DOC: add comparison to TF wrappers (#133)

adriangb · Nov 27, 2020 · 94960da · 94960da
1 parent 8d47423
commit 94960da
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 2 deletions.
diff --git a/README.md b/README.md
@@ -6,6 +6,22 @@
 
 Scikit-Learn compatible wrappers for Keras Models.
 
+## Why SciKeras
+
+SciKeras is derived from and API compatible with `tf.keras.wrappers.scikit_learn`. The original TensorFlow (TF) wrappers are not actively mantained,
+and [may be deprecated](https://github.com/tensorflow/tensorflow/pull/37201#pullrequestreview-391650001) at some point.
+In addition, they have many incompatibilities with both the Keras ecosystem and the Scikit-Learn ecosystem.
+SciKeras attempts to resolve these issues by providing maintained, documented wrappers that are fully compatible with the
+entire Scikit-Learn and Keras ecosystems. Some advantages over the TF wrappers are:
+
+* Full compatibility with the Scikit-Learn API, including grid searches, ensembles, transformers, etc.
+* Support for pre-trained models.
+* Support for dynamically set Keras parameters depending on inputs (e.g. input shape).
+* Support for hyperparameter tuning of optimizers and losses.
+* Support for multi-input and multi-ouput Keras models.
+* Functional `random_state` for reproducible training.
+* Many more that you will discover as you use SciKeras!
+
 ## Installation
 
 This package is available on PyPi:
@@ -14,7 +30,7 @@ This package is available on PyPi:
 pip install scikeras
 ```
 
-The only dependencies are `scikit-learn>=0.22` and `TensorFlow>=2.1.0`.
+The only dependencies are `scikit-learn>=0.22` and `TensorFlow>=2.2.0`.
 
 ## Documentation
 

diff --git a/docs/source/index.rst b/docs/source/index.rst
@@ -3,7 +3,17 @@ Welcome to SciKeras's documentation!
 
 The goal of scikeras is to make it possible to use Keras/TensorFlow with sklearn. 
 This is achieved by providing a wrapper around Keras that has an Scikit-Learn interface.
-SciKeras is the successor to ``~tensorflow.keras.wrappers.scikit_learn``.
+SciKeras is the successor to ``tf.keras.wrappers.scikit_learn``.
+
+Some advantages over the ``tf.keras.wrappers.scikit_learn`` wrappers are:
+
+* Full compatibility with the Scikit-Learn API, including grid searches, ensembles, transformers, etc.
+* Support for pre-trained models.
+* Support for dynamically set Keras parameters depending on inputs (e.g. input shape).
+* Support for hyperparameter tuning of optimizers and losses.
+* Support for multi-input and multi-ouput Keras models.
+* Functional `random_state` for reproducible training.
+* Many more that you will discover as you use SciKeras!
 
 SciKeras tries to make things easy for you while staying out of your way.
 If you are familiar with Scikit-Learn and Keras, you don’t have to learn any new concepts, and the syntax should be well known.