Internal change

PiperOrigin-RevId: 373175082

Author: achoum

README.md

@@ -50,6 +50,7 @@ The following resources are available:
 -   [Colabs](documentation/tutorials)
 -   [Migration guide from Neural Network to Decision Forests](documentation/migration.md)
 -   [Issue tracker](https://github.com/tensorflow/decision-forests/issues)
+-   [Known issues](documentation/known_issues.md)
 -   [Changelog](CHANGELOG.md)
 -   [Discuss on TensorFlow.Org](https://discuss.tensorflow.org)
 -   [Yggdrasil documentation](https://github.com/google/yggdrasil-decision-forests)

documentation/_index.yaml

@@ -49,6 +49,7 @@ landing_page:
           <li><a href="tutorials/index">Tutorial</a></li>
           <li><a href="migration">Migration guide from Neural Network to Decision Forests</a></li>
           <li><a href="https://github.com/tensorflow/decision-forests/issues">Issue tracker</a></li>
+          <li><a href="https://github.com/tensorflow/decision-forests/documentation/known_issues.md">Known issues</a></li>
           <li><a href="https://github.com/tensorflow/decision-forests/blob/master/CHANGELOG">Changelog</a></li>
           <li><a href="https://discuss.tensorflow.org">Discuss on TensorFlow.Org</a></li>
           <li><a href="https://github.com/google/yggdrasil-decision-forests">Yggdrasil Decision Forest</a> (for advanced users)</li>

documentation/known_issues.md

@@ -0,0 +1,35 @@
+# Known Issues
+
+The underlying engine behind the decision forests algorithms used by TensorFlow
+Decision Forests have been extensively production-tested. But this API to
+TensorFlow and Keras is new, and some issues are expected -- we are trying to
+fix them as quickly as possible.
+
+See also the
+[known issues of Yggdrasil Decision Forests](https://github.com/google/yggdrasil-decision-forests/documentation/known_issues.md).
+
+## Windows Pip package is not available
+
+TensorFlow Decision Forest is not yet available as a Windows Pip package.
+
+**Workarounds:**
+
+-   *Solution #1:* Install
+    [Windows Subsystem for Linux (WSL)](https://en.wikipedia.org/wiki/Windows_Subsystem_for_Linux)
+    on your Windows machine and follow the Linux instructions.
+
+## No support for in-model preprocessing with input of rank 1
+
+Keras expands the dimension of input tensors to rank 2 (using `tf.expand_dims`).
+If your `preprocessing` model argument only support rank 1 tensors, you will get
+an error complaining about tensor shape.
+
+**Workarounds:**
+
+-   *Solution #1:* Apply your preprocessing before the model, for example using
+    the dataset's
+    [`map`](https://www.tensorflow.org/api_docs/python/tf/data/Dataset#map)
+    method.
+
+-   *Solution #2:* Wrapps your preprocessing function into another function that
+    [squeeze](https://www.tensorflow.org/api_docs/python/tf/squeeze) its inputs.

documentation/tutorials/advanced_colab.ipynb

@@ -55,7 +55,12 @@
         "import pandas as pd\n",
         "import tensorflow as tf\n",
         "import math\n",
-        "import collections"
+        "import collections\n",
+        "\n",
+        "try:\n",
+        "  from wurlitzer import sys_pipes\n",
+        "except:\n",
+        "  from colabtools.googlelog import CaptureLog as sys_pipes"
       ]
     },
     {