karansikka1/Foodx

Cannot map integer class ids to class texts

Opened this issue · 3 comments

I've noticed that a mapping between class_ids and class_texts is not provided.

In the Data Download and Format section, it's said that:

  • The tar contains 4 files
    • class_list.txt: Contains the names of 211 class labels. This can be used to map class_ids with class names.

However, the annot.tar file actually only contains three files and we cannot find a class_list.txt. The class_list.txt in this repo is just an alphabetically sorted list of class text and it's not the correct mapping.

For instance, train_059329.jpg has a class_id of 0. The first line of class_list.txt is adobo but the food in that image is clearly a macaron.

Dear derekhh

Thanks for pointing out the issue. I have updated the class_list.txt on the Github and will shortly include that again in the annotations.tar. This new file has the right mapping. Let me know if there is any other problems.

Thanks
Karan

Thanks Karan! One more question - it seems the iFood challenge on Kaggle is set up as "
InClass Prediction Competition". This seems to have reduced the potential number of participants by a lot. Is this intentional?

That is a valid question. We decided to have this competition a little late and with all the prep needed we did not have much time to inform Kaggle about the competition in time (which is booked prior to CVPR). This is our first edition of food challenge and we plan to host it on Kaggle from next year. Hope to have your participation this year.