美文网首页
图文识别(Faster R-CNN for Tensorflow

图文识别(Faster R-CNN for Tensorflow

作者: Asen_十足坏蛋 | 来源:发表于2020-01-16 16:01 被阅读0次

环境情况:
1、ubuntu 18.04.3(虚拟环境)
2、python 3.6.8(如果有多个python,需要删除一些,并只保留一个)

https://www.tensorflow.org/install/source

1、安装TensorFlow

安装 Python 和 TensorFlow 软件包依赖项
   sudo apt install python-dev python-pip  # or python3-dev python3-pip
  安装 TensorFlow pip 软件包依赖项(如果使用虚拟环境,请省略 --user 参数):
   pip install -U --user pip six numpy wheel setuptools mock future>=0.17.1
   pip install -U --user keras_applications==1.0.6 --no-deps
   pip install -U --user keras_preprocessing==1.0.5 --no-deps

实际:

sudo apt install python3-dev python3-pip
pip3 install -U pip six numpy wheel setuptools mock future>=0.17.1
pip3 install -U --user keras_applications==1.0.6 --no-deps
pip3 install -U --user keras_preprocessing==1.0.5 --no-deps

2、安装Bazel

  • 步骤1:将Bazel发行版URI添加为包源
sudo apt install curl
curl https://bazel.build/bazel-release.pub.gpg | sudo apt-key add -
echo "deb [arch=amd64] https://storage.googleapis.com/bazel-apt stable jdk1.8" | sudo tee /etc/apt/sources.list.d/bazel.list
  • 步骤2:安装和更新Bazel
sudo apt update && sudo apt install bazel(无效)
sudo apt update && sudo apt install bazel-1.1.0(变更为此版本)
sudo apt update && sudo apt full-upgrade
sudo apt install openjdk-11-jdk

3、安装tensorflow

image.png

1、安装git

sudo apt install git
git clone https://github.com/tensorflow/tensorflow.git
cd tensorflow

2、配置编译系统

./configure

选是否的可以用 'n'

3、编译

bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package

变更为(上面那句受限于内存大小,会中途终止):

bazel build -c opt --copt=-mavx --copt=-mavx2 --copt=-mfma --copt=-msse4.1 --copt=-msse4.2 -k //tensorflow/tools/pip_package:build_pip_package --local_resources 3072,.5,1.0

当提示如下错误,但又明明安装了numpy时,可能是多个python版本造成的。

Traceback (most recent call last):
 File "<string>", line 1, in <module>
ModuleNotFoundError: No module named 'numpy'

编译时,可能会出现网络404的错误,多试几次,或尝试翻墙。

./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg

pip install /tmp/tensorflow_pkg/tensorflow-2.0.0-cp36-cp36m-linux_x86_64.whl

PS: 1、最后的文件,看自己电脑中“tmp/tensorflow_pkg/”中的文件名
2、第二个命令执行时,需要pip进行在线下载,可能由于网络原因而中断,多试几次

此时会报错:

Could not find a version that satisfies the requirement grpcio>=1.8.6 (from tf-nightly==2.0.0) (from versions: none)

当出现“Successfully installed ....”则属于安装完成。

3 安装tf-faster-rcnn

需要先安装:

// pip安装cython
pip install --user --upgrade cython
// 安装cuda(不用安装,这个是gpu用的,虚拟机装不上)
wget http://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run
sudo sh cuda_10.2.89_440.33.01_linux.run

下载并编译

git clone https://github.com/endernewton/tf-faster-rcnn.git
cd ./tf-faster-rcnn/lib
make clean
make

下载下来的版本,默认采用gpu进行编译(如果本身系统带gpu的,直接编译即可),需要做如下调整变更为cpu模式:

  • /lib/model/config.py
# Use GPU implementation of non-maximum suppression
__C.USE_GPU_NMS = False
  • lib/model/nms_wrapper.py
# --------------------------------------------------------
# Fast R-CNN
# Copyright (c) 2015 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Ross Girshick
# --------------------------------------------------------
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from model.config import cfg
#from nms.gpu_nms import gpu_nms
from nms.cpu_nms import cpu_nms

def nms(dets, thresh, force_cpu=False):
  """Dispatch to either CPU or GPU NMS implementations."""

  if dets.shape[0] == 0:
    return []
  #if cfg.USE_GPU_NMS and not force_cpu:
    #return gpu_nms(dets, thresh, device_id=0)
  else:
    return cpu_nms(dets, thresh)
  • lib/setup.py


    image.png
image.png image.png

4 安装 Python COCO API,这是为了使用COCO数据库

cd data
git clone https://github.com/pdollar/coco.git
cd coco/PythonAPI
make
cd ../../..

5 下载预训练模型

# Resnet101 for voc在07 + 12上进行了预训练设置 
./data/scripts/fetch_faster_rcnn_models.sh

faster_rcnn路径下执行

NET=res101
TRAIN_IMDB=voc_2007_trainval+voc_2012_trainval
mkdir -p output/${NET}/${TRAIN_IMDB}
cd output/${NET}/${TRAIN_IMDB}
ln -s ../../../data/voc_2007_trainval+voc_2012_trainval ./default
cd ../../..

运行示例程序(转到最下面)

GPU_ID=01
CUDA_VISIBLE_DEVICES=${GPU_ID} ./tools/demo.py

其他缺少的安装内容

pip install --user easydict
pip install --user opencv-python
pip install --user matplotlib

安装旧版本的tensorflow,新版本有个内容移除了,但r-cnn并没有变更,会报错

pip uninstall tensorflow
pip install --user tensorflow==1.13.2

6 结束

至此环境搭建完成,demo运行效果如下图:


image.png

参考地址:

1、https://docs.bazel.build/versions/master/install-ubuntu.html
2、https://www.tensorflow.org/install/source#setup_for_linux_and_macos

相关文章

网友评论

      本文标题:图文识别(Faster R-CNN for Tensorflow

      本文链接:https://www.haomeiwen.com/subject/qvecoctx.html