Dockerfile

ARG CUDA_IMAGE="12.4.0-runtime-ubuntu22.04"
FROM nvidia/cuda:${CUDA_IMAGE}

RUN <<EOF
apt update
apt install --no-install-recommends -y \
    build-essential \
    git
apt clean && rm -rf /var/lib/apt/lists/*
EOF

# 安装Node.js和npm
RUN curl -fsSL https://deb.nodesource.com/setup_lts.x | bash - && \
    apt-get install -y nodejs

# 安装pnpm
RUN npm install -g pnpm

# 克隆前端仓库并构建
RUN git clone https://github.com/SolomonLeon/web-rwkv-realweb.git /frontend
WORKDIR /frontend
RUN pnpm install && pnpm run build

COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/

COPY . .

RUN useradd -m -u 1000 user
USER user

ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH

WORKDIR $HOME/app

COPY --chown=user . $HOME/app
COPY --chown=user /frontend/dist $HOME/app/dist-frontend

RUN << EOF
cat > $HOME/app/config.local.yaml<<EOL
HOST: "0.0.0.0"
PORT: 7860
STRATEGY: "cuda fp16"
RWKV_CUDA_ON: False
CHUNK_LEN: 256
MODELS:
  - SERVICE_NAME: "RWKV-x070-World-0.1B-v2.8-20241210-ctx4096"
    DOWNLOAD_MODEL_FILE_NAME: "RWKV-x070-World-0.1B-v2.8-20241210-ctx4096.pth"
    DOWNLOAD_MODEL_REPO_ID: "BlinkDL/rwkv-7-world"
    DOWNLOAD_MODEL_DIR: "./"
    REASONING: False
    DEFAULT: True
    DEFAULT_SAMPLER:
      max_tokens: 512
      temperature: 1.0
      top_p: 0.3
      presence_penalty: 0.5
      count_penalty: 0.5
      penalty_decay: 0.996
      stop:
        - "\n\n"
  - SERVICE_NAME: "RWKV7-G1-0.1B-68%trained-20250303-ctx4k"
    DOWNLOAD_MODEL_FILE_NAME: "RWKV7-G1-0.1B-68%trained-20250303-ctx4k.pth"
    DOWNLOAD_MODEL_REPO_ID: "BlinkDL/temp-latest-training-models"
    DOWNLOAD_MODEL_DIR: "./"
    REASONING: True
    DEFAULT: True
    DEFAULT_SAMPLER:
      max_tokens: 4096
      temperature: 1.0
      top_p: 0.3
      presence_penalty: 0.5
      count_penalty: 0.5
      penalty_decay: 0.996
      stop:
        - "\n\n"
EOL
EOF

RUN uv sync --frozen --extra cu124

CMD ["uv","run","app.py",]