matrix-pygmalion-bot/runpod/runpod-worker-oobabooga-api/Dockerfile

ARG DEV_IMAGE=nvidia/cuda:11.8.0-cudnn8-devel-ubuntu22.04
ARG BASE_IMAGE=nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04
#ARG BASE_IMAGE=nvidia/cuda:11.6.2-cudnn8-devel-ubuntu20.04
#ARG BASE_IMAGE=runpod/pytorch:3.10-2.0.0-117
#ARG BASE_IMAGE=nvcr.io/nvidia/pytorch:23.03-py3
FROM ${DEV_IMAGE} as builder

# https://github.com/runpod/containers/blob/main/oobabooga/Dockerfile

# DOCKER_BUILDKIT=1 docker build --build-arg MODEL_NAME="PygmalionAI/pygmalion-350m" -t magn418/runpod-oobabooga-pygmalion:test .
# docker builder prune

WORKDIR /
SHELL ["/bin/bash", "-o", "pipefail", "-c"]
ENV DEBIAN_FRONTEND noninteractive\
    SHELL=/bin/bash

RUN --mount=type=cache,target=/var/cache/apt,sharing=locked apt-get update --yes && \
    # - apt-get upgrade is run to patch known vulnerabilities in apt-get packages as
    #   the ubuntu base image is rebuilt too seldom sometimes (less than once a month)
    apt-get upgrade --yes && \
    apt install --yes --no-install-recommends \
      build-essential \
      cmake \
      ca-certificates \
      git \
      git-lfs \
      wget \
      curl \
      bash \
#      libgl1 \
      software-properties-common \
      openssh-server
#      apt-get clean && rm -rf /var/lib/apt/lists/* && \
#      echo "en_US.UTF-8 UTF-8" > /etc/locale.gen

#RUN apt-key del 7fa2af80 && \
#    apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/3bf863cc.pub
#
#RUN add-apt-repository ppa:deadsnakes/ppa && \
#    apt-get install python3.10 python3.10-dev python3.10-venv python3-pip -y --no-install-recommends && \
#    update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.10 1 && \
#    update-alternatives --install /usr/bin/python python /usr/bin/python3 1 && \
##    update-alternatives --install /usr/bin/pip pip /usr/bin/pip3 1 && \
#    update-alternatives --set python3 /usr/bin/python3.10 && \
#    update-alternatives --set python /usr/bin/python3 && \
#    apt-get clean && rm -rf /var/lib/apt/lists/*

RUN --mount=type=cache,target=/var/cache/apt,sharing=locked apt-get update --yes && \
    apt install --yes --no-install-recommends \
      python3 python3-dev python3-venv python3-pip
#    apt-get clean && rm -rf /var/lib/apt/lists/*

#ARG TORCH_CUDA_ARCH_LIST="3.5;5.0;6.0;6.1;7.0;7.5;8.0;8.6+PTX"
ARG TORCH_CUDA_ARCH_LIST="7.5;8.0;8.6+PTX"
ENV TORCH_CUDA_ARCH_LIST=${TORCH_CUDA_ARCH_LIST}
RUN --mount=type=cache,target=/root/.cache,sharing=locked pip3 install --upgrade pip setuptools && \
    pip3 install cuda-python==11.8.0 && \
    pip3 install torch torchvision torchaudio --extra-index-url=https://download.pytorch.org/whl/cu118 && \
    pip3 install bitsandbytes && \
    pip3 install safetensors && \
    pip3 install sentencepiece && \
    pip3 install diffusers && \
    pip3 install deepspeed && \
    pip3 install accelerate xformers triton && \
    pip3 install git+https://github.com/huggingface/transformers.git && \
    pip3 install huggingface-hub && \
    pip3 install runpod
#    pip3 cache purge

RUN mkdir -p /workspace
WORKDIR /workspace

#RUN mkdir /workspace && 
RUN --mount=type=cache,target=/root/.cache,sharing=locked \
    cd /workspace && git clone --depth 1 https://github.com/oobabooga/text-generation-webui.git && \
    cd /workspace/text-generation-webui && pip3 install -r requirements.txt && \
    cd extensions/api && pip3 install -r requirements.txt
#  pip3 cache purge

RUN --mount=type=cache,target=/root/.cache,sharing=locked \
    cd /workspace/text-generation-webui/ && mkdir repositories && cd repositories && \
    # Triton
    git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa.git && \
    git reset --hard 508de42ff45ec560a4504e12b0d42114d599cf38 && \
    (cd GPTQ-for-LLaMa && pip3 install -r requirements.txt && rm -rf .git)
    # Cuda
#    git clone --branch cuda --single-branch https://github.com/qwopqwop200/GPTQ-for-LLaMa.git && \
#    (cd GPTQ-for-LLaMa && python3 setup_cuda.py install && python3 setup_cuda.py bdist_wheel -d .)
    # Cuda (Oobabooga's fork)
#    git clone --branch cuda --single-branch --depth 1 https://github.com/oobabooga/GPTQ-for-LLaMa.git && \
#    (cd GPTQ-for-LLaMa && python3 setup_cuda.py bdist_wheel -d .)


FROM builder AS modeldownloader

ARG MODEL_NAME="PygmalionAI/pygmalion-350m"
ENV MODEL_NAME=${MODEL_NAME}

RUN cd /workspace/text-generation-webui && python3 download-model.py ${MODEL_NAME}
#RUN git lfs install && \
#    git clone --depth 1 https://huggingface.co/${MODEL_NAME}

#RUN wget -P /workspace/text-generation-webui/models/ https://raw.githubusercontent.com/BlinkDL/ChatRWKV/main/v2/20B_tokenizer.json && \
#    wget -P /workspace/text-generation-webui/models/ https://huggingface.co/BlinkDL/rwkv-4-raven/resolve/main/RWKV-4-Raven-7B-v9-Eng99%25-Other1%25-20230412-ctx8192.pth

    #https://huggingface.co/BlinkDL/rwkv-4-pile-7b/resolve/main/RWKV-4-Pile-7B-20230406-ctx8192-test949.pth

#RUN cd /workspace/text-generation-webui && python3 download-model.py MetaIX/GPT4-X-Alpaca-30B-Int4 --text-only && \
#    wget -P /workspace/text-generation-webui/models/MetaIX_GPT4-X-Alpaca-30B-Int4/ https://huggingface.co/MetaIX/GPT4-X-Alpaca-30B-Int4/resolve/main/gpt4-x-alpaca-30b-4bit.safetensors

FROM ${BASE_IMAGE}
#ENV TORCH_CUDA_ARCH_LIST="3.5;5.0;6.0;6.1;7.0;7.5;8.0;8.6+PTX"
RUN mkdir -p /workspace
WORKDIR /workspace

RUN --mount=type=cache,target=/var/cache/apt,sharing=locked apt-get update --yes && \
    apt install --yes --no-install-recommends \
      python3 python3-dev python3-venv python3-pip \
      make g++ \
      git
#    apt-get clean && rm -rf /var/lib/apt/lists/*

RUN --mount=type=cache,target=/root/.cache,sharing=locked pip3 install --upgrade pip setuptools && \
    pip3 install cuda-python==11.8.0 && \
    pip3 install torch torchvision torchaudio --extra-index-url=https://download.pytorch.org/whl/cu118 && \
    pip3 install bitsandbytes && \
    cp /usr/local/lib/python3.10/dist-packages/bitsandbytes/libbitsandbytes_cuda118.so /usr/local/lib/python3.10/dist-packages/bitsandbytes/libbitsandbytes_cpu.so && \
    pip3 install safetensors && \
    pip3 install sentencepiece && \
    pip3 install diffusers && \
    pip3 install deepspeed && \
    pip3 install accelerate xformers triton && \
    pip3 install git+https://github.com/huggingface/transformers.git && \
    pip3 install rwkv && \
    pip3 install huggingface-hub && \
    pip3 install runpod
#    pip3 cache purge

#RUN mkdir -p /workspace/text-generation-webui/repositories
COPY --from=builder /workspace/text-generation-webui /workspace/text-generation-webui
COPY --from=modeldownloader /workspace/text-generation-webui/models /workspace/text-generation-webui/models

RUN --mount=type=cache,target=/root/.cache,sharing=locked \
    cd /workspace/text-generation-webui && pip3 install -r requirements.txt && \
    (cd extensions/api && pip3 install -r requirements.txt) && \
    (cd repositories/GPTQ-for-LLaMa && pip3 install -r requirements.txt)
#    pip3 install /workspace/text-generation-webui/repositories/GPTQ-for-LLaMa/*.whl
#    pip3 cache purge

COPY runpod_infer.py /workspace/
COPY start.sh /
RUN chmod +x /start.sh

CMD [ "/start.sh" ]