CNN Model performance with 3,26,410 trainable parameters

CNN Model performance with 2,41,930 trainable parameters

Project information

  • Category: ML Classification
  • Project date: 11 Nov, 2023
  • Project URL: Project Link

Project Description

  • Main Aim: The primary goal of the project was to develop a classification model specifically designed for digit recognition. This involved creating a model capable of accurately identifying numerical digits from various inputs
  • Use of CNN Models: The project extensively employed multiple Convolutional Neural Networks (CNNs). These models are well-suited for image and pattern recognition tasks, like digit recognition.
  • Ensemble Methods: In an innovative approach, the project utilized ensemble methods. This involved taking the mean of all the model predictions to improve overall accuracy
  • Kaggle Competition Ranking: The project achieved significant recognition by securing the 419th rank out of 1800 teams in a Kaggle competition.