Deep Learning
文章平均质量分 60
js_sjtu
这个作者很懒,什么都没留下…
展开
-
Caffe安装要点:Ubuntu 16.04+双GTX1080+Opencv3.3+Anaconda3
包含Ubuntu 16.04+双GTX1080+Opencv3.3+Anaconda3环境下caffe的安装要点和报错解决方式。安装已完全成功,可顺利运行Caffe。本教程里均为绿色安装,即不使用sudo命令。维护环境养成良好习惯从我做起:)原创 2017-11-24 17:09:51 · 244 阅读 · 0 评论 -
多GPU环境下使用caffe遇到的问题
两张GTX 1080 Ti, ubuntu sever 16.04需要nccl的支持。但是nccl的test跑不通,曾在github上提交相关issue:https://github.com/NVIDIA/nccl/issues/120现已解决。原创 2017-11-24 21:59:26 · 1075 阅读 · 0 评论 -
安装最新版tensorflow(r1.5)遇到的各种坑
坑1. r1.5版本的tf若从源码编译,须用pip另外安装tensorboard: pip install tb-nightlytensorboard --logdi="./TFlog" 似乎不能识别~符号否则会有如下报错:ImportError: cannot import name 'run_main'下面这个解决方式我没试过:"python /usr/local原创 2018-01-03 16:40:45 · 2182 阅读 · 0 评论 -
吴恩达课程深度学习错题集
1.Logistic regression’s weights w should be initialized randomly rather than to all zeros, because if you initialize to all zeros, then logistic regression will fail to learn a useful decision boundar...原创 2018-02-08 10:42:17 · 723 阅读 · 0 评论 -
吴恩达课程深度学习笔记
1. Train / Dev / Test setsThen traditionally you might take all the data you have and carve off some portion of it to be your training set. Some portion of it to be your hold-out cross validation set,...原创 2018-02-08 11:22:52 · 313 阅读 · 0 评论