《学习&&工作笔记》Che_Hongshu个人笔记1

更改学习笔记2019.1.17
没用的全部清理掉接下来每天的安排是（因为最近是寒假在家）
保持一天或者两天一篇paper reading
每天背200个单词
每天读英语听英语各30分钟，精听
写代码
重新整理基础numpy,pandas,
特征工程
打比赛，阿里
pro
博客笔记
tf和pytorch

最近在公司做项目，具体过程，自己笔记本装个双系统win/linux，之后装caffe_ssd，下载Mobilenet，熟练掌握MobilenetSSD的使用，制作VOC数据集，训练model，运用剪枝方式进行剪裁，缩小加速model，最后用腾讯ncnn前向推理框架进行移植MobilenetSSD的model，以下是过程中出现问题解决问题的正确打开方式，和很多有一定参考价值的文章。

还在继续做当中，挖个坑，等优化ok掉，速度和大小都比较好的情况下，再来从头到尾进行一波个人笔记。
-----------------------分界线--------------------------------------------------
完成了mobilenetssd的ncnn前向推理框架的android移植，效果和预想有一定差距用 不同的线程最快差不多一帧300ms左右(问题在于打开方式不对，自己的图片预处理出现问题了，已解决了)，准备尝试tensorflow深度学习框架tflite的移植，目前已跑成功mobilenetv1的识别demo 速度差不多100ms，准备做mobilenetssd的tf版本

2018.11.3

2018.11.5

Pro1:怎么进入adb shell模式
https://blog.csdn.net/number1killer/article/details/79192914
Pro2:mobilefacenet-mxnet
https://github.com/moli232777144/mobilefacenet-mxnet
Pro3:cmake- release
https://blog.csdn.net/u012525096/article/details/78950979
Pro4:ncnn -benchmark使用
https://swift.ctolib.com/article/wiki/85269
Pro5:adb use
https://blog.csdn.net/henni_719/article/details/62229263
ubuntu如何切换到root用户

2018.11.6

Pro1:一文读懂VGG网络
https://blog.csdn.net/amusi1994/article/details/81461968
Pro2:mobilenet网络的理解
https://blog.csdn.net/wfei101/article/details/78310226

2018.11.7

Pro1:Could not install packages due to an EnvironmentError: [Errno 13] 权限不够: ‘/usr/local/lib/python2.7/dist-packages/astor’
Consider using the --user option or check the permissions.
https://blog.csdn.net/a781751136/article/details/80231406
Pro2:failed to get default registry endpoint from daemon (Got permission denied while trying to connect
https://blog.coderstory.cn/ubuntu下docker的常见安装方式/
Pro3:linux杀死进程
https://www.cnblogs.com/liutao1122/p/8268932.html

2018.11.8

Pro1:配置 docker 的tensorflow-gpu的环境
https://blog.csdn.net/qq_37341836/article/details/80303357
Pro2:docker 退出容器
https://blog.csdn.net/leoe_/article/details/78685186
Pro3:docker 进入容器
https://blog.csdn.net/u012124564/article/details/70230358
Pro4:docker 删除容器
https://blog.csdn.net/qq_32447301/article/details/79387649
Pro5:docker 安装tensorflow(更全)
https://blog.csdn.net/wc781708249/article/details/78787994
Pro6:Win10下用Anaconda安装TensorFlow
https://blog.csdn.net/u010858605/article/details/64128466
Pro7:pycharm更换为Anaconda中另外建立的tensorflow的python的环境
https://blog.csdn.net/weixin_39223665/article/details/80706154
Pro8:完美解决 No toolchains found in the NDK toolchains folder for ABI with prefix: mips64el-linux-android
https://blog.csdn.net/qq_24118527/article/details/82867864

2018.11.9

Pro1:.Tensorflow Lite初探（Android）tflite demo运行 https://blog.csdn.net/bhneo/article/details/78835224

2018.11.11

Pro1:量化的算法笔记
https://note.youdao.com/share/?id=829ba6cabfde990e2832b048a4f492b3&type=note#/
Pro2:量化的tools的使用
https://github.com/BUG1989/caffe-int8-convert-tools
Pro3:量化后如何使用.table文件运用ncnn转换
https://github.com/Tencent/ncnn/wiki/quantized-int8-inference

2018.11.14

Pro1:android导入library 项目时target pattern contain no % 的解决方法
https://blog.csdn.net/onerain88/article/details/6969235

2018.11.15

Pro1: linux下截图
https://blog.csdn.net/qq_39609900/article/details/79224347

2018.11.16

Pro1:caffeprortxt中初始化参数Xavier与MSRA的区别和对比
https://blog.csdn.net/Fire_Light_/article/details/79556192
Pro2:caffe参数介绍
https://blog.csdn.net/cyh_24/article/details/51537709
Pro3：caffe深度学习参数调节
https://blog.csdn.net/dongapple/article/details/75499390
Pro4:优化方法总结：SGD，Momentum，AdaGrad，RMSProp，Adam（优化器）
https://blog.csdn.net/u010089444/article/details/76725843
Pro5:深度学习-超参数调整总结
http://www.cnblogs.com/andre-ma/p/8676220.html
Pro6:caffe protxt soler.protext中学习率的策略的选择问题
https://blog.csdn.net/sunshine_in_moon/article/details/53563611
Pro7:目标检测衡量标准mAP
https://blog.csdn.net/jay463261929/article/details/78734383
Pro8:tensorflow的model
https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md

---

git push git 中github clone “Permission denied (publickey). fatal: Could not read from remote repository.”
https://blog.csdn.net/chaoyueziji123/article/details/54669555
11: /etc/default/grub: splash: not found
http://forum.ubuntu.org.cn/viewtopic.php?t=278240## 2018.10.22
Por1:浙大blog嵌入式AI：
http://hey-yahei.cn/archives/
Dl(pytorch) repo:
https://github.com/bharathgs/Awesome-pytorch-list
Ok blog:
https://yikunhaocong.com
Pro1：Channel Pruning for Accelerating Very Deep Neural Networks 算法笔记
https://blog.csdn.net/u014380165/article/details/79811779
https://blog.csdn.net/zhongshaoyy/article/details/78551728?utm_source=blogxgwz1
Pro1:计算机视觉中山大学硕士博客
https://www.cnblogs.com/skyfsm/archive/2018/10.html
Pro1:Ubuntu14.04升级cmake版本的方法
https://blog.csdn.net/learning_tortosie/article/details/80593956
Pro2:Android NDK APP_STL gnustl_shared is no longer supported 报错处理
https://blog.csdn.net/strikedragon/article/details/83090105
Pro3:Android Studio常见问题 – uses-sdk:minSdkVersion 8 cannot be smaller than version 9 declared in library
https://www.cnblogs.com/kangyi/p/4409336.html
Pro1:解决linux下android studio真机运行问题xxx[null]
https://blog.csdn.net/qq_35244529/article/details/79754133
Pro2: 无法确定Bitmap签名
https://blog.csdn.net/GXH_APOLOGIZE/article/details/77930966
qt5.5程序打包发布以及依赖