Visual Question Answering Demo on pretrained model
Jupyter Notebook Python
Switch branches/tags
Nothing to show
Clone or download
iamaaditya Merge pull request #18 from akshitac8/master
updated question and image features function
Latest commit 6295d23 Jun 25, 2018

VQA Demo

Updated to work with Keras 2.0 and TF 1.2 and Spacy 2.0 This code is meant for education thus focus is on simplicity and not speed.

This is a simple Demo of Visual Question answering which uses pretrained models (see models/CNN and models/VQA) to answer a given question about the given image.


  1. Keras version 2.0+

    • Modular deep learning library based on python
  2. Tensorflow 1.2+ (Might also work with Theano. I have not tested Theano after the recent commit, use commit 0f89007 for Theano)

  3. scikit-learn

    • Quintessential machine library for python
  4. Spacy version 2.0+

    • Used to load Glove vectors (word2vec)
    • To upgrade & install Glove Vectors
      • python -m spacy download en_vectors_web_lg
  5. OpenCV

    • OpenCV is used only to resize the image and change the color channels,
    • You may use other libraries as long as you can pass a 224x224 BGR Image (NOTE: BGR and not RGB)
  6. VGG 16 Pretrained Weights


python -image_file_name path_to_file -question "Question to be asked"


python -image_file_name test.jpg -question "Is there a man in the picture?"

if you have prefer to use Theano backend and if you have GPU you may want to run like this

THEANO_FLAGS='floatX=float32,device=gpu0,lib.cnmem=1,mode=FAST_RUN' python -image_file_name test.jpg -question "What vechile is in the picture?"

Expected Output : 095.2 % train 00.67 % subway 00.54 % mcdonald's 00.38 % bus 00.33 % train station


  • GPU (Titan X) Theano optimizer=fast_run : 51.3 seconds
  • GPU (Titan X) Theano optimizer=fast_compile : 47.5 seconds
  • CPU (i7-5820K CPU @ 3.30GHz : 35.9 seconds (Is this strange or not ?)

iPython Notebook

Jupyter/iPython Notebook has been provided with more examples and interactive tutorial.

NOTE: See the comments on for more information on the model and methods

VQA Training