Oscar issues

Can you share the full NoCaps results on the test data?

1

The OSCAR paper reports a CIDEr score of 78.8 and 80.9 for OSCAR base and large, resp. Since it's not clarified, I assume these are scores on the NoCaps test...

YovaKem

gqa_bal_qla_testdev.json and gqa_all_qla_testdev.json missing

2

Hey, Thanks for this amazing repo! I have downloaded the preprocessed GQA dataset with the two different strategies: ``` wget https://biglmdiag.blob.core.windows.net/oscar/datasets/gqa.zip unzip gqa.zip -d ./ ``` and ``` path/to/azcopy copy...

JoPfeiff

VinVL features for datasets not available

2

Hi there, Thanks a lot for your code release. I noticed that the VinVL features are not available anymore: https://github.com/microsoft/Oscar/blob/master/VinVL_DOWNLOAD.md#pre-exacted-image-features Could you please advise?

aleSuglia

Tagging Apex version

9

INSTALL.md recommends installing Apex#master. However, Apex does not build any more against pytorch 1.2.0 (see https://github.com/NVIDIA/apex/issues/956). So which version of apex is known to work?

benob

How are the files in "train_logs/" and "test_logs/" encoded?

1

Hi, I download the pre-trained models as indicated in the "VinVL_DOWNLOAD.md". When I tried to view the files in "train_logs/" and "test_logs", like the file "events.out.tfevents.1605353916.container-e2250-1602032654055-124610-02-000003", I found it was...

yx3266

How to run captioning on any image? E.g. how to prepare the test.yaml and other required files required by the run_captioning.py?

11

The image contains classes seen in coco captioning dataset but I do not know how to extract the features for captioning using oscar+

aliencaocao

what is checkpoint-29-66420

2

From this page https://github.com/microsoft/Oscar/blob/master/MODEL_ZOO.md, For Oscar on image captioning task, there is only one checkpoint model: checkpoint-29-66420. This checkpoint model is trained by only cross-entropy or it's finetuned by cross-entropy...

Vincent-Ww