search

Seth-Park / MultimodalExplanations

Code release for Park et al. Multimodal Multimodal Explanations: Justifying Decisions and Pointing to the Evidence. in CVPR, 2018
BSD 2-Clause "Simplified" License
49 stars 10 forks source link

Dataset files #12

Open OanaMariaCamburu opened 4 years ago

OanaMariaCamburu commented 4 years ago

Hi,

Could you please what are the files v2_mscoco_train2014_annotations.json and the v2_mscoco_val2014_annotations.json meant to be in Annotations/ and the v2_OpenEnded_mscoco_train2014_questions.json and v2_OpenEnded_mscoco_val2014_questions.json files from the Questions/ ?

I couldn't find these exact names in the links mentioned and I want to make sure I'm not using incorrect files if I just guess myself what they would be. Also, to be sure, should visual/val/ and visual/test be empty in case we retrain the model from scratch or use the files you in https://drive.google.com/drive/u/0/folders/1Cr9JRXDmjks_wmi-a9eIe4SWSwWKcCk7 ?

Thanks, Oana

Seth-Park commented 4 years ago

The json files can be downloaded in https://visualqa.org/download.html. They are json files that contain question and annotations for the VQA task. In the above link, you can find good overview of how the files are formatted.

Also, the visual/val and visual/test files should be downloaded from the google drive link. They are used for evaluation purposes so they are not going to be used for training.

OanaMariaCamburu commented 4 years ago

Thanks, Seth-Park.

In https://visualqa.org/download.html I can only see the versions of 2015 and 2017, and you mentioned the 2014 version in the repo. I could find some 2014 files here https://www.floydhub.com/sominw/datasets/vqa_data/3/v2, are they the correct ones?