuda:12.2.0-runtime-ubuntu22.04

RUN apt-get update && \
    apt-get install --no-install-recommends -y python3 python3-pip && \
    rm -rf /var/lib/apt/lists/*

RUN --mount=type=cache,target=/root/.cache/pip pip3 install virtualenv
RUN mkdir /app

WORKDIR /app

RUN virtualenv /venv
RUN . /venv/bin/activate && \
    pip3 install --upgrade pip

COPY requirements_versions.txt /app/requirements_versions.txt
RUN . /venv/bin/activate && \
    pip3 install -r requirements_versions.txt

COPY . /app/

ENTRYPOINT [ "bash", "-c", ". /venv/bin/activate && exec \"$@\"", "--" ]
CMD [ "python3", "launch.py", "--listen" ]

version: "3.3"
services:
  testui:
    container_name: testui
    build:
      context: .
    ports:
      - "7860:7860"
    stdin_open: true
    tty: true
    volumes:
      - ./:/app
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              device_ids: ['0']
              capabilities: [gpu]

[mabushey]

Thank you @dstrop , I was unable to get fooocus work via the Linux python directions, but this is working perfect.

[gusmolinabr]

It appears the instructions are wrong. I could not do it also, and I'd prefer not to run on docker.

[mabushey]

@gusmolinabr Docker is by far the easiest way to get anything python or java to work.

[yairopro]

I've found this
https://github.com/camenduru/Fooocus-docker

[brianetaveras]

I ran into some issues using the config posted by @dstrop - thank you for sharing!

Here is my adjusted Dockerfile.

FROM nvidia/cuda:12.3.0-runtime-ubuntu22.04

RUN apt-get update && \
    apt-get install --no-install-recommends -y python3 python3-pip libgl1-mesa-glx libglib2.0-0 libsm6 libxrender1 libxext6 && \
    rm -rf /var/lib/apt/lists/*

RUN --mount=type=cache,target=/root/.cache/pip pip3 install virtualenv
RUN mkdir /app

WORKDIR /app

RUN virtualenv /venv
RUN . /venv/bin/activate && \
    pip3 install --upgrade pip

COPY requirements_versions.txt /app/requirements_versions.txt
RUN . /venv/bin/activate && \
    pip3 install -r requirements_versions.txt

COPY . /app/

# may need to adjust the arch
RUN ln -s /usr/lib/x86_64-linux-gnu/mesa/libGL.so.1 /usr/lib/libGL.so.1
RUN ln -s /usr/lib/x86_64-linux-gnu/libgthread-2.0.so.0 /usr/lib/libgthread-2.0.so.0

ENTRYPOINT [ "bash", "-c", ". /venv/bin/activate && exec \"$@\"", "--" ]
CMD [ "python3", "launch.py", "--listen" ]

The project runs on port 7865 by default so people may want to adjust the docker-compose file as well

[salatfreak]

Stripping away the venv, adding the installation of torch and torchvision that is otherwise done when running launch.py for the first time, making the CUDA version configurable, and flagging the models directory as volume, I arrived at this:

ARG CUDA_VERSION=12.3.0

FROM docker.io/nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu22.04

RUN apt-get update && \
    apt-get install --no-install-recommends -y \
      libgl1-mesa-glx libglib2.0-0 libsm6 libxrender1 libxext6 \
      git python3 python3-pip && \
    rm -rf /var/lib/apt/lists/*

RUN git clone https://github.com/lllyasviel/Fooocus /app
WORKDIR /app

RUN pip install --no-cache -r requirements_versions.txt
RUN pip install --no-cache torch==2.1.0 torchvision==0.16.0 \
    --extra-index-url "https://download.pytorch.org/whl/cu121"

VOLUME /app/models

ENTRYPOINT [ "python3", "launch.py", "--listen" ]

On debian stable the latest CUDA image doesn't seem to work, so I use --build-arg CUDA_VERSION=12.0.1. It runs rootless via podman run --rm -v fooocus-models:/app/models -p 7865:7865 fooocus.

[mambari]

I tried with portainer on my windows.

(base) root@85a74394deb6:/Fooocus# conda activate fooocus
(fooocus) root@85a74394deb6:/Fooocus# python launch.py
[System ARGV] ['launch.py']
Python 3.10.12 (main, Jul 5 2023, 18:54:27) [GCC 11.2.0]
Fooocus version: 2.1.824
Running on local URL: http://127.0.0.1:7865

To create a public link, set share=True in launch().
Total VRAM 24564 MB, total RAM 15878 MB
Set vram state to: NORMAL_VRAM
Disabling smart memory management
Device: cuda:0 NVIDIA GeForce RTX 4090 : native
VAE dtype: torch.bfloat16
Using pytorch cross attention
Refiner unloaded.
model_type EPS
adm 2816
Using pytorch attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using pytorch attention in VAE
extra keys {'cond_stage_model.clip_l.text_projection', 'cond_stage_model.clip_g.transformer.text_model.embeddings.position_ids', 'cond_stage_model.clip_l.logit_scale'}
Base model loaded: /Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors
Request to load LoRAs [['sd_xl_offset_example-lora_1.0.safetensors', 0.1], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0]] for model [/Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors].
Loaded LoRA [/Fooocus/models/loras/sd_xl_offset_example-lora_1.0.safetensors] for UNet [/Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors] with 788 keys at weight 0.1.
Fooocus V2 Expansion: Vocab with 642 words.
Fooocus Expansion engine loaded for cuda:0, use_fp16 = True.
Requested to load SDXLClipModel
Requested to load GPT2LMHeadModel
Loading 2 new models
[Fooocus Model Management] Moving model(s) has taken 0.56 seconds
App started successful. Use the app with http://127.0.0.1:7865/ or 127.0.0.1:7865

And I have

But I can not test it:

[dgsiegel]

For AMD GPUs I've had success with the following Docker/Podman file:

FROM ubuntu:22.04

RUN apt-get update && \
    apt-get install --no-install-recommends -y \
      libgl1-mesa-glx libglib2.0-0 libsm6 libxrender1 libxext6 \
      wget aria2 \
      git python3 python3-pip && \
    rm -rf /var/lib/apt/lists/*

RUN git clone https://github.com/lllyasviel/Fooocus /app
WORKDIR /app

RUN pip install --no-cache -r requirements_versions.txt
RUN pip uninstall -y torch torchvision torchaudio torchtext functorch xformers
RUN pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm5.7

VOLUME /app/models

ENV HSA_OVERRIDE_GFX_VERSION="10.3.0"

ENTRYPOINT [ "python3", "launch.py", "--listen" ]

Build it with docker build -t fooocus ., then run it with docker run -it --rm -v fooocus-models:/app/models --device=/dev/kfd --device=/dev/dri -p 7865:7865 fooocus

[mashb1t]

Feel free to also contribute your suggestions to the PR #1418.

[realies]

[mashb1t]

I can relate... This is a feature in Phase 2 - Features now, starting to work on merging PRs tomorrow for Phase 1 - Bugfixes (now finally permitted to do so!)
See #2154