danieljf24

I guess the file named vec500flickr30m.tar.gz (3.0G) has not been downloaded completely.

What's the version of pytorch you used? We recommend the environment with PyTorch 0.3.1, Python 2.7.

Sorry, I just trained the model on 1 GPU. The results you posted are interesting. I think it may be caused by the triplet loss with hard example mining. Additionally,...

I checked the evaluation using the provided trained model and found no problem. Maybe the software version leads to the problem. Did you use the pytorch 0.3.1?

I also try using pytorch 0.4, but the result is still the same as reported in the readme. I recommend you use the environment suggested in [requirements.txt](https://github.com/danieljf24/dual_encoding/blob/master/requirements.txt) first. If there...

The script do_test_dual_encoding_msrvtt10ktest.sh is generated automatically after training, have you trained the model successfully?

Hi,andyyuan78, I met the same problem, have your fixed it?

Sorry for the late reply. The data can be downloaded from [Google Drive](https://drive.google.com/open?id=1AgFzl7OvfmEQAOYzdbNbbL2r9ox6TpJA) and [Baidu Pan](https://pan.baidu.com/s/1aoGDxmEUH6XRSigQ64KFTA).

What do you want to visualize? If you would like to visualize tags, we used the Tag cloud toolkit.

I have tried to access all the baiduyun URL, and found the password is correct. Cloud you give which baiduyun URL you can not access?