search

clin1223 / VLDet

[ICLR 2023] PyTorch implementation of VLDet （https://arxiv.org/abs/2211.14843）
Other
179 stars 11 forks source link

Can not get the dataset. #6

Closed JiuqingDong closed 1 year ago

JiuqingDong commented 1 year ago

You suggest following OVR-CNN to create the open-vocabulary COCO split. However, I cam not got coco/ zero-shot/ instances_train2017_seen_2.json instances_val2017_all_2.json

Besides, I didn't see 'coco_65_concepts.txt' in your Google driver.

Could you provide these files? Thank you very much.

JiuqingDong commented 1 year ago

We look forward to your directly providing these files, which will be very helpful for us to reproduce your work, and will also make your work more influential.

Besides, I would appreciate it if you could provide some description about the config file.

clin1223 commented 1 year ago

Hi,

Sorry, can you provide the error information when you generate json file following https://github.com/alirezazareian/ovr-cnn/blob/master/ipynb/003.ipynb

'coco_65_concepts.txt' is not necessary for running the code.

Thanks!

--Chuang

QiueY514 commented 1 year ago

Hi,

Sorry, can you provide the error information when you generate json file following https://github.com/alirezazareian/ovr-cnn/blob/master/ipynb/003.ipynb

'coco_65_concepts.txt' is not necessary for running the code.

Thanks!

--Chuang

Hi! Could you release the files instances_train2017_seen_2.json and instances_val2017_all_2.json? There is something wrong when I try to generate these files.