search

shuangshuangguo / caffe2pytorch-tsn

Transform the caffe model to pytorch model for Temporal Segment Network
73 stars 16 forks source link

Mapping of label #19

Open wsy19961129 opened 5 years ago

wsy19961129 commented 5 years ago

hello, I am just wondering how do you map the label to one-hot encoding. Is it just sort the label(string) in alphabetical order? E.g. in UCF101, ApplyEyeMakeup=0 and ApplyLipstick=1. I test the model using your pretrianed weights while on UCF101/split1 is 76% and on HMDB51 is about 43.3% and I do not know what the problem is. Thanks!

linshuheng6 commented 5 years ago

@wsy19961129 I got the same problem that the accuracy on UCF101/split1 is only 75.53%. Have you solved this problem?