/lenet-5-traffic-light

Primary LanguagePython

lenet-5-minst

Train a LeNet-5 neural network with Keras for minst classification

Training

To train the network: python train.py

Monitoring with tensorboard: tensorboard --logdir=logs/

and open http://localhost:6006

Classification

To evaluate for the 3 images: python classify.py

Dataset

  • Udacity simulator traffic light images (173 images)
  • Udactiy parking lot loop traffic light images (48 images)
  • Udactiy parking lot traffic light images (118 images)
  • Bosch traffic light train images (10698 images)
  • Bosch traffic light test images (13486 images)
  • Bosch traffic light additional images (320 images)