search

lukemelas / image-paragraph-captioning

[EMNLP 2018] Training for Diversity in Image Paragraph Captioning
90 stars 23 forks source link

pre-trained model #9

Open volquelme opened 5 years ago

volquelme commented 5 years ago

Thanks for your studying and sharing! I am just wondering what dataset is used for pre-trained model Which one(MS-COCO or Visual Genome) is used? if MS-COCO used, can I get the model pre-trained on visual genome dataset?

arjung128 commented 5 years ago

The Visual Genome was most likely used to train the given pre-trained model as the Visual Genome provides captions which are of paragraph-length, while the MS-COCO dataset does not.

volquelme commented 5 years ago

ah... i got it thanks for your reply