CS663ML Frameworks
There are a number of machine learning frameworks that implement many of the recognition algorithms we will learn about in this class
In fact OpenCV implements a number of them for you.
However, the most popular frameworks out there can work with many kinds of data (not just images).
More recently, there is a strong use of Python in many of the frameworks but the most popular frameworks support multiple languages like Java, C++, Python and more.
In this class we will be learning about TensorFlow ---why? It is popular and a good framework See the following image
-
TensorFlow is WITH PyTorch considered the mainstream of deep learning frameworks
-
It is develped by Google
PyTorch and Tensorflow have similar capabilities. We are using Tensorflow for its good on (Android) device support.
FUNNY Video
image from from https://medium.com/@valkont/battle-of-the-giants-tensorflow-vs-pytorch-2023-fd8274210a38 (2023)
There are Frameworks on top of Tensorflow: Keras and TF Learn
....why??? well Tensorflow has more code than you might want to write and groups have developed frameworks that sit on top to make setting up and training and running of networks easier --- read this article for a discussion.
....some of the most popular are TFlearn and Keras..... some definitions of the 2
-
TFLearn = a modular and transparent deep learning library built on top of Tensorflow. It was designed to provide a higher-level API to TensorFlow in order to facilitate and speed-up experimentations, while remaining fully transparent and compatible with it
-
Keras = is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano. It was developed with a focus on enabling fast experimentation.
TF Keras --Keras now integrated into Tensorflow (2+)
As of mid-2017, Keras was actually fully adopted and integrated into TensorFlow ( tf.keras)