Merge commit for internal changes

docs/_book.yaml

@@ -26,10 +26,6 @@ upper_tabs:
         path: /lattice/tutorials/keras_layers
       - title: Keras Premade Models
         path: /lattice/tutorials/premade_models
-      - title: Canned Estimators
-        path: /lattice/tutorials/canned_estimators
-      - title: Custom Estimators
-        path: /lattice/tutorials/custom_estimators
       - title: Aggregate Function Models
         path: /lattice/tutorials/aggregate_function_models
 

docs/_index.yaml

@@ -15,8 +15,7 @@ landing_page:
         <a href="./tutorials/shape_constraints">shape constraints</a>. This is done using a
         collection of <a href="./tutorials/keras_layers">Keras layers</a> that can satisfy
         constraints such as monotonicity, convexity and how features interact. The library also
-        provides easy to setup <a href="./tutorials/premade_models">premade models</a> and
-        <a href="./tutorials/canned_estimators">canned estimators</a>.</p>
+        provides easy to setup <a href="./tutorials/premade_models">premade models</a>.</p>
         <p>With TF Lattice you can use domain knowledge to better extrapolate to the parts of the
         input space not covered by the training dataset. This helps avoid unexpected model behaviour
         when the serving distribution is different from the training distribution.</p>

docs/install.md

@@ -5,7 +5,7 @@ There are several ways to set up your environment to use TensorFlow Lattice
 
 *   The easiest way to learn and use TFL requires no installation: run the any
     of the tutorials (e.g.
-    [canned estimators tutorial](tutorials/canned_estimators.ipynb)).
+    [premade models](tutorials/premade_models.ipynb)).
 *   To use TFL on a local machine, install the `tensorflow-lattice` pip package.
 *   If you have a unique machine configuration, you can build the package from
     source.
@@ -18,6 +18,8 @@ Install using pip.
 pip install --upgrade tensorflow-lattice
 ```
 
+Note that you will need to have `tf_keras` package installed as well.
+
 ## Build from source
 
 Clone the github repo:

docs/overview.md

@@ -6,7 +6,7 @@ knowledge into the learning process through common-sense or policy-driven
 [shape constraints](tutorials/shape_constraints.ipynb). This is done using a
 collection of [Keras layers](tutorials/keras_layers.ipynb) that can satisfy
 constraints such as monotonicity, convexity and pairwise trust. The library also
-provides easy to setup [canned estimators](tutorials/canned_estimators.ipynb).
+provides easy to setup [premade models](tutorials/premade_models.ipynb).
 
 ## Concepts
 
@@ -59,7 +59,7 @@ following show examples such calibrations functions with 10 keypoints:
 </p>
 
 It is often a good idea to use the quantiles of the features as input keypoints.
-TensorFlow Lattice [canned estimators](tutorials/canned_estimators.ipynb) can
+TensorFlow Lattice [premade models](tutorials/premade_models.ipynb) can
 automatically set the input keypoints to the feature quantiles.
 
 For categorical features, TensorFlow Lattice provides categorical calibration
@@ -209,8 +209,7 @@ include embeddings or other Keras layers.
 ## Tutorials and API docs
 
 For common model architectures, you can use
-[Keras premade models](tutorials/premade_models.ipynb) or
-[canned Estimators](tutorials/canned_estimators.ipynb). You can also create
+[Keras premade models](tutorials/premade_models.ipynb). You can also create
 custom models using [TF Lattice Keras layers](tutorials/keras_layers.ipynb) or
 mix and match with other Keras layers. Check out the
 [full API docs](https://www.tensorflow.org/lattice/api_docs/python/tfl) for