This mini-project are composed of three parts.
- data
Please go to these website to download data: https://www.kaggle.com/c/facial-keypoints-detection/data
trainning data: 8832 96*96 gray-scale images
test data: 1783 images
IdLookupTable: tells you which point they need in the test data. Be careful, you are not going to get all of the 30 points in one images. It depends on the information on IdLookupTable.csv.
- models/code
fully_running.py
run it on ubuntu:
for example: python fully_running.py trainning
python fully_running.py test
Be careful, before running it, please change the route of data. It should be on the top of python code.
- configuring amazon GPU instance
please run setup.sh, just type setup.sh on the ubuntu, it will work automatically
baidu onlin disk : 链接: http://pan.baidu.com/s/1b6Oooe 密码: 4ype They are PPT and a record of vedio.