basic_vqa

Pytorch implementation of the paper - VQA: Visual Question Answering (https://arxiv.org/pdf/1505.00468.pdf).

Usage

1. Clone the repositories.

$ git clone https://github.com/tbmoon/basic_vqa.git

2. Download and unzip the dataset from official url of VQA: https://visualqa.org/download.html.

$ cd basic_vqa/utils
$ chmod +x download_and_unzip_datasets.csh
$ ./download_and_unzip_datasets.csh

3. Preproccess input data for (images, questions and answers).

$ python resize_images.py --input_dir='../datasets/Images' --output_dir='../datasets/Resized_Images'  
$ python make_vacabs_for_questions_answers.py --input_dir='../datasets'
$ python build_vqa_inputs.py --input_dir='../datasets' --output_dir='../datasets'

4. Train model for VQA task.

$ cd ..
$ python train.py

Results

• Comparison Result

Model	Metric	Dataset	Accuracy	Source
Paper Model	Open-Ended	VQA v2	54.08	VQA Challenge
My Model	Multiple Choice	VQA v2	54.72

• Loss and Accuracy on VQA datasets v2

References

• Paper implementation

  • Paper: VQA: Visual Question Answering
  • URL: https://arxiv.org/pdf/1505.00468.pdf

• Pytorch tutorial

  • URL: https://pytorch.org/tutorials/
  • Github: https://github.com/yunjey/pytorch-tutorial
  • Github: https://github.com/GunhoChoi/PyTorch-FastCampus

• Preprocessing

  • Tensorflow implementation of N2NNM
  • Github: https://github.com/ronghanghu/n2nmn