Skip to content

ndtands/CNN_Color_Car_Pytorch

Repository files navigation

Build model CNN for Detection Color Car

1. Architecture

Link paper

CNN architecture

3. Build model

  • Implementation paper with pytorch
  • Custorm with metrics F1-score, Early Stopping, Scheduler learning rate.
  • You can try with mobinet or resnet18. But it using pretrain not neccessary for simple classification.

2. Evaluation

  • Confusion Matrix of Train dataset classification_report confusion matrix
  • confusion matrix of Test dataset classification_report confusion matrix

3. Prediction

KQ

  • Inference Time: 0.028s