HelloRicky123 / Siamese-RPN

Full reimplementation of siamese rpn, has 0.24 eao on vot2017.
MIT License
223 stars 44 forks source link

The trained model has too low accuracy. #57

Open lyp22 opened 4 years ago

lyp22 commented 4 years ago

I train the code only on VID dataset for 50 epoches and I don't modify the code. But the OTB AUC result is only about 0.28. And you said in another issue that the result on VID dataset should be about 0.53.