From 57918d67115684250986ffa654e9a09321e9f383 Mon Sep 17 00:00:00 2001
From: Adrian Garcia Badaracco <1755071+adriangb@users.noreply.github.com>
Date: Mon, 29 Nov 2021 14:06:47 -0800
Subject: [PATCH] doc: add note on KerasRegressor.score() return value to
 migration guide (#255)

---
 docs/source/migration.rst | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/docs/source/migration.rst b/docs/source/migration.rst
index 4e369578..e039b398 100644
--- a/docs/source/migration.rst
+++ b/docs/source/migration.rst
@@ -60,6 +60,13 @@ Calling ``fit`` resets this attribute, calling ``partial_fit`` on the other hand
 .. note::
    Unlike the TensorFlow wrappers, SciKeras normalizes the names of the keys, so that if you use `metrics=["mae"]` you will get a key named `"mean_absolute_error"`.
 
+``score()`` return R^2 by default for regressors
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Keras returns your model's mean loss when ``KerasRegressor.score()`` is called.
+SciKeras, like most Scikit-Learn estimators, returns the coefficient of determination of the prediction (also known as R^2 score).
+This is however customizable.
+
 One-hot encoding of targets for categorical crossentropy losses
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^