feat: support Docker installation

Dockerfile

@@ -0,0 +1,64 @@
+# Copyright (c) 2023 Amphion.
+#
+# This source code is licensed under the MIT license found in the
+# LICENSE file in the root directory of this source tree.
+
+# Other version: https://hub.docker.com/r/nvidia/cuda/tags
+FROM nvidia/cuda:11.8.0-cudnn8-devel-ubuntu18.04
+
+ARG DEBIAN_FRONTEND=noninteractive
+ARG PYTORCH='2.0.0'
+ARG CUDA='cu118'
+ARG SHELL='/bin/bash'
+ARG MINICONDA='Miniconda3-py39_23.3.1-0-Linux-x86_64.sh'
+
+ENV LANG=en_US.UTF-8 PYTHONIOENCODING=utf-8 PYTHONDONTWRITEBYTECODE=1 CUDA_HOME=/usr/local/cuda CONDA_HOME=/opt/conda SHELL=${SHELL}
+ENV PATH=$CONDA_HOME/bin:$CUDA_HOME/bin:$PATH \
+    LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH \
+    LIBRARY_PATH=$CUDA_HOME/lib64:$LIBRARY_PATH \
+    CONDA_PREFIX=$CONDA_HOME \
+    NCCL_HOME=$CUDA_HOME
+
+# Install ubuntu packages
+RUN sed -i 's/archive.ubuntu.com/mirrors.cloud.tencent.com/g' /etc/apt/sources.list \
+    && sed -i 's/security.ubuntu.com/mirrors.cloud.tencent.com/g' /etc/apt/sources.list \
+    && rm /etc/apt/sources.list.d/cuda.list \
+    && apt-get update \
+    && apt-get -y install \
+    python3-pip ffmpeg git less wget libsm6 libxext6 libxrender-dev \
+    build-essential cmake pkg-config libx11-dev libatlas-base-dev \
+    libgtk-3-dev libboost-python-dev vim libgl1-mesa-glx \
+    libaio-dev software-properties-common tmux \
+    espeak-ng
+
+# Install miniconda with python 3.9
+USER root
+# COPY Miniconda3-py39_23.3.1-0-Linux-x86_64.sh /root/anaconda.sh
+RUN wget -t 0 -c -O /tmp/anaconda.sh https://repo.anaconda.com/miniconda/${MINICONDA} \
+    && mv /tmp/anaconda.sh /root/anaconda.sh \
+    && ${SHELL} /root/anaconda.sh -b -p $CONDA_HOME \
+    && rm /root/anaconda.sh
+
+RUN conda create -y --name amphion python=3.9.15
+
+WORKDIR /app
+COPY env.sh env.sh
+RUN chmod +x ./env.sh
+
+RUN ["conda", "run", "-n", "amphion", "-vvv", "--no-capture-output", "./env.sh"]
+
+RUN conda init \
+    && echo "\nconda activate amphion\n" >> ~/.bashrc
+
+CMD ["/bin/bash"]
+
+# *** Build ***
+# docker build -t realamphion/amphion .
+
+# *** Run ***
+# cd Amphion
+# docker run --runtime=nvidia --gpus all -it -v .:/app -v /mnt:/mnt_host realamphion/amphion
+
+# *** Push and release ***
+# docker login
+# docker push realamphion/amphion

README.md

@@ -81,6 +81,10 @@ Amphion unifies the data preprocess of the open-source datasets including [Audio
 
 ## 📀 Installation
 
+Amphion can be installed through either Setup Installer or Docker Image.
+
+### Setup Installer
+
 ```bash
 git clone https://github.com/open-mmlab/Amphion.git
 cd Amphion
@@ -93,6 +97,21 @@ conda activate amphion
 sh env.sh
 ```
 
+### Docker Image
+
+1. Install [Docker](https://docs.docker.com/get-docker/), [NVIDIA Driver](https://www.nvidia.com/download/index.aspx), [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html), and [CUDA](https://developer.nvidia.com/cuda-downloads).
+
+2. Run the following commands:
+```bash
+git clone https://github.com/open-mmlab/Amphion.git
+cd Amphion
+
+docker pull realamphion/amphion
+docker run --runtime=nvidia --gpus all -it -v .:/app realamphion/amphion
+```
+Mount dataset by argument `-v` is necessary when using Docker. Please refer to [Mount dataset in Docker container](egs/datasets/docker.md) and [Docker Docs](https://docs.docker.com/engine/reference/commandline/container_run/#volume) for more details.
+
+
 ## 🐍 Usage in Python
 
 We detail the instructions of different tasks in the following recipes:

egs/datasets/README.md

@@ -23,6 +23,8 @@ Amphion support the following academic datasets (sort alphabetically):
 
 The downloading link and the file structure tree of each dataset is displayed as follows.
 
+> **Note:** When using Docker to run Amphion, mount the dataset to the container is necessary after downloading. Check [Mount dataset in Docker container](./docker.md) for more details.
+
 ## AudioCaps
 
 AudioCaps is a dataset of around 44K audio-caption pairs, where each audio clip corresponds to a caption with rich semantic information.

egs/datasets/docker.md

@@ -0,0 +1,19 @@
+# Mount dataset in Docker container
+
+When using Docker to run Amphion, mount the dataset to the container first is needed. It is recommend to mounte dataset to `/mnt/<dataset_name>` in the container, where `<dataset_name>` is the name of the dataset.
+
+When configuring the dataset in `exp_config.json`, you should use the path `/mnt/<dataset_name>` as the dataset path instead of the actual path on your host machine. Otherwise, the dataset will not be found in the container.
+
+## Mount Example
+
+```bash
+docker run --runtime=nvidia --gpus all -it -v .:/app -v <dataset_path1>:/mnt/<dataset_name1> -v <dataset_path2>:/mnt/<dataset_name2> amphion
+```
+
+For example, if you want to use the `LJSpeech` dataset, you can mount the dataset to `/mnt/LJSpeech` in the container.
+
+```bash
+docker run --runtime=nvidia --gpus all -it -v .:/app -v /home/username/datasets/LJSpeech:/mnt/LJSpeech amphion
+```
+
+If you want to use multiple datasets, you can mount them to different directories in the container by adding more `-v` options.

env.sh

@@ -3,6 +3,9 @@
 # This source code is licensed under the MIT license found in the
 # LICENSE file in the root directory of this source tree.
 
+# Raise error if any command fails
+set -e
+
 # Install ffmpeg in Linux
 conda install -c conda-forge ffmpeg