search

WuJie1010 / Facial-Expression-Recognition.Pytorch

A CNN based pytorch implementation on facial expression recognition (FER2013 and CK+), achieving 73.112% (state-of-the-art) in FER2013 and 94.64% in CK+ dataset
MIT License
1.78k stars 548 forks source link

Training acc #66

Closed jcjavaismgood closed 4 years ago

jcjavaismgood commented 4 years ago

I retrained the model with my own data, i am surprised to find that model between your profile1-10,1 0th will be better than most of them,and output acc will look bad if sum them uo and average in confusion_matrix， i wonder the relationship of 10 folder model. .....appreaciate your guidance and help !!

WuJie1010 commented 4 years ago

They are generally independent and have different performance for different samples. So I used the integrated model to improve the accuracy.