Van-Quang Nguyen

Please find here for all the files. In fact, you don't need `train_ann_lmdb`. Just check the code of Dataset and turn off them. https://drive.google.com/drive/folders/1c7eWTjrlo_UJuKH3GwHIvTYMwzTkbIRd?usp=share_link

I think the paths to images are wrong. Please check it carefully.

@auh-12 You are right. The pretrained object detectors shared in this repo were finetuned on OD and Attribute prediction using VG. I think you can find the code in the...

Thanks for your question. There are two main purposes: 1. To show that trade-off between finetuning the entire model (including the backbone) and finetuning the model partially (excluding the backbone)....

Thanks for asking! May you find the evaluation code/notebook from M2 Trasnformer repo? I remember that I took the code from there.

Sorry for my bad memory. Luckily, I've just checked the old code and found that the evaluation based on this repo, not M2 transformer: https://github.com/salaniz/pycocoevalcap. To get 'Object Attr. Relation...

I can't find my old code for CLIP score. However, you can easily follow this repo to compute: https://github.com/jmhessel/clipscore

Thank you for asking. I am not sure which errors you encounter. * `itos` is a short name of `index_to_string` (or `to_token`). * Similarly, `stoi` is a short name of...

I think in this case, you need to do a little bit effort in the generate code. For example, if the logit is the highest value, then choose the second...

May you provide the structure of the annotations you have? BTW, I have an example to train on custom dataset here: https://github.com/davidnvq/grit#finetune--retrain-grit-on-your-own-dataset