README.md

Project: Build a Traffic Sign Recognition Program

Steps to run code

1. Clone starter kit repo to setup environment https://github.com/udacity/CarND-Term1-Starter-Kit
2. Follow these directions: https://github.com/udacity/CarND-Term1-Starter-Kit/blob/master/doc/configure_via_anaconda.md and create the gpu environment
3. Activate the carnd-term1 environment

To run all the experiments you can take the following steps (Note: I would not recommend running the experiments because it is training a model for 4 epochs 430 different times with different hyperparameter combinations):

1. Run python fetch_data.py
2. Run python run_experiments.py
3. Wait 5 hours
4. Run python run_top_five.py

The Jupyter notebook and writeup both contain explanations for the experiments being ran. Additionally the Jupyter notebook contains a model that can be trained with the set of hyperparmeters that I found had the highest accuracy rate from the experiments.

Writeup

Traffic Sign Classifier Writeup