From 1c972bff8da2ba1e39d4bc488d0c5bfa81e2fb47 Mon Sep 17 00:00:00 2001
From: MengqingCao <cmq0113@163.com>
Date: Mon, 24 Jun 2024 03:40:44 +0000
Subject: [PATCH] add docker for llamafactory

---
 sources/llamafactory/Dockerfile          | 37 +++++++++++++++++
 sources/llamafactory/docker-compose.yaml | 31 ++++++++++++++
 sources/llamafactory/install.rst         | 52 ++++++++++++++++++++++--
 3 files changed, 116 insertions(+), 4 deletions(-)
 create mode 100644 sources/llamafactory/Dockerfile
 create mode 100644 sources/llamafactory/docker-compose.yaml

diff --git a/sources/llamafactory/Dockerfile b/sources/llamafactory/Dockerfile
new file mode 100644
index 0000000..955256c
--- /dev/null
+++ b/sources/llamafactory/Dockerfile
@@ -0,0 +1,37 @@
+# Using ubuntu 22.04 images with cann 8.0.rc1
+# More options can be found at https://hub.docker.com/r/cosdt/cann/tags
+FROM cosdt/cann:8.0.rc1-910b-ubuntu22.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+# Define installation arguments
+ARG INSTALL_DEEPSPEED=false
+ARG PIP_INDEX=https://pypi.org/simple
+
+# Set the working directory
+WORKDIR /app/LLaMA-Factory
+
+RUN cd /app && \
+    git config --global http.version HTTP/1.1 && \
+    git clone https://github.com/hiyouga/LLaMA-Factory.git && \
+    cd /app/LLaMA-Factory
+
+RUN pip config set global.index-url $PIP_INDEX
+RUN python3 -m pip install --upgrade pip
+
+# Install the LLaMA Factory
+RUN EXTRA_PACKAGES="torch-npu,metrics"; \
+    if [ "$INSTALL_DEEPSPEED" = "true" ]; then \
+        EXTRA_PACKAGES="${EXTRA_PACKAGES},deepspeed"; \
+    fi; \
+    pip install -e .[$EXTRA_PACKAGES] && \
+    pip uninstall -y transformer-engine flash-attn
+
+# Set up volumes
+VOLUME [ "/root/.cache/huggingface/", "/app/data", "/app/output" ]
+
+# Expose port 7860 for the LLaMA Board
+EXPOSE 7860
+
+# Expose port 8000 for the API service
+EXPOSE 8000
diff --git a/sources/llamafactory/docker-compose.yaml b/sources/llamafactory/docker-compose.yaml
new file mode 100644
index 0000000..93eb671
--- /dev/null
+++ b/sources/llamafactory/docker-compose.yaml
@@ -0,0 +1,31 @@
+services:
+  llamafactory:
+    build:
+      dockerfile: Dockerfile
+      context: .
+      args:
+        INSTALL_DEEPSPEED: false
+        PIP_INDEX: https://pypi.org/simple
+    container_name: llamafactory
+    volumes:
+      - ../../hf_cache:/root/.cache/huggingface/
+      - ../../data:/app/LLaMA-Factory/data
+      - ../../output:/app/LLaMA-Factory/output
+      - /usr/local/dcmi:/usr/local/dcmi
+      - /usr/local/bin/npu-smi:/usr/local/bin/npu-smi
+      - /usr/local/Ascend/driver/lib64:/usr/local/Ascend/driver/lib64
+      - /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info
+      - /etc/ascend_install.info:/etc/ascend_install.info
+    ports:
+      - "7860:7860"
+      - "8000:8000"
+    ipc: host
+    tty: true
+    stdin_open: true
+    command: bash
+    devices:
+      - /dev/davinci0
+      - /dev/davinci_manager
+      - /dev/devmm_svm
+      - /dev/hisi_hdc
+    restart: unless-stopped
diff --git a/sources/llamafactory/install.rst b/sources/llamafactory/install.rst
index 3c63e33..75a270a 100644
--- a/sources/llamafactory/install.rst
+++ b/sources/llamafactory/install.rst
@@ -1,11 +1,16 @@
 安装指南
 ==============
 
-本教程面向使用 LLAMA-Factory & 昇腾的开发者，帮助完成昇腾环境下 LLaMA-Factory 的安装。
+本教程面向使用 LLAMA-Factory & 昇腾的开发者，帮助完成昇腾环境下 LLaMA-Factory 的安装。此处提供 :ref:`install` 和 :ref:`docker` 两种安装方式。
 
-昇腾环境安装
+.. _install:
+
+直接安装
 ------------
 
+昇腾环境安装
+~~~~~~~~~~~~~~~
+
 请根据已有昇腾产品型号及CPU架构等按照 :doc:`快速安装昇腾环境指引 <../ascend/quick_install>` 进行昇腾环境安装，或使用已安装好昇腾环境及 LLaMA-Factory 的 docker 镜像：
 
 - TODO
@@ -16,7 +21,7 @@
   安装 CANN 时，请同时安装 Kernel 算子包。
 
 Python 环境创建
-----------------------
+~~~~~~~~~~~~~~~
 
 .. note::
   如果你已经选择使用上述 docker 镜像，可忽略此步骤，直接开始使用 LLaMA-Factory。
@@ -31,7 +36,7 @@ Python 环境创建
 
 
 LLaMA-Factory 安装
-----------------------
+~~~~~~~~~~~~~~~~~~~~~~~~
 
 使用以下指令安装带有 torch-npu 的 LLaMA-Factory：
 
@@ -40,6 +45,45 @@ LLaMA-Factory 安装
 
     pip install -e .[torch_npu,metrics]
 
+.. _docker:
+
+使用 Docker
+------------
+
+进入 docker-npu 目录：
+
+.. code-block:: shell
+  
+  cd docker/docker-npu
+
+该目录下为 Dockerfile 及 docker compose 配置文件：
+
+Dockerfile:
+
+.. literalinclude:: ./Dockerfile
+    :language: docker
+    :linenos:
+
+docker-compose.yaml:
+
+.. literalinclude:: ./docker-compose.yaml
+    :language: yaml
+    :linenos:
+
+.. note::
+
+  默认镜像为 `cosdt/cann:8.0.rc1-910b-ubuntu22.04 <https://hub.docker.com/layers/cosdt/cann/8.0.rc1-910b-ubuntu22.04/images/sha256-29ef8aacf6b2babd292f06f00b9190c212e7c79a947411e213135e4d41a178a9?context=explore>`_。
+  更多选择见 `cosdt/cann <https://hub.docker.com/r/cosdt/cann/tags>`_.
+
+使用以下指令构建及启动 docker 容器：
+
+.. code-block:: shell
+  
+  docker build -f ./Dockerfile \
+      --build-arg INSTALL_DEEPSPEED=false \
+      --build-arg PIP_INDEX=https://pypi.org/simple \
+      -t llamafactory:latest 
+
 安装校验
 ----------------------