This project aims to share some image classification experimentation.
I recommend to install Tensorflow by using Anaconda and create a specific Anaconda environnement for Tensorflow as it's describe here :
https://www.tensorflow.org/install/install_windows#id=installing_with_anaconda (example for windows)
This repo may also use TFLearn which is a Python library on top of Tensorflow and provide High-level API usage. It also provides a lot of examples.
If you have a Nvidia graphic card, you can use it to speed up some image processing while running deep learning algorithms.
Tensorflow works well with the Nvidia CUDA Toolkit, you will need to install the following tools in this order :
- Microsoft Visual Studio : If you're on Windows, it's needed for better integration with CUDA
- CUDA Toolkit 9.0 : be carefull to chosse the CUDA 9.0 version as it's for now the latest version Tensorflow fully supports
- NVIDIA cuDNN : The GPU Accelerated Deep Learning library by Nvidia