Skip to content

Latest commit

 

History

History
12 lines (7 loc) · 721 Bytes

File metadata and controls

12 lines (7 loc) · 721 Bytes

Story-Telling-Using-Show-Attend-and-tell

augmenting the show attend and tell image captioning with story telling

implement paragraph by using https://github.com/tensorpack/tensorpack/tree/master/examples/FasterRCNN

Files and what they do

GetDataset.ipynb : Split the dataset into the training and test and valuation pathfiles and save it to corresponding files "imgs_test_path.txt"

MakeParagraphandimage.ipynb : take the paragraph json file and map the dataset to the given paragraph "img2paragraph"

ParagraphPreprocessing.ipynb : take the paragraph json file and build the vocabulary and then from these built vocabulary append padding, bos and eos and unknown characters and build idxtoword and wordtoindex