[CVPR23] A cascaded diffusion captioning model with a novel semantic-conditional diffusion process that upgrades conventional diffusion model with additional semantic prior.
I run tools/create_feats.py, but it throw out this problem. I turn the mode into read text, but still not work.
Could you please tell me what is wrong?
Hi, @catfish132 ,
tools/create_feats.py is modified from the read_tsv.py in the original BUTD repo. I suggest you to run this script in python2, please refer to this issue for more information.
I run tools/create_feats.py, but it throw out this problem. I turn the mode into read text, but still not work. Could you please tell me what is wrong?