/image-captioning

Arithmetic operation to be performed on the MNIST data

Primary LanguageJupyter Notebook

slide from Rachal's TEDx.

All about mnist

Contents

Project Overview

Data

MNIST

  • Using MNIST dataset, which is bacis to machine learning, you can learn the core concepts of NLP and Computer vision.
  • Generated arithmetic equations which has random length, code
    • SerializationDataset: Arrange the data
    • CalculationDataset: Automatically make operations
    • ClassificationDataset: Making data lable, which is result of arithmetic
  • if you want to use jupyter notebook and test it, click code

Algorithms

Architecture

sequence to sequence model (Enncoder-Decoder)

  • Seq2seq model, pytorch code

  • Seq2seq with attention, pytorch code
  • VGG classification model, tensorflow code

Reference

Language modeling paper/article which is applied in these project