An AI to identify which traffic sign appears in a photograph using neural networks.

At first, I wanted to achieve the highest accuracy with the dataset, so I used many layers. But it took too long to train, and the results weren't great. So, I tried removing layers one by one and adjusting settings. Surprisingly, this improved accuracy each time. So I noticed that more layers does not mean better accuracy and drop out affects the accuracy very much.

Download requirements: pip3 install -r requirements.txt

Start application: python traffic.py gtsrb

• Download the data set:
  • Data set
  • Small data set