Keras-based LSTM/CNN models for Visual Question Answering
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is 19 commits behind avisingh599:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
3rdParty
data
features
models
results
scripts
.gitignore
LICENSE.txt
README.md

README.md

#Deep Learning for Visual Question Answering

Click here to go to the accompanying blog post.

This project uses Keras to train a variety of Feedforward and Recurrent Neural Networks for the task of Visual Question Answering. It is designed to work with the VQA dataset.

Models Implemented:

BOW+CNN Model LSTM + CNN Model
alt text alt text

##Requirements

  1. Keras 0.20
  2. spaCy 0.94
  3. scikit-learn 0.16
  4. progressbar
  5. Nvidia CUDA 7.5 (optional, for GPU acceleration)

Tested with Python 2.7 on Ubuntu 14.04 and Centos 7.1.

Notes:

  1. Keras needs the latest Theano, which in turn needs Numpy/Scipy.
  2. spaCy is currently used only for converting questions to a vector (or a sequence of vectors), this dependency can be easily be removed if you want to.
  3. spaCy uses Goldberg and Levy's word vectors by default, but I found the performance to be much superior with Stanford's Glove word vectors.

##Using Pre-trained models Take a look at scripts/demo_batch.py. An LSTM-based pre-trained model has been released. It currently works only on the images of the MS COCO dataset (need to be downloaded seperately). I do intend to add a pipeline for it to work for all images in general.

Caution: Use the pre-trained model with 300D Common Crawl Glove Word Embeddings

Do not the word embeddings which are the default spaCy embeddings (Goldberg and Levy 2014). If you try to use these models with any embeddings except Glove, your results woulf be garbage.

##The Numbers Performance on the validation set of the VQA Challenge:

Model Accuracy
BOW+CNN 44.30%
LSTM-Language only 42.51%
LSTM+CNN 47.80%

There is a lot of scope for hyperparameter tuning here. Experiments were done for 100 epochs.

Training Time on various hardware:

Model GTX 760 Intel Core i7
BOW+CNN 140 seconds/epoch 900 seconds/epoch
LSTM+CNN 200 seconds/epoch 1900 seconds/epoch

The above numbers are valid when using a batch size of 128, and training on 215K examples in every epoch.

##Get Started Have a look at the get_started.sh script in the scripts folder. Also, have a look at the readme present in each of the folders.

##Feedback All kind of feedback (code style, bugs, comments etc.) is welcome. Please open an issue on this repo instead of mailing me, since it helps me keep track of things better.

##License MIT