ML-CaPsule icon indicating copy to clipboard operation
ML-CaPsule copied to clipboard

MNIST data Classification

Open Saswatsusmoy opened this issue 9 months ago • 3 comments

MNIST Dataset

The MNIST dataset is a large collection of handwritten digits that's widely used for training and testing in machine learning, especially in the area of image processing. Here's a quick rundown:

  • Content: Black and white images of handwritten digits (0-9)
    
  • Size: 60,000 training images and 10,000 testing images
    
  • Properties: Images are all normalized to a fixed size (28x28 pixels) and centered
    
  • Origin: Created by Yann LeCun et al. by remixing samples from two National Institute of Standards and Technology (NIST) databases
    

MNIST is a popular choice for beginners in machine learning due to its simplicity and availability. It's a good starting point to learn image classification techniques.

Models

Add models with different architecture ranging from simple CNNs to complex ESNs

Models used (Tentative):

  • CNN along it's various architectures
  • Echo State Networks
  • Hybrid Models

Documentation

Proper documentation with rundowns on the Models, re-usable model classes, Comparisons on the models should be done

Saswatsusmoy avatar May 10 '24 21:05 Saswatsusmoy

Thanks for creating the issue

github-actions[bot] avatar May 10 '24 21:05 github-actions[bot]

Hello! I have worked on the project by MNIST data classification- handwritten digits("https://github.com/shrehs/Handwritten_Digits_Classifier-with-PyTorch.git"). I thnk I will be able to understand the data Loaders, CNN architecture, Optimization and validation and more. I believe I will be able to contribute to this issue. Please consider my request to be assigned to this task. Thank you very much!

shrehs avatar May 11 '24 13:05 shrehs

Hi, I'm a GSSOC'24 contributor and would like to contribute to this issue. Could you please assign me to this task?

anushkalahare avatar May 13 '24 05:05 anushkalahare

Hi, I am a GSSOC'24 contributor and would like to contribute to the issue. Please assign me this issue.

SayantikaLaskar avatar May 14 '24 12:05 SayantikaLaskar

Hey , @Saswatsusmoy I've already worked on this dataset and created a cnn-architecture from scartch using C++ which gives accuracy of 97% or more..I could uitlize similar approch to work on different models for this dataset particularly or make my model work on it.

Link to repo :- https://github.com/kyra-09/CNN_for_MNIST_from_scratch

Please assign this issue to me

Full Name :- Aaradhya Singh , Contributor/GSSOC(Girl Script Summer of Code) - 2024

aaradhyasinghgaur avatar May 14 '24 17:05 aaradhyasinghgaur