https://github.com/geektrovert/deepnumta

A Comprehensive Study on Deep Learning Approaches for Bengali Handwritten Digit Recognition
https://github.com/geektrovert/deepnumta

bangla bangla-digit-recognition deep-learning digit-recognition handwritten-digit-recognition keras python3 tensorflow

Last synced: about 2 months ago
JSON representation

A Comprehensive Study on Deep Learning Approaches for Bengali Handwritten Digit Recognition

• Host: GitHub
• URL: https://github.com/geektrovert/deepnumta
• Owner: Geektrovert
• License: mit
• Created: 2018-07-30T17:44:21.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2019-08-24T19:25:26.000Z (almost 7 years ago)
• Last Synced: 2025-03-20T00:57:22.070Z (over 1 year ago)
• Topics: bangla, bangla-digit-recognition, deep-learning, digit-recognition, handwritten-digit-recognition, keras, python3, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 23.9 MB
• Stars: 2
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# DeepNumta

A Comprehensive Study on Deep Learning Approaches for Bengali Handwritten Digit Recognition

## Approaches

* [x] A custom built 10 layer Neural Network

  * Code: [DeepNumtaCNN.ipynb](./DeepNumtaCNN.ipynb)

  * Leaderboard position: [4th](https://www.kaggle.com/samnan/competitions)

  * Gets 99.4% accuracy on validation set and 96.68% test accuracy on Kaggle test set of NumtaDB

* [ ] ResNet34

* [ ] ResNeXt

## Dataset

NumtaDB - [https://www.kaggle.com/BengaliAI/numta](https://www.kaggle.com/BengaliAI/numta)