• Linux 和 macOS
• Python 3.6+
• PyTorch 1.3+
• CUDA 9.2+
• GCC 5+
• MMCV

1. Create a virtual environment with conda and enter the virtual environment.

   conda create -n mmdet python=3.7 -y
   conda activate mmdet

2. Install PyTorch and torchvision based on PyTorch website, for example:

   conda install pytorch torchvision -c pytorch

   conda install pytorch cudatoolkit=10.1 torchvision -c pytorch

   conda install pytorch=1.3.1 cudatoolkit=9.2 torchvision=0.4.2 -c pytorch

1. To install mmcv-full, we recommend using the pre-built package to install:

   pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/{cu_version}/{torch_version}/index.html

   You need to replace '{cu_version}' and '{torch_version}' on the command line with the corresponding version. For example, in CUDA 11 and PyTorch 1.7.0 environments, you can install the latest version of MMCV with the following command:

   pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu110/torch1.7.0/index.html

   Refer to [MMCV] (https://mmcv.readthedocs.io/en/latest/#installation) access to different versions of the compatible to different MMCV PyTorch and CUDA version. At the same time, you can also compile MMCV from source by using the following command line:

   git clone https://github.com/open-mmlab/mmcv.git
   cd mmcv
   MMCV_WITH_OPS=1 pip install -e .  
   cd ..

   pip install mmcv-full

   pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu110/torch1.7/index.html
   

2. Insatll MMDetection：

   pip install -r requirements/build.txt
   pip install -v -e .  # or "python setup.py develop"

conda create -n mmdet python=3.7 -y
conda activate mmdet

conda install pytorch==1.6.0 torchvision==0.7.0 cudatoolkit=10.1 -c pytorch -y

# Install mmcv
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu101/torch1.6.0/index.html (这里要选择对应当前mmdet的版本的mmcv版本，具体版本查看可查看mmdet/version.py)

# Insatll MMDetection
pip install -r requirements/build.txt
pip install -v -e .

To verify that MMDetection and the required environment are installed correctly, we can run the sample Python code to initialize the detector and reason about a demonstration image:

from mmdet.apis import init_detector, inference_detector

config_file = 'configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
#http://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
checkpoint_file = 'checkpoints/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth'
device = 'cuda:0'
# Initialize the detector
model = init_detector(config_file, checkpoint_file, device=device)
#  Reasoning demo image
inference_detector(model, 'demo/demo.jpg')

If MMDetection is installed successfully, the above code runs completely.

# The model is trained using the command line
python tools/train.py work_dirs/esnet/mask_rcnn_r101_fpn_2x_coco.py ```

## Test code
```python
# Test the model using the command line
python tools/test.py work_dirs/esnet/mask_rcnn_r101_fpn_2x_coco.py work_dirs/esnet/epoch_40.pth```

## Single picture reasoning
```python
# Single-image reasoning on the model using the command line
python demo/image_demo.py demo/test.tif work_dirs/esnet/mask_rcnn_r101_fpn_2x_coco.py /TEST/xiuyu.li/esnet/epoch_40.pth