Hi, I'm trying to train the pytorch version model. However, in the "/code/img_to_text.py" file, line 28: coco_dataset_path = "./coco/coco_dataset_full_rm_style.json",
the json file is not provided.
It seems to me that, as your paper indicated, this file includes sentences after preprossing.
Could you please release preprocessing codes for generating such json file? Thx!
Hi, I'm trying to train the pytorch version model. However, in the "/code/img_to_text.py" file, line 28:
coco_dataset_path = "./coco/coco_dataset_full_rm_style.json",the json file is not provided.
It seems to me that, as your paper indicated, this file includes sentences after preprossing.
Could you please release preprocessing codes for generating such json file? Thx!