Fei Wang

Hi What kind of network do you train? (attention-56 or attention-92?)

We train the attention-56 with memory optimization, and the usage of memory is about 6G with 32 images per card. I think you bn setting is not suitable for classification,...

hi, thanks for your attention to our work. We will release our trained model as soon as possible.

Thanks for your reminding. The interp layer register bug has been fixed.

I have also found this problem. The URL is crawled in the last year and maybe some URLs have been blocked. We cannot directly send the original images to the...

@hustzeyu we do not have the copyright of the original images.

@seesemichaelj thanks for your detailed explanation.

@zhongyy hi，could you tell me the number of invalid URL?

Yes. Thanks a lot.

We found that there are few error images, which may result from the data parsing progress. But the proportion of error data is very low.