Giter Site home page Giter Site logo

ttkunt / image-captioning-with-keras Goto Github PK

View Code? Open in Web Editor NEW

This project forked from rahulrt7/image-captioning-with-keras

0.0 1.0 0.0 64.45 MB

An Image Caption generator based on google's Show and Tell (2015) paper

Jupyter Notebook 62.35% Shell 0.01% Python 0.96% HTML 36.68%

image-captioning-with-keras's Introduction

Image-Captioning-with-keras

An Image Caption generator based on google's Show and Tell (2015) paper

Dependencies

  • keras (2.0) [use tensorflow backend]
  • tensorflow (1.0)

Running the code

Executing the Final+Model.py file will run 100 epochs over the whole MS COCO dataset. The data can be downloded by running .sh scripts provided in dataset folder. If you don't have a GPU the computation might take too long. You can disable callbacks (for logging tensorboard files and checkpoints to save weights) to reduce computation time. For disabling callbacks just remove callback parameter from the call model.fit().

Files:

  • Final model training : HTML file downloaded from actual .ipynb file after training on complete MS COCO dataset. This file includes training loss and validation loss after each epochs. Total Epochs around 130 and maximum accuracy 98.64. Time taken to train the model -> approx 7 hours.
  • Final Model : Final code where predictions are made for images in validation data. Probabitlies are also computed for all words in a caption(caption_length = 17). Predicted and Actual capiton is displayed alongside each other just below the funciton call. Image and its corresponding 5 captions are also displayed.
  • Final+Model.py: Python code for FinalModel.ipynb notebook
  • GPUstats: .png image file displaying a screenshot while training RNN on gpu. Can give an idea on GPU memory, GPU graphics and CPU usage on training time.

GPU used for training

Single Nvidia GTX 970

Reserch Paper Reffered

Show and Tell: A Neural Image Caption Generator Oriol Vinyals, Alexander Toshev, Samy Bengio, Dumitru Erhan (Submitted on 17 Nov 2014, last revised 20 Apr 2015)

image-captioning-with-keras's People

Contributors

rahulrt7 avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.