Kevin Ke-Yun Lin

W is the network parameters. BTW, this repo does not support multi-label version.

We didn't observe this problem. We used the same codes and model converged.

Thanks for pointing out this problem. I am confused because I cannot reproduce this error..

@lucmichalski Thank you for your suggestion. We don't have such docker. We will create one and make it available in future. Since our repo doesn't require additional dependencies, you may...

Unfortunately this project was developed before the birth of tensorflow and pytorch, thus we didn't plan to develop our algorithm using the two tools. Currently we do not have an...

Since this caffe is an old version, it doesn't support batch normalization. One of the possible solutions is adding our loss functions to a newer caffe. Then, you can modify...

Sorry for the late reply. I am too busy with current projects. If you want to generate binary codes quickly and avoid additional implementations, you can just train the model...

Thank you for this!

@RishabGargeya: Current github implementation does not scale the objective function. We have conducted the parameter selection and study the influence of individual terms in the learning objective, but our latest...

How about removing your loss function, and training with only the softmax loss function at first? 2017-07-19 23:25 GMT-07:00 Sarthak Yadav : > Hello > > I am trying to...