TypeError: 'numpy.float64' object cannot be interpreted as an index

ubuntu 16.04 cuda8.0 cudnn5.0

faster_rcnn出现的问题:TypeError: 'numpy.float64' object cannot be interpreted as an index
Solution: Try to install numpy 1.11.0: sudo pip install -U numpy==1.11.0
安装完numpy之后,首先删除py-faster-rcnn文件,然后在github上再重新下载py-faster-rcnn文件

按照github上步骤重新编译。

(经验:编译faster-rcnn出现问题,尽量删除之前编译好的,重新下载py-faster-rcnn)

./experiments/scripts/faster_rcnn_end2end.sh 0 VGG16 pascal_voc
./experiments/scripts/faster_rcnn_alt_opt.sh 0 VGG16 pascal_voc

你可能感兴趣的:(caffe)