/Automatic-Image-Captioning

A neural network architecture(CNN+LSTM) that automatically generates captions from the images. The model uses ResNet architecture to train the Encoder while DecoderRNN has to be trained with our choice of trainable parameters. I have trained the model on the Microsoft Common Objects in COntext (MS COCO) dataset and have tested the network on fictitious images!

Primary LanguageJupyter NotebookMIT LicenseMIT

Stargazers