DeepHash-pytorch icon indicating copy to clipboard operation
DeepHash-pytorch copied to clipboard

Implementation of Some Deep Hash Algorithms, Including DPSH、DSH、DHN、HashNet、DSDH、DTSH、DFH、GreedyHash、CSQ.

Results 12 DeepHash-pytorch issues
Sort by recently updated
recently updated
newest added

Hi, swuxyj. Nice work for this community. It is noted that the training loss of DTSH contains a for loop which is somewhat time-consuming. Is there any change to speed...

拿计算mAP@3为例, case1: [+ - +] AP@3= (1 + 2/3) / 2 = 5/6 case2: [+ - - ] AP@3 = (1) / 1 = 1 有两个case,case1 返回两个正样本,case2 返回一个正样本,但是计算出来的AP@3,case1却低于case2, 这合理吗?

您好,非常感谢您的贡献! 在跑实验的过程中,发现coco数据集的训练集和原始HashNet不一致。具体地,比如COCO_val2014_000000480663.jpg,没有出现在原始HashNet的训练集中,可以帮忙检查一下吗,非常感谢啦~

非常感谢您的工作。我想问一下为什么贝叶斯网络这一系列方法,hashnet,DCH等,都达不到那个结果,请问这跟使用什么GPU有关系嘛,我看您复现的结果是相近的,我也用了您的代码,但是还不行,是有什么细节嘛??

dataset.txt the file composition is: image name + binary label? How do I get a binary label? How do I define the file .txt make dataset when using my own...

(Database0 represents unique images that are not in the train set) When we test our model, we assume that it corresponds to a real-world situation when we can train seen...

非常感谢你的project,实在太棒了。有个问题请教,Greedy Hash是基于交叉熵目标分类损失,但是NUS和COCO数据集是多标签图像,如何训练Greedy Hash,看到你给了结果,但不知道怎么训练?

您好,非常感谢您的贡献,请问imagenet数据集的train.txt、test.txt和database.txt之间有什么样的关系,数量比例是多少,我想构建一个imagenet格式的自己的数据集,可是程序会报维度的错误,只有当训练、测试和数据库都用同样的数据集才不报错,期待您的回复,谢谢!

Line 163-165 for bit in config["bit_list"]: train_val(config, bit) orror for obj in iterable: TypeError: 'int' object is not iterable