Problems encountered in various deep learning processes:
From scratch~Ubuntu16.04 install caffe+cuda9.1+cndnn9.1+OpenCV2.4.9
http://blog.csdn.net/cpfelix/article/details/79086580
How to install TensorFlow under Ubuntu (note that the installed version is the CPU version , cuda9.1 is not compatible)
http://blog.csdn.net/tina_ttl/article/details/51762471
ubuntu 14.04 builds GPU deep learning tensorflow platform (18.3.12 is installed successfully in cuda7.5, use sudo pip install -- The ignore-installed installation has been importerror)
http://blog.csdn.net/myjiayan/article/details/52084662
Install Tensorflow based on Anaconda in Ubuntu environment
http://blog.csdn.net/hgdwdtt/article/details/78633232
Under Ubuntu Install caffe and tensorflow at the same time
http://www.cnblogs.com/heubme/p/7850348.html
Ubuntu14.04 install CUDA-7.5 (deb installation) +tensorflow
http://blog.csdn.net/zhongshaoyy/article/details /54575116
Faster R-CNN installation and testing experience:
http://blog.csdn.net/wopawn/article/details/52751614
py-faster-rcnn trains the model with its own data:
http://blog.csdn.net/princepaul3/article/details/52313582
Ubuntu16.04 environment Install caffe and run faster-rcnn
http://blog.csdn.net/zoro_lov3/article/details/60581174
pva-faster-rcnn configure, install and train your own dataset
http://blog.csdn.net/samylee/ article/details/52860038
CharlesShang/FastMaskRCNN
https://github.com/CharlesShang/FastMaskRCNN
Installation tutorial (1): Using Anaconda to create a coexistence environment of caffe and tensorflow | Jack Cuihttp://cuijiahua.com/blog/2017/12/dl_1.html
Installation tutorial: Using Anaconda to create a coexistence environment of caffe and tensorflow
http://blog .csdn.net/primezpy/article/details/78819249
make: *** [.build_release/lib/libcaffe.so.1.0.0-rc3] Error 1
https://www.cnblogs.com/zjutzz/p/5716453 .html?utm_source=itdadao&utm_medium=referral
How to switch between Anaconda's python and the system's own python
http://blog.csdn.net/teeyohuang/article/details/79076821
A protobuf problem of compiling caffe (not adopted)
http://blog.csdn.net/teeyohuang/article/details/79076821 ://blog.csdn.net/qq_17278169/article/details/64128525
Installation Tutorial: Using Anaconda to create a coexistence environment of caffe and tensorflowhttp
:
//blog.csdn.net/primezpy/article/details/78819249 KeyError: 'chair'
http://www.bubuko.com/infodetail-2096570.html
py-faster-rcnn] [training own data] Notes on parameters that need to be modified
http://blog.csdn.net/amor_tila/article/details/78452261
Caffe Implementation of Google's MobileNets (v1 and v2)
https://github.com/shicai/MobileNet-Caffe
faster rcnn The problem of testing VOC2012https://www.cnblogs.com/xthrough/p/6386651.htmlThere is a problem in training
GoogLeNet
:check failed:mdb_status==0(2vs.0)No such file or directiry [
http://bbs.csdn.net/topics/391931961
MobileNet Tutorial (2): Building an Image Classification App on Android Phones with TensorFlow
https: //baijiahao.baidu.com/s?id=1574957535863211&wfr=spider&for=pc
TensorFlow installation tutorial
http://blog.csdn.net/bitcarmanlee/article/details/52749488
Various pits encountered by Tensorflow running the MaskRCNN framework (1)
http://blog.csdn.net/aimer_chen/article/details/78528000
matterport Mask_RCNN configurationhttp
: //blog.csdn.net/wei_guo_xd/article/details/78579534
Train Faster-RCNN with your own data, tensorflow version (2)
http://blog.csdn.net/lanyuelvyun/article/details/78094003
pycocotools/_mask.so: undefined symbol: _Py_ZeroStruct error report on coco dataset
http: //blog.csdn.net/ab0902cd/article/details/79085797
protobuf 3.1.0 installation example and usage guide description
http://blog.csdn.net/mircosheng/article/details/70141704
/autogen.sh: 4: autoreconf: not foundhttps://blog.csdn.net/ww506772362/article/details/42265381
protobuf installation and uninstallation (mainly)
http://blog.csdn.net/tuhuolong/article/details/78353519
Install google protobuf under linux (change the installation location)
https://blog.csdn.net/xiexievv/article/details/47396725
protobuf installation and use
http://blog.csdn.net/csqingchen/article/details/50669725
Caffe training error: malloc(): memory corruptionhttp://blog.csdn.net/u011070171/article/details/52846306
Ubuntu16.04+CUDA9.1+cudnn-..-v7 environment install caffe runtest problem
http:/ /blog.csdn.net/pzh16789/article/details/78913107
BatchReindexLayer fails GPU gradient tests under CUDA v9.1 #
https://github.com/BVLC/caffe/issues/6164
tensorflow tf.python_io module source code read
http:/ /blog.csdn.net/fengwuxuan/article/details/77482404
Ubuntu + CUDA9.0 + tensorflow-gpu installation processhttp://blog.csdn.net/qq_35976351/article/details/79325476 Installation
method
of OpenCV in Anaconda environmenthttp
://blog.csdn.net/u011501388/article/details/76320974
ImportError: dlopen: cannot load any more object with static TLS
http://blog.csdn.net/u012609509/article/details/78432323
[Tensorflow] Plant recognition based on the inception model under the slim frameworkhttp
:
//blog.csdn.net/wlzard/article/details/77689311 FastRCNN trains its own data Set (1 compilation configuration)
http://www.cnblogs.com/louyihang-loves-baiyan/p/4885659.html
fast-rcnn trains its own data set and demo code interpretation and summary (for fast-rcnn beginners)
http: //blog.csdn.net/u010668907/article/details/50991664
sergeyk/selective_search_ijcv_with_python
https://github.com/sergeyk/selective_search_ijcv_with_python
selective search generates .mat files
https://www.cnblogs.com/ymjyqsx/p/6959328 .html
python opencv entry Hough circle transform (28)
http://blog.csdn.net/tengfei461807914/article/details/77507820
facebookresearch/Detectron
https://github.com/facebookresearch/Detectron
caffe learning (seven): use py-faster-rcnn to train the voc2007 dataset (Ubuntu)
http://blog.csdn.net/hongbin_xu/article/details/77278329
faster rcnn Error: TypeError: slice indices must be integers or None or have an __index__ method
http://blog.csdn.net/qq_27637315/article/details/78849756
Python script runs a syntax error: IndentationError:unexpected indent
https://www. cnblogs.com/gary-bao/p/5142102.html
Python image cropping example code (such as avatar cropping)
http://www.jb51.net/article/116789.htm
TypeError: slice indices must be integers or None or have an __index__ method
https://www.jianshu.com/p/9137efc1bd8f
File "/home/scau2/Downloads/py-faster-rcnn/tools/../lib/datasets/voc_eval.py", line 148, in voc_eval
BB = BB[sorted_ind, :]
IndexError: too many indices for array
https://blog.csdn.net/lhanchao/article/details/70807912
ImportError: numpy.core.multiarray failed to import
https://www.jianshu. com/p/9137efc1bd8f
source installation tensorflow Cuda Configuration Error: Cannot find libdevice.10.bc under /usr/local/cuda-7.5
INFO: Elapsed time: 0.071s
FAILED: Build did NOT complete successfully (0 packages loaded)
currently loading: tensorflow/tools/pip_package
https://blog.csdn.net/hedongya/article/details/79671469
TensorFlow learning 1: source code installation (reference) combined with tensorflow combat books (tensorflow has also been installed for a long time, the previous installed version has always been the cpu version, Although the GPU version is downloaded, the CPU is used to run the experiment, because conda or pip is used for installation, and there are many potential version problems. There are not so many problems with source code installation, but python and pip are mainly used. Whether the installation location is the same)
https://blog.csdn.net/u013832707/article/details/73161071
How to switch between Anaconda's python and the system's own python The installation is successful on cuda9.1cudnn7.0.5, but it always fails to import tensorflow, showing that there is no such module. The key reason is that after the manager has successfully installed countless times and failed to import, as for the discovery, it is because the last step is to use pip to install it in the system. With python directory, and the default Python uses Python under anaconda2, so it causes an error) /usr/bin/python
https://blog.csdn.net/teeyohuang/article/details/79076821
https://www.zhihu .com/question/48330791
facebookresearch/Detectron https://github.com/facebookresearch/Detectron
install Caffe2 https://caffe2.ai/docs/getting-started.html?platform=ubuntu&configuration=compile
Install caffe2 with anaconda2 under ubuntu without root privileges
https://blog.csdn.net/xuluohongshang/article/details/79280100
Repair in ubuntu16.04 grub rescue mode http://www.cnblogs.com/sddai/p/7214330.html
Ubuntu16.04 boot bug fix https://blog.csdn.net/abcsunl/article/details/72875983
(The disk space in ubuntu 16.04 is full of error logs.) I modified the permissions of usr before, which made it impossible to enter the boot. I guessed a lot of reasons and tried many methods. I encountered a stuck boot before because of a graphics card driver conflict. This time I almost have to uninstall and reinstall caffe again. . . It turned out that the usr permission was set to 777 by me to copy the file in, and then the boot entry was modified after booting, but the boot could not be booted because the space was occupied by the error log. . . During the period, it also appeared that after repairing grub, the wireless network could not be searched, and after restarting a few times, the system could not be entered. Now it seems that the exponential growth of the log file after the usr permission has been modified causes the system disk to explode. The solution to the sudo command that I searched before was an unusable method that led to the pit. . . https://blog.csdn.net/yiscola/article/details/61638012
sudo:must be setuid root problem https://www.linuxidc.com/Linux/2013-06/86245.htm
Caffe2 - Initial use of Detectron https://blog.csdn.net/zziahgf/article/details/79142978
Preparation of own dataset before using Mask RCNN
https://baijiahao.baidu.com/s?id=1593798547829047940&wfr=spider&for=pc
labelme https://github.com/wkentaro/labelme
[Object detection] Install the image labeling tool labelme under Windows (install labelme to use) ( key )
https://blog.csdn.net/disiwei1012/article/details/79508954