ubuntu18.04安装pytorch，配置faster-rcnn记录

运行环境

• ubuntu18.04
• cuda10.1
• python3.6（ Anaconda）
• pytorch1.4
• torchvision0.5.0

一. anaconda创建虚拟环境

1. 创建虚拟环境

create -n python3.6 python==3.6

2. 激活虚拟环境

 conda activate python3.6

二. pytorch安装

1. 进入官网https://pytorch.org/，选择自己需要的版本安装
   

不建议安装最新版本的pytorch，否则安装编译faster-rcnn的时候会编译失败，我选择的pytorch1.4

2. 建议选择pip安装pytorch，不容易失败，conda容易安装失败

pip install torch==1.4.0 torchvision==0.5.0

3. 验证pytorch是否安装成功

(python3.6) xxX@OMEN:~$ python
Python 3.6.13 |Anaconda, Inc.| (default, Feb 23 2021, 21:15:04) 
[GCC 7.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> import torchvision
>>> torch.cuda.is_available()
True
>>> 

三. 下载faster-rcnn.pytorch代码

git clone https://github.com/jwyang/faster-rcnn.pytorch.git

此时代码下载到主目录下faster-rcnn.pytorch

cd faster-rcnn.pytorch && mkdir data
#安装python依赖项
pip install -r requirements.txt
#编译cuda依赖项
cd lib
sh make.sh

报错ImportError: torch.utils.ffi is deprecated. Please use cpp extensions instead.
原因是当前使用的是pytorch0.4版本的，需要切换到pytorch1.0，参考博客中的“Step1、安装”重新安装编译安装。

四. 训练模型
参考博客

1. 下载特征提取层的预训练(pre-trained)权重
   VGG 16：
   ResNet 101 ：
   下载完成后把他们放到data/pretrained_model目录下。
2. 准备数据集

#Beyond the demo: installation for training and testing models

#Download the training, validation, test data and VOCdevkit

    wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar
    wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar
    wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCdevkit_08-Jun-2007.tar

#Extract all of these tars into one directory named VOCdevkit

    tar xvf VOCtrainval_06-Nov-2007.tar
    tar xvf VOCtest_06-Nov-2007.tar
    tar xvf VOCdevkit_08-Jun-2007.tar

#It should have this basic structure

#    $VOCdevkit/                           # development kit
#    $VOCdevkit/VOCcode/                   # VOC utility code
#    $VOCdevkit/VOC2007                    # image sets, annotations, etc.
    # ... and several other directories ...

#    Create symlinks for the PASCAL VOC dataset

    cd $FRCN_ROOT/data
    ln -s $VOCdevkit VOCdevkit2007

#    Using symlinks is a good idea because you will likely want to share the same PASCAL dataset installation between multiple projects.

3. 训练模型

CUDA_VISIBLE_DEVICES=0 python trainval_net.py --dataset pascal_voc --net res101 --cuda