Jetson TX1 install py-faster-rcnn

Install py-faster-rcnn following the official version

https://github.com/rbgirshick/py-faster-rcnn

  • Build the Cython modules
cd $FRCN_ROOT/lib
make

Errors:

Traceback (most recent call last):
  File "setup.py", line 58, in <module>
  CUDA = locate_cuda()
  File "setup.py", line 55, in locate_cuda
  raise EnvironmentError(‘The CUDA %s path could not be located in %s‘ % (k, v))
EnvironmentError: The CUDA lib path could not be located in /usr/local/cuda-8.0/lib

Reason: the cuda path is not correct in ./lib/setup.py

Solver:

gedit 打开 setup.py

cudaconfig = {‘home‘:home, ‘nvcc‘:nvcc,
                  ‘include‘: pjoin(home, ‘include‘),
                  ‘lib‘: pjoin(home, ‘lib‘)}

把上面的最后一个lib改为lib64即可

cudaconfig = {‘home‘:home, ‘nvcc‘:nvcc,
                  ‘include‘: pjoin(home, ‘include‘),
                  ‘lib‘: pjoin(home, ‘lib64‘)}
  • Build Caffe and pycaffe
cd $FRCN_ROOT/caffe-fast-rcnn
# Now follow the Caffe installation instructions here:
#   http://caffe.berkeleyvision.org/installation.html

# If you‘re experienced with Caffe and have all of the requirements installed
# and your Makefile.config in place, then simply do:
make -j8 && make pycaffe

Errors:

[email protected]:~/caffe$ make all
PROTOC src/caffe/proto/caffe.proto
CXX .build_release/src/caffe/proto/caffe.pb.cc
CXX src/caffe/data_transformer.cpp
CXX src/caffe/common.cpp
CXX src/caffe/internal_thread.cpp
CXX src/caffe/blob.cpp
CXX src/caffe/data_reader.cpp
CXX src/caffe/parallel.cpp
CXX src/caffe/util/hdf5.cpp
In file included from src/caffe/util/hdf5.cpp:1:0:
./include/caffe/util/hdf5.hpp:6:18: fatal error: hdf5.h: No such file or directory
compilation terminated.
Makefile:572: recipe for target ‘.build_release/src/caffe/util/hdf5.o‘ failed
make: *** [.build_release/src/caffe/util/hdf5.o] Error 1
[email protected]:~/caffe$ 

Solver:

Makefile.config
INCLUDE_DIRS
/usr/include/hdf5/serial/
Makefile
LIBRARIES
hdf5_hl and hdf5 改为 hdf5_serial_hl ,hdf5_serial

  • Demo

Errors:

Loaded network /home/ubuntu/Projects/py-faster-rcnn/data/faster_rcnn_models/VGG16_faster_rcnn_final.caffemodel
Killed

Reason:

The vgg16 model is too large

Solver:

Change to zf model

  • Results:
Loaded network /home/ubuntu/Projects/py-faster-rcnn/data/faster_rcnn_models/ZF_faster_rcnn_final.caffemodel
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Demo for data/demo/000456.jpg
Detection took 0.530s for 300 object proposals
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Demo for data/demo/000542.jpg
Detection took 0.722s for 135 object proposals
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Demo for data/demo/001150.jpg
Detection took 0.643s for 231 object proposals
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Demo for data/demo/001763.jpg
Detection took 0.764s for 200 object proposals
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Demo for data/demo/004545.jpg
Detection took 0.661s for 300 object proposals

时间: 2024-10-11 04:55:18

Jetson TX1 install py-faster-rcnn的相关文章

Jetson TX1 install Opencv3

https://jkjung-avt.github.io/opencv3-on-tx2/ 注意:在编译的时候会遇到内存空间不足的情况,可以插入U盘,将程序拷贝到U盘内编译,然后安装到Jetson上.U盘格式化采用NTFS,其他格式可能无法识别. Installation Steps I’d start by cleaning up older opencv packages and installing necessary dependencies for building opencv. Re

使用faster rcnn 跑vot2015的数据集

本周老师给的任务: 一是将VOT15数据集(世华已传到服务器上)上每个序列的第1,11,21,31,41帧分别运行Faster R-CNN检测器并保存在图片上显示的检测结果: 二是将这5帧的ground truth bounding box作为proposal得到其对应的检测器分类结果(比如网络要检测20类物体,那包括背景就是得到21类对应的检测分数值),并将每个序列的检测结果分别存成一个文本文档. 注意,使用代码的时候,可能会有路径错误,还可能是,我贴上的代码,博客园的网站给在某些语句后加了

Windows下TensorFlow+Faster Rcnn 详细安装步骤

参考:https://cloud.tencent.com/developer/news/303081 实现步骤也很简单,实现流程如下: 1.安装前准备: 安装前请确保TensorFlow和相关的依赖库安装成功 2.下载 faster rcnn程序包 连接:https://github.com/dBeker/Faster-RCNN-TensorFlow-Python3 解压,然后可以看一下里面的readme 3.安装python依赖库 注意还是要在TensorFlow虚拟环境下安装,命令: pip

Faster RCNN代码理解(Python)

转自http://www.infocool.net/kb/Python/201611/209696.html#原文地址 第一步,准备 从train_faster_rcnn_alt_opt.py入: 初始化参数:args = parse_args() 采用的是Python的argparse 主要有–net_name,–gpu,–cfg等(在cfg中只是修改了几个参数,其他大部分参数在congig.py中,涉及到训练整个网络). cfg_from_file(args.cfg_file) 这里便是代用

Faster rcnn代码理解(1)

这段时间看了不少论文,回头看看,感觉还是有必要将Faster rcnn的源码理解一下,毕竟后来很多方法都和它有相近之处,同时理解该框架也有助于以后自己修改和编写自己的框架.好的开始吧- 这里我们跟着Faster rcnn的训练流程来一步一步梳理,进入tools\train_faster_rcnn_alt_opt.py中: 首先从__main__入口处进入,如下: 上图中首先对终端中的命令行进行解析,获取相关的命令参数:然后利用mp.Queue()创建一个多线程的对象,再利用get_solvers

faster r-cnn 在CPU配置下训练自己的数据

因为没有GPU,所以在CPU下训练自己的数据,中间遇到了各种各样的坑,还好没有放弃,特以此文记录此过程. 1.在CPU下配置faster r-cnn,参考博客:http://blog.csdn.net/wjx2012yt/article/details/52197698#quote 2.在CPU下训练数据集,需要对py-faster-rcnn内的roi_pooling_layer和smooth_L1_loss_layer改为CPU版本, 并重新编译.这位博主对其进行了修改,可直接进行替换:htt

[转]Jetson TX1 开发教程(1)配置与刷机

开箱 Jetson TX1是英伟达公司新出的GPU开发板,拥有世界上先进的嵌入式视觉计算系统,提供高性能.新技术和极佳的开发平台.在进行配置和刷机工作之前,先来一张全家福: 可以看到,Jetson TX1 包含内容为:一块开发板.WIFI天线.AC电源(缺了插座线,需要自己去买).Micro USB线.USB-OTG线.排线连接器(感觉没什么用). Jetson TX1找不到太多的教程,我也只能参考一些CSDN博客,摸索着进行配置刷机工作. 开机并安装驱动 开发板连接好AC电源线,使用HDMI线

纯C++版500VIP源码下载的Faster R-CNN(通过caffe自定义RPN层实现)

这里500VIP源码下载 dsluntan.com 介绍的是通过添加自定义层(RPN层)代替python层,实现c++版的Faster R-CNN,因为去掉python了,所以部署时不会因为牵扯到python库等其它的莫名其妙的错误,使用起来就跟单纯的caffe一样,更简单方便. 核心代码,借鉴的是这篇博客,这里的话,我们不扣具体的代码细节(比如rpn层是怎么产出候选框啊,非极大值抑制是具体怎么实现的等等),有兴趣的可以自己查下资料,所以主要是走一个步骤,从而完成c++版Faster R-CNN

Faster RCNN算法代码解析

一. Faster-RCNN代码解释 先看看代码结构: Data: This directory holds (after you download them): Caffe models pre-trained on ImageNet Faster R-CNN models Symlinks to datasets demo 5张图片 scripts 下载模型的脚本 Experiments: logs scripts/faster_rcnn_alt_opt.sh cfgs/faster_rcn