Repo Contains the code for different hands-on demontrated.
- Download glove embedding from https://www.kaggle.com/datasets/subrata2019/glovedata and put it in the emojify folder before running the code.
- Emojify demo is in emojify_notebook.ipynb .
- Data will be dowloaded directly using pytorch library "from torchtext.datasets import Multi30k", in the seq2seq.py script
- Demonstartion can be found in .ipynb file
- Download the dataset used: https://www.kaggle.com/dataset/e1cd22253a9b23b073794872bf565648ddbe4f17e7fa9e74766ad3707141adeb Then set images folder, captions.txt inside a folder Flickr8k.
- train.py: For training the network
- model.py: creating the encoderCNN, decoderRNN
- get_loader.py: Loading the data, creating vocabulary
- utils.py: Load model, save model, printing few test cases downloaded online
- whole demo is given in demo.ipynb