Copyright: Attribution, allow others to create paper-based, and must distribute paper (based on the original license agreement with the same license Creative Commons )
So far, the training data set has been used voc_2007 better faster-rcnn model, try the following with their own data sets to train faster-rcnn model.
First, the environment configuration:
Ubuntu 16.04 Linux, GTX1080 Ti graphics card GeForce;
tensorflow-GPU 1.3.0, Cuda8.0, Cudnn 6.0;
Python 3.5.4
Second, the production data collection
The category pascal_voc.py line 25 lines of code changed their category
folder to open the terminal where the picture file, and then enter the following line of code in the command line can batch will be converted to .jpg format pictures .JPG format.
ls *.JPG|sed -r 's#(.*).JPG#mv & \1.jpg#'|bash
