Skip to content

Latest commit

 

History

History
27 lines (19 loc) · 995 Bytes

README.md

File metadata and controls

27 lines (19 loc) · 995 Bytes
Raw

Car_logo-Classifier

This is a neural network image classifier that helps to identify 10 different classes of car brands.

Neural Network - 98% Validation_acc and 0.37 Validation_error

The neural network used is a Convoluted Neural network (CNN) with an architecture similar to the AlexNet (made by Prof. Alex Krizhevsky of University of Toronto). The neural network was trained on 10 categories containing 1000 images each. The test set consisted of 150 images per category. After running for 15 epoch the the model achieved a validation accuracy of 98% and validation loss of 0.37!

The architecture

Neural Network Architecture

The categories

  1. Buick
  2. Chery
  3. Citroen
  4. Honda
  5. Hyundai
  6. Lexus
  7. Mazda
  8. Peugeot
  9. Toyota
  10. Volkswagen

The data

The data was contributed by BinhMinhs10 on kaggle and you can find it here: https://www.kaggle.com/binhminhs10/10-class-for-car-logo?