Clone the git repo:
$ git clone git@github.com:ttseriotou/image-captioning.git
Create a conda environment:
$ conda env create --file=image_captioning.yml
Actviate the conda environment
$ conda activate image_captioning
The notebook defaults to the git-large-coco image captioning model for inference. There is though a wide range of different image captioning models in Huggingface. Examples of such models are enlisted here
Flickr Sample Data were obtained through: https://www.kaggle.com/datasets/hsankesara/flickr-image-dataset
Credits to the following sources for processing and GIT captioning: https://www.kaggle.com/code/hsankesara/image-captioning