aimagelab / meshed-memory-transformer

Meshed-Memory Transformer for Image Captioning. CVPR 2020

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

work with other datasets

Akbari-A opened this issue · comments

how to train your model for real images instead of features prepared for COCO dataset? for example, using Flickr8k images and annotations to train this model

Additionally, is there any chance for us to test the model with new datasets or custom images? Do we need to extract features of the image(s) first manually and then use these features to test the model?

I have the same question as that of @barisbatuhan.
Any idea @baraldilorenzo @marcellacornia @MatteoStefanini Would really appreciate your help

I have the same question.I dont know how to make .hdf5