search

devaansh100 / CLIPTrans

Official implementation for the paper "Transferring Visual Knowledge with Pre-trained Models for Multimodal Machine Translation", published at ICCV'23.
19 stars 3 forks source link

Datasets #3

Open Roaxy opened 5 months ago

Roaxy commented 5 months ago

Hi,Thanks for your contribution! I am very interested in your paper，but i cannot download images_mscoco.task1.tar.gz for the MSCOCO images,I'm not sure if you can provide the dataset directly？Looking forward to your answer very much！

devaansh100 commented 5 months ago

Thanks for your interest in our work! It seems that the link on statmt is no longer working. You can get the corpora from this repository.

Regarding the raw images, let me get back to you in some time.

Roaxy commented 5 months ago

Thanks for your reply,I anticipate your follow-up progress extremely.

dhawan98 commented 3 months ago

did you find where to download the tar.gz file for mscoco(images_mscoco.task1.tar.gz)? Also, does the data folder need to have only the tar files, or the whole image dataset as well?

devaansh100 commented 3 months ago

Hi, unfortunately I have not been able to get in touch with the authors of the dataset. What you can do is fetch the images from the MSCOCO dataset (since we already have the names of the files).

The data directory only needs the tarballs - run the corresponding bash scripts to setup the dataset. That should extract them