弱监督学习框架 Detectron2/DRN-WSOD-pytorch 在服务器/windows上配置安装及使用

最近做弱监督学习研究，进行相关分析。发现Detectron2是一个不错的框架，其中也有model zoo相当多种类的预训练模型可以拿来直接用。但是安装配置使用中碰到了许多坑。跟各位小伙伴们分享。

推荐使用Linux Ubuntu16.04以上版本安装，虚拟机不太好使

我的环境：

GPU：4 x RTX2080

Linux: Ubuntu 16.04 x64 预装doc

Nvidia驱动：440.82

CUDA:10.1

Usr: 无sudo权限

服务器安装指南：

1.强烈推荐先安装anaconda

找到合适的的anaconda安装包，推荐使用国内镜像
Index of /anaconda/archive/ | 北京外国语大学开源软件镜像站 | BFSU Open Source Mirror
推荐使用2019以上的版本，ubutun的话选择linux_x64
为方便在服务器端，使用xshell操作下载，使用wget命令+复制对应的版本下载地址直接下载。
小提示：安装完后需要在是否init conda时输入yes。如果没有进行这一步，参考博文安装conda时没有执行conda init_weixin_43364556的博客-CSDN博客_conda init
```
/home/XXXXXX/anaconda3/bin/conda shell.bash hook >>~/.bashrc
```
- 直接在终端执行，将XXXXX替换成自己的用户名
- 之后执行source .bashrc
显示(base)后完成安装

2.安装PyTorch

确保驱动和cuda已经安装好了，进入pytorch官网，选择对应版本，这里推荐使用conda安装方式，以便整个conda下的所有环境都可以使用pytorch不必重复安装

https://pytorch.org/get-started/locally/

如果cuda版本比较老，需要自己认证去寻找previous version

https://pytorch.org/get-started/previous-versions/

3.下载并配置Detectron2

Git链接：GitHub - shenyunhang/DRN-WSOD-pytorch: Enabling Deep Residual Networks for Weakly Supervised Object Detection创建合适的项目路径，通过以下执行clone

git clone https://github.com/shenyunhang/DRN-WSOD-pytorch.git

以下是官方给出的安装指导和配置要求，供参考。

Installation

Our Colab Notebook has step-by-step instructions that install detectron2. The Dockerfile also installs detectron2 with a few simple commands.

Requirements

Linux or macOS with Python ≥ 3.6
PyTorch ≥ 1.4 and torchvision that matches the PyTorch installation. You can install them together at pytorch.org to make sure of this
OpenCV is optional and needed by demo and visualization

Build Detectron2 from Source

gcc & g++ ≥ 5 are required. ninja is recommended for faster build. After having them, run:
python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
# (add --user if you don't have permission)# Or, to install it from a local clone:
git clone https://github.com/facebookresearch/detectron2.git
python -m pip install -e detectron2# Or if you are on macOS
CC=clang CXX=clang++ python -m pip install ......
To rebuild detectron2 that's built from a local clone, use rm -rf build/ **/*.so to clean the old build first. You often need to rebuild detectron2 after reinstalling PyTorch.

Install Pre-Built Detectron2 (Linux only)

Choose from this table:

CUDA torch 1.5 torch 1.4
10.2
install

10.1
install

install

10.0
install

9.2
install

install

cpu
install

install

Note that:

The pre-built package has to be used with corresponding version of CUDA and official PyTorch release. It will not work with a different version of PyTorch or a non-official build of PyTorch.
Such installation is out-of-date w.r.t. master branch of detectron2. It may not be compatible with the master branch of a research project that uses detectron2 (e.g. those in projects or meshrcnn).

CUDA	torch 1.5	torch 1.4
10.2	install
10.1	install	install
10.0		install
9.2	install	install
cpu	install	install

发现安装需要执行

python3 -m pip install -e .

其实对应的就是根目录下执行

pip install setup.py

如果在安装中有错误，请根据setup.py内要求的依赖包逐个安装

安装完后准备试一下他的功力，这里直接选用demo/中的demo.py做实验

以下是GETSTART.MD的部分内容

Getting Started with Detectron2

This document provides a brief intro of the usage of builtin command-line tools in detectron2.

For a tutorial that involves actual coding with the API, see our Colab Notebook which covers how to run inference with an existing model, and how to train a builtin model on a custom dataset.

For more advanced tutorials, refer to our documentation.

Inference Demo with Pre-trained Models

Pick a model and its config file from model zoo, for example, mask_rcnn_R_50_FPN_3x.yaml.
We provide demo.py that is able to run builtin standard models. Run it with:
cd demo/
python demo.py --config-file ../configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml \--input input1.jpg input2.jpg \[--other-options]--opts MODEL.WEIGHTS detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl
The configs are made for training, therefore we need to specify MODEL.WEIGHTS to a model from model zoo for evaluation. This command will run the inference and show visualizations in an OpenCV window.

For details of the command line arguments, see demo.py -h or look at its source code to understand its behavior. Some common arguments are:

To run on your webcam, replace --input files with --webcam.
To run on a video, replace --input files with --video-input video.mp4.
To run on cpu, add MODEL.DEVICE cpu after --opts.
To save outputs to a directory (for images) or a file (for webcam or video), use --output.

在其他博客中detectron2 + ubuntu + cpu - 走看看有对其的一点点翻译大家可以参看。

注意：这里要说的是在线下载模型文件不太好用，无论是用代理还是不用，都下载不了

所以选择model zoo中直接下载pkl与训练模型，到demo/目录https://github.com/shenyunhang/DRN-WSOD-pytorch/blob/DRN-WSOD/MODEL_ZOO.mdhttps://github.com/shenyunhang/DRN-WSOD-pytorch/blob/DRN-WSOD/MODEL_ZOO.md注意，id要对应上，然后将你要识别的图片也放在demo/下，执行命令

python demo.py --config-file ../configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml --input ies.jpg --output 1.jpg --opts MODEL.WEIGHTS model_final_f10217.pkl

[12/22 01:35:08 detectron2]: Arguments: Namespace(confidence_threshold=0.5, config_file='../configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml', input=['ies.jpg'], opts=['MODEL.WEIGHTS', 'model_final_f10217.pkl'], output='1.jpg', video_input=None, webcam=False)
[12/22 01:35:12 fvcore.common.checkpoint]: [Checkpointer] Loading from model_final_f10217.pkl ...
[12/22 01:35:12 fvcore.common.checkpoint]: Reading a file from 'Detectron2 Model Zoo'
0%| | 0/1 [00:00<?, ?it/s]/home/zhf/PJ/wsod/DRN-WSOD-pytorch/detectron2/layers/wrappers.py:226: UserWarning: This overload of nonzero is deprecated:
nonzero()
Consider using one of the following signatures instead:
nonzero(*, bool as_tuple) (Triggered internally at /opt/conda/conda-bld/pytorch_1607370141920/work/torch/csrc/utils/python_arg_parser.cpp:882.)
return x.nonzero().unbind(1)
[12/22 01:35:13 detectron2]: ies.jpg: detected 10 instances in 0.36s
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 2.49it/s]

完成后1.png即为输出文件，导出后查看，效果还可以，输入图片可以是任意尺寸的！