techbookfest-16

docker-compose.yml

version: "3"

services:
  app:
    build:
      context: .
      dockerfile: Dockerfile
    image: kohya:latest
    stop_grace_period: 0s
    ipc: host
    tty: true
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              device_ids: ['1']
              capabilities: [gpu]
    volumes:
      - ./:/app
      - ./cache:/root/.cache
      - /home/thr3a/lora-trains:/data
    # command: sleep infinity
    # command: python tools/resize_images_to_resolution.py --max_resolution 1024x1024 --save_as_png --interpolation lanczos4 --copy_associated_files /data/ebifly/orig /data/ebifly/train/
    command: accelerate launch --num_cpu_threads_per_process 8 sdxl_train_network.py --config_file=/data/ebifly/config.toml --dataset_config /data/ebifly/database.toml

Dockerfile

# sha256:21196d81f56b48dbee70494d5f10322e1a77cc47ffe202a3bf68eab81533c20f
FROM --platform=linux/x86_64 nvcr.io/nvidia/cuda:12.1.1-cudnn8-devel-ubuntu22.04

# マウント時にパーミッションエラーを避けるためにホスト側のUID,GIDに合わせてください
# IDはidコマンドで確認できます。 例えばuid=2000(thr3a) gid=2000(thr3a) …と表示された場合2箇所を2000に変更します。
ARG UID=1000
ARG GID=1000
ARG PYTHON_VERSION=3.11
ARG PACKAGES="git curl ca-certificates vim wget unzip build-essential cmake jq libopencv-dev"

ENV DEBIAN_FRONTEND=noninteractive
ENV TZ=Asia/Tokyo
ENV PYTHONUNBUFFERED=1
ENV PIP_NO_CACHE_DIR=on
ENV PYTHONDONTWRITEBYTECODE=1

RUN apt-key adv --keyserver keyserver.ubuntu.com --recv f23c5a6cf475977595c89f51ba6932366a755776 \
 && echo "deb http://ppa.launchpadcontent.net/deadsnakes/ppa/ubuntu jammy main" > /etc/apt/sources.list.d/python.list \
 && echo "deb-src http://ppa.launchpad.net/deadsnakes/ppa/ubuntu jammy main" >> /etc/apt/sources.list.d/python.list

RUN apt-get update \
 && apt-get install -y --no-install-recommends ${PACKAGES} python${PYTHON_VERSION} \
 && ln -nfs /usr/bin/python${PYTHON_VERSION} /usr/bin/python \
 && ln -nfs /usr/bin/python${PYTHON_VERSION} /usr/bin/python3 \
 && rm -rf /var/lib/apt/lists/* \
 && curl -sSL https://bootstrap.pypa.io/get-pip.py | python -

RUN groupadd -g $UID user && \
    useradd -m -s /bin/bash -u $UID -g $GID user

WORKDIR /app
ADD ./setup.py ./
ADD ./requirements.txt ./
RUN pip install -U torch torchvision xformers==0.0.24 --index-url https://download.pytorch.org/whl/cu121
RUN pip install -U pip && pip install -r requirements.txt

USER user