Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Missing wordnet files for processing imaginet data #65

Closed
brenowca opened this issue May 3, 2021 · 6 comments
Closed

Missing wordnet files for processing imaginet data #65

brenowca opened this issue May 3, 2021 · 6 comments
Labels
question Further information is requested

Comments

@brenowca
Copy link

brenowca commented May 3, 2021

Hi, thanks for sharing the work and for, well the amazing work.

I am trying to pre-process the ilsvrc_2012 data but it seems that two links (for the wordnet info) are no longer available. I searched for them in the image-net.org webpage, but with no success.

Download the following two files into ILSVRC2012_img_train/:
http://www.image-net.org/archive/wordnet.is_a.txt
http://www.image-net.org/archive/words.txt

Do you have any other links to access those files? (Or can you share them somewhere else if it does not violate any license?)

Thanks in advance for your time.
And thanks again for sharing the dataset, I believe it is an important step towards more practical meta-learning.

Best wishes,
Breno

@brenowca
Copy link
Author

brenowca commented May 3, 2021

The instructions I quoted is from the file dataset_conversion.md#ilsvrc_2012.

@lamblin lamblin added the question Further information is requested label May 4, 2021
@lamblin
Copy link
Collaborator

lamblin commented May 4, 2021

It appears the ImageNet website has changed the organization of its files.

If someone has a valid login to image-net.org, maybe they could try to see if these files are available to download in a different way and report it, we would be happy to update the instructions.

I am not sure whether we have the permission to share them directly, but some copies seem to exist on the Internet.

@brenowca
Copy link
Author

brenowca commented May 4, 2021

I got a login. It is an "external researcher who agreed with the terms of access" kind of login.

But still I was not able to find it.

This is the file structure that I have access so far:

ImageNet download page

ImageNet Large Scale Visual Recognition Challenge 2012 (ILSVRC2012) download page

@lamblin
Copy link
Collaborator

lamblin commented May 4, 2021

Thanks for the update.
If you would like the official source, I suggest you email the maintainers of image-net.org (imagenet.help.desk@gmail.com) and ask them about it.
Otherwise, the ones you can find with https://www.google.com/search?q=wordnet.is_a.txt seem to be correct.

@brenowca
Copy link
Author

brenowca commented May 5, 2021

Hey, I got a reply from them and those are the new paths for the files:

Basically we only got to replace "archive" with "data" and all is good.

@lamblin
Copy link
Collaborator

lamblin commented May 5, 2021

Thanks for figuring it out!
I'll make the update.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
question Further information is requested
Development

No branches or pull requests

2 participants