Update install.md

docs/en/install.md

@@ -1,8 +1,8 @@
 ## Prerequisites
 
-- Linux (Windows is not officially supported)
+- Linux / Windows / Mac
 - Python 3.6+
-- PyTorch 1.3 or higher
+- PyTorch 1.5 or higher
 - CUDA 9.0 or higher
 - NCCL 2
 - GCC 5.4 or higher
@@ -13,17 +13,22 @@
 a. Create a conda virtual environment and activate it.
 
 ```shell
-conda create -n open-mmlab python=3.7 -y
+conda create -n open-mmlab python=3.8 -y
 conda activate open-mmlab
 ```
 
-b. Install PyTorch and torchvision following the [official instructions](https://pytorch.org/), e.g.,
+b. Install PyTorch and torchvision following the [official instructions](https://pytorch.org/), then install mmcv-full
+
+e.g. `cuda 10.1` & `pytorch 1.7`,
 
 ```shell
-conda install pytorch torchvision -c pytorch
+conda install pytorch==1.7.1 torchvision cudatoolkit=10.1 -c pytorch
+pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7/index.html "opencv-python<=4.5.4.60"
 ```
 
-Note: Make sure that your compilation CUDA version and runtime CUDA version match.
+Note 1: Higher version `opencv-python` is not supported.
+
+Note 2: Make sure that your compilation CUDA version and runtime CUDA version match.
 You can check the supported CUDA version for precompiled packages on the [PyTorch website](https://pytorch.org/).
 
 `E.g. 1` If you have CUDA 10.1 installed under `/usr/local/cuda`, you need to install the prebuilt PyTorch with CUDA 10.1.
@@ -39,9 +44,6 @@ PyTorch 1.3.1., you need to install the prebuilt PyTorch with CUDA 9.2.
 conda install pytorch=1.3.1 cudatoolkit=9.2 torchvision=0.4.2 -c pytorch
 ```
 
-If you build PyTorch from source instead of installing the prebuilt package,
-you can use more CUDA versions such as 9.0.
-
 c. Clone the mmediting repository.
 
 ```shell
@@ -62,6 +64,18 @@ If you build mmediting on macOS, replace the last command with
 CC=clang CXX=clang++ CFLAGS='-stdlib=libc++' pip install -e .
 ```
 
+e. Verify installation
+
+After the installation is complete, you can switch to another directory (such as `/home`) and try to import mmedit in python. If the import is successful, the installation is successful.
+
+```shell
+$ cd ~
+$ python
+>>> import mmedit
+>>> mmedit.__version__
+'0.12.0'
+```
+
 Note:
 
 1. The git commit id will be written to the version number with step d, e.g. 0.6.0+2e7045c. The version will also be saved in trained models.
@@ -85,6 +99,13 @@ you can install it before installing MMCV.
 ### Install with CPU only
 The code can be built for CPU only environment (where CUDA isn't available).
 
+Accordingly, install the CPU version of PyTorch and MMCV.
+
+```shell
+conda install pytorch==1.7.1 torchvision cudatoolkit=10.1 -c pytorch
+pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7/index.html "opencv-python<=4.5.4.60"
+```
+
 <!-- In CPU mode you can run the demo/webcam_demo.py for example. -->
 However some functionality is gone in this mode:
 

docs/zh_cn/install.md

@@ -1,29 +1,34 @@
 ## 依赖
 
-- Linux (目前 Windows 暂无官方支持)
+- Linux / Windows / Mac
 - Python 3.6+
-- PyTorch 1.3 或更高
+- PyTorch 1.5 或更高
 - CUDA 9.0 或更高
 - NCCL 2
 - GCC 5.4 或更高
 - [mmcv](https://github.com/open-mmlab/mmcv)
 
 ## 安装
 
-a. 创建并激活 conda 虚拟环境，如：
+a. 创建并激活 conda 虚拟环境，如 `python 3.8`：
 
 ```shell
-conda create -n open-mmlab python=3.7 -y
-conda activate open-mmlab
+conda create -n mmedit python=3.8 -y
+conda activate mmedit
 ```
 
-b. 按照 [PyTorch 官方文档](https://pytorch.org/) 安装 PyTorch 和 torchvision，如：
+b. 按照 [PyTorch 官方文档](https://pytorch.org/) 安装 PyTorch 和 torchvision，然后安装对应路径下的 `mmcv-full`
+
+如 `cuda 10.1` & `pytorch 1.7`：
 
 ```shell
-conda install pytorch torchvision -c pytorch
+conda install pytorch==1.7.1 torchvision cudatoolkit=10.1 -c pytorch
+pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7/index.html "opencv-python<=4.5.4.60"
 ```
 
-注：确保 CUDA 编译版本和 CUDA 运行版本相匹配。 用户可以参照 [PyTorch 官网](https://pytorch.org/) 对预编译包所支持的 CUDA 版本进行核对。
+注 1：过高版本的 `opencv-python` 在使用中存在一些问题，因此在安装时限制其版本。
+
+注 2：确保 CUDA 编译版本和 CUDA 运行版本相匹配。 用户可以参照 [PyTorch 官网](https://pytorch.org/) 对预编译包所支持的 CUDA 版本进行核对。
 
 `例1`：如果 `/usr/local/cuda` 文件夹下已安装了 CUDA 10.1 版本，则需要安装 CUDA 10.1 下预编译的 PyTorch。
 
@@ -36,8 +41,6 @@ conda install pytorch cudatoolkit=10.1 torchvision -c pytorch
 conda install pytorch=1.3.1 cudatoolkit=9.2 torchvision=0.4.2 -c pytorch
 ```
 
-如果你从源码编译 PyTorch 而不是安装的预编译版本的话，你可以使用更多 CUDA 版本（例如9.0）。
-
 c. 克隆 MMEditing 仓库
 
 ```shell
@@ -58,6 +61,19 @@ pip install -v -e .  # or "python setup.py develop"
 CC=clang CXX=clang++ CFLAGS='-stdlib=libc++' pip install -e .
 ```
 
+e. 验证安装
+
+安装完成后，可以切换到其他目录（例如 `/home` 目录），并尝试在 python 中导入 mmedit，导入成功则证明安装成功
+
+```shell
+$ cd ~
+$ python
+
+>>> import mmedit
+>>> mmedit.__version__
+'0.12.0'
+```
+
 注：
 1. git commit 的 id 将会被写到版本号中，如 0.6.0+2e7045c。这个版本号也会被保存到训练好的模型中。 推荐用户每次在对本地代码和 github 上的源码进行同步后，执行一次步骤 b。如果 C++/CUDA 代码被修改，就必须进行这一步骤。
 
@@ -81,6 +97,13 @@ CC=clang CXX=clang++ CFLAGS='-stdlib=libc++' pip install -e .
 
 MMEditing 也可以在只有 CPU 的环境下安装（即无法使用 GPU 的环境）。
 
+相应的，安装 CPU 版本的 PyTorch 和 MMCV
+
+```shell
+conda install pytorch==1.7.1 torchvision cudatoolkit=10.1 -c pytorch
+pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.7/index.html "opencv-python<=4.5.4.60"
+```
+
 然而在该环境下，有些功能将被移除，例如：
 - Deformable Convolution（可变形卷积）