Traffic Sign Classification
This is a convolutional neural network trained with TensorFlow to classify Traffic Signs.
- Create the following folders: Save, Models and Log.
- Download both training and testing GTSRB datasets, add Final_Test and Final_Training to the GTSRB folder.
- Go to Convert to JPG notebook.
- Create TFRecords using build_image_data.py. Feel free to change flags like train_directory, validation_directory, output_directory and labels_file.
- Go to Train ConvNet.
- (Optional) Go to Freeze Model notebook to create a .pb file that can be used on iOS and Android apps.