Merge pull request #4355 from MengqingCao/npu

Add docker-npu

README.md

@@ -383,11 +383,6 @@ source /usr/local/Ascend/ascend-toolkit/set_env.sh
 | torch-npu    | 2.1.0   | 2.1.0.post3 |
 | deepspeed    | 0.13.2  | 0.13.2      |
 
-Docker image:
-
-- 32GB: [Download page](http://mirrors.cn-central-221.ovaijisuan.com/detail/130.html)
-- 64GB: [Download page](http://mirrors.cn-central-221.ovaijisuan.com/detail/131.html)
-
 Remember to use `ASCEND_RT_VISIBLE_DEVICES` instead of `CUDA_VISIBLE_DEVICES` to specify the device to use.
 
 If you cannot infer model on NPU devices, try setting `do_sample: false` in the configurations.
@@ -424,17 +419,33 @@ llamafactory-cli webui
 
 ### Build Docker
 
-#### Use Docker
+For CUDA users:
 
 ```bash
-docker build -f ./Dockerfile \
+docker-compose -f ./docker/docker-cuda/docker-compose.yml up -d
+docker-compose exec llamafactory bash
+```
+
+For Ascend NPU users:
+
+```bash
+docker-compose -f ./docker/docker-npu/docker-compose.yml up -d
+docker-compose exec llamafactory bash
+```
+
+<details><summary>Build without Docker Compose</summary>
+
+For CUDA users:
+
+```bash
+docker build -f ./docker/docker-cuda/Dockerfile \
     --build-arg INSTALL_BNB=false \
     --build-arg INSTALL_VLLM=false \
     --build-arg INSTALL_DEEPSPEED=false \
     --build-arg PIP_INDEX=https://pypi.org/simple \
     -t llamafactory:latest .
 
-docker run -it --gpus=all \
+docker run -dit --gpus=all \
     -v ./hf_cache:/root/.cache/huggingface/ \
     -v ./data:/app/data \
     -v ./output:/app/output \
@@ -443,15 +454,43 @@ docker run -it --gpus=all \
     --shm-size 16G \
     --name llamafactory \
     llamafactory:latest
+
+docker exec -it llamafactory bash
 ```
 
-#### Use Docker Compose
+For Ascend NPU users:
 
 ```bash
-docker-compose up -d
-docker-compose exec llamafactory bash
+# Change docker image upon your environment
+docker build -f ./docker/docker-npu/Dockerfile \
+    --build-arg INSTALL_DEEPSPEED=false \
+    --build-arg PIP_INDEX=https://pypi.org/simple \
+    -t llamafactory:latest .
+
+# Change `device` upon your resources
+docker run -dit \
+    -v ./hf_cache:/root/.cache/huggingface/ \
+    -v ./data:/app/data \
+    -v ./output:/app/output \
+    -v /usr/local/dcmi:/usr/local/dcmi \
+    -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
+    -v /usr/local/Ascend/driver:/usr/local/Ascend/driver \
+    -v /etc/ascend_install.info:/etc/ascend_install.info \
+    -p 7860:7860 \
+    -p 8000:8000 \
+    --device /dev/davinci0 \
+    --device /dev/davinci_manager \
+    --device /dev/devmm_svm \
+    --device /dev/hisi_hdc \
+    --shm-size 16G \
+    --name llamafactory \
+    llamafactory:latest
+
+docker exec -it llamafactory bash
 ```
 
+</details>
+
 <details><summary>Details about volume</summary>
 
 - hf_cache: Utilize Hugging Face cache on the host machine. Reassignable if a cache already exists in a different directory.

README_zh.md

@@ -360,7 +360,7 @@ pip install https://github.com/jllllll/bitsandbytes-windows-webui/releases/downl
 
 <details><summary>昇腾 NPU 用户指南</summary>
 
-在昇腾 NPU 设备上安装 LLaMA Factory 时，需要指定额外依赖项，使用 `pip install -e '.[torch-npu,metrics]'` 命令安装。此外，还需要安装 **[Ascend CANN Toolkit and Kernels](https://www.hiascend.com/developer/download/community/result?module=cann)**，安装方法请参考[安装教程](https://www.hiascend.com/document/detail/zh/CANNCommunityEdition/80RC2alpha002/quickstart/quickstart/quickstart_18_0004.html)或使用以下命令：
+在昇腾 NPU 设备上安装 LLaMA Factory 时，需要指定额外依赖项，使用 `pip install -e ".[torch-npu,metrics]"` 命令安装。此外，还需要安装 **[Ascend CANN Toolkit and Kernels](https://www.hiascend.com/developer/download/community/result?module=cann)**，安装方法请参考[安装教程](https://www.hiascend.com/document/detail/zh/CANNCommunityEdition/80RC2alpha002/quickstart/quickstart/quickstart_18_0004.html)或使用以下命令：
 
 ```bash
 # 请替换 URL 为 CANN 版本和设备型号对应的 URL
@@ -383,11 +383,6 @@ source /usr/local/Ascend/ascend-toolkit/set_env.sh
 | torch-npu    | 2.1.0   | 2.1.0.post3 |
 | deepspeed    | 0.13.2  | 0.13.2      |
 
-Docker 镜像：
-
-- 32GB：[下载地址](http://mirrors.cn-central-221.ovaijisuan.com/detail/130.html)
-- 64GB：[下载地址](http://mirrors.cn-central-221.ovaijisuan.com/detail/131.html)
-
 请使用 `ASCEND_RT_VISIBLE_DEVICES` 而非 `CUDA_VISIBLE_DEVICES` 来指定运算设备。
 
 如果遇到无法正常推理的情况，请尝试设置 `do_sample: false`。
@@ -424,17 +419,33 @@ llamafactory-cli webui
 
 ### 构建 Docker
 
-#### 使用 Docker
+CUDA 用户：
 
 ```bash
-docker build -f ./Dockerfile \
+docker-compose -f ./docker/docker-cuda/docker-compose.yml up -d
+docker-compose exec llamafactory bash
+```
+
+昇腾 NPU 用户：
+
+```bash
+docker-compose -f ./docker/docker-npu/docker-compose.yml up -d
+docker-compose exec llamafactory bash
+```
+
+<details><summary>不使用 Docker Compose 构建</summary>
+
+CUDA 用户：
+
+```bash
+docker build -f ./docker/docker-cuda/Dockerfile \
     --build-arg INSTALL_BNB=false \
     --build-arg INSTALL_VLLM=false \
     --build-arg INSTALL_DEEPSPEED=false \
     --build-arg PIP_INDEX=https://pypi.org/simple \
     -t llamafactory:latest .
 
-docker run -it --gpus=all \
+docker run -dit --gpus=all \
     -v ./hf_cache:/root/.cache/huggingface/ \
     -v ./data:/app/data \
     -v ./output:/app/output \
@@ -443,15 +454,43 @@ docker run -it --gpus=all \
     --shm-size 16G \
     --name llamafactory \
     llamafactory:latest
+
+docker exec -it llamafactory bash
 ```
 
-#### 使用 Docker Compose
+昇腾 NPU 用户：
 
 ```bash
-docker-compose up -d
-docker-compose exec llamafactory bash
+# 根据您的环境选择镜像
+docker build -f ./docker/docker-npu/Dockerfile \
+    --build-arg INSTALL_DEEPSPEED=false \
+    --build-arg PIP_INDEX=https://pypi.org/simple \
+    -t llamafactory:latest .
+
+# 根据您的资源更改 `device`
+docker run -dit \
+    -v ./hf_cache:/root/.cache/huggingface/ \
+    -v ./data:/app/data \
+    -v ./output:/app/output \
+    -v /usr/local/dcmi:/usr/local/dcmi \
+    -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
+    -v /usr/local/Ascend/driver:/usr/local/Ascend/driver \
+    -v /etc/ascend_install.info:/etc/ascend_install.info \
+    -p 7860:7860 \
+    -p 8000:8000 \
+    --device /dev/davinci0 \
+    --device /dev/davinci_manager \
+    --device /dev/devmm_svm \
+    --device /dev/hisi_hdc \
+    --shm-size 16G \
+    --name llamafactory \
+    llamafactory:latest
+
+docker exec -it llamafactory bash
 ```
 
+</details>
+
 <details><summary>数据卷详情</summary>
 
 - hf_cache：使用宿主机的 Hugging Face 缓存文件夹，允许更改为新的目录。

docker/docker-cuda/Dockerfile

@@ -12,13 +12,14 @@ ARG PIP_INDEX=https://pypi.org/simple
 WORKDIR /app
 
 # Install the requirements
-COPY requirements.txt /app/
+COPY requirements.txt /app
 RUN pip config set global.index-url $PIP_INDEX
+RUN pip config set global.extra-index-url $PIP_INDEX
 RUN python -m pip install --upgrade pip
 RUN python -m pip install -r requirements.txt
 
 # Copy the rest of the application into the image
-COPY . /app/
+COPY . /app
 
 # Install the LLaMA Factory
 RUN EXTRA_PACKAGES="metrics"; \
@@ -38,10 +39,9 @@ RUN EXTRA_PACKAGES="metrics"; \
 VOLUME [ "/root/.cache/huggingface/", "/app/data", "/app/output" ]
 
 # Expose port 7860 for the LLaMA Board
+ENV GRADIO_SERVER_PORT 7860
 EXPOSE 7860
 
 # Expose port 8000 for the API service
+ENV API_PORT 8000
 EXPOSE 8000
-
-# Launch LLaMA Board
-CMD [ "llamafactory-cli", "webui" ]

docker/docker-cuda/docker-compose.yml

@@ -1,8 +1,8 @@
 services:
   llamafactory:
     build:
-      dockerfile: Dockerfile
-      context: .
+      dockerfile: ./docker/docker-cuda/Dockerfile
+      context: ../..
       args:
         INSTALL_BNB: false
         INSTALL_VLLM: false

docker/docker-npu/Dockerfile

@@ -0,0 +1,41 @@
+# Use the Ubuntu 22.04 image with CANN 8.0.rc1
+# More versions 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
+
+# Install the requirements
+COPY requirements.txt /app
+RUN pip config set global.index-url $PIP_INDEX
+RUN pip config set global.extra-index-url $PIP_INDEX
+RUN python -m pip install --upgrade pip
+RUN python -m pip install -r requirements.txt
+
+# Copy the rest of the application into the image
+COPY . /app
+
+# Install the LLaMA Factory
+RUN EXTRA_PACKAGES="torch-npu,metrics"; \
+    if [ "$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
+ENV GRADIO_SERVER_PORT 7860
+EXPOSE 7860
+
+# Expose port 8000 for the API service
+ENV API_PORT 8000
+EXPOSE 8000

docker/docker-npu/docker-compose.yml

@@ -0,0 +1,30 @@
+services:
+  llamafactory:
+    build:
+      dockerfile: ./docker/docker-npu/Dockerfile
+      context: ../..
+      args:
+        INSTALL_DEEPSPEED: false
+        PIP_INDEX: https://pypi.org/simple
+    container_name: llamafactory
+    volumes:
+      - ./hf_cache:/root/.cache/huggingface/
+      - ./data:/app/data
+      - ./output:/app/output
+      - /usr/local/dcmi:/usr/local/dcmi
+      - /usr/local/bin/npu-smi:/usr/local/bin/npu-smi
+      - /usr/local/Ascend/driver:/usr/local/Ascend/driver
+      - /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