ImageCaptioning-Pytorch

A minimal pytorch image captioning system trained on Flickr8k dataset

Dataset Link

caption.py

python caption.py --root 'test_img.jpg'

Argument	Parse	Default
Number of words to generate	--n	25
Hidden Size	--hidden_size	512
Image to Caption	--root

Pretrained model was too heavy to upload

Output

a dog is running through a field of grass and flowers .

a man in a red jacket and black pants is standing next to a man in a red jacket .