willard-yuan / hashing-baseline-for-image-retrieval

:octocat:Various hashing methods for image retrieval and serves as the baselines
http://yongyuan.name/habir/
615 stars 255 forks source link

不同bit的结果差距过大 #21

Open LCH1238 opened 4 years ago

LCH1238 commented 4 years ago

您好,我用这个工具包测试CIFAR-10 VGG16特征的数据集,ITQ的16,32,64,128bit的mAP分别是0.1993,0.3397,0.4562,0.5218。因为不同位之间的mAP差距过大,我又找了ITQ的pytorch版本测试。同样配置,16,32,64,128bit的mAP为0.2993,0.3263,0.3425,0.3662。按理说不同位之间的结果并不会有几倍的差距才对,是因为matlab的缘故还是算法有错误?