Search code examples
tensorflowtflearn

What is the relationship between tflearn and tf.contrib.learn?

What is the relationship between tflearn and tf.contrib.learn?

Solution

  • Both, tflearn and tf.contrib.learn are abstractions from pure Tensorflow.

    When you look at the Github projects (tflearn and tf.contrib.learn) you can compare them a bit:

                  tflearn        tf.contrib.learn
-------------------------------------------------
origins                      skflow (Scikit Flow)
first commit  31.05.2016     19.04.2016 (?)
commits       549            853 (*1)
contributors  91             39 (*2)

    While tf.contrib.learn describes itself as a "High level API for learning" and contains linear and logistic regression, kmeans, ... it does not offer a convolutional layer, for example. tf.contrib.learn aims to have a similar interface to scikit-learn.

    In contrast, tflearn describes itself as a "a modular and transparent deep learning library built on top of Tensorflow". It contains all basic deep learning building blocks.

    Methods

    • *1: git log --pretty=format:"%h" . | wc -l
    • *2: git log --pretty=format:"%an" . | sort | uniq | wc -l

    More, similar packages