Release TF-DF 1.6.0 and YDF 1.6.0

PiperOrigin-RevId: 569106996

CHANGELOG.md

@@ -1,14 +1,14 @@
 # Changelog
 
-## 1.6.0 - rc0 2023-08-22
+## 1.6.0 2023-09-27
 
 ### Breaking Changes
 
 -   TF-DF no longer supports Python 3.8 since Tensorflow dropped its support.
 
 ### Features
 
--   Compatibility with Tensorflow 2.14.0 rc0
+-   Compatibility with Tensorflow 2.14.0
 -   Contrib: Training preprocessing jointly on the input features, labels and 
     weights
 

README.md

@@ -120,6 +120,11 @@ the following paper:
 Yggdrasil Decision Forests: A Fast and Extensible Decision Forests Library,
 Guillame-Bert et al., KDD 2023: 4068-4077. doi:10.1145/3580305.3599933
 
+## Contact
+
+You can contact the core development team at
+[[email protected]](mailto:[email protected]).
+
 ## Credits
 
 TensorFlow Decision Forests was developed by:

WORKSPACE

@@ -11,12 +11,12 @@ load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
 # absl used by tensorflow.
 http_archive(
     name = "org_tensorflow",
-    strip_prefix = "tensorflow-2.13.0",
+    strip_prefix = "tensorflow-2.14.0",
     sha256 = "447cdb65c80c86d6c6cf1388684f157612392723eaea832e6392d219098b49de",
-    urls = ["https://github.com/tensorflow/tensorflow/archive/v2.13.0.zip"],
+    urls = ["https://github.com/tensorflow/tensorflow/archive/v2.14.0.zip"],
     # Starting with TF 2.14, disable hermetic Python builds.
-    # patch_args = ["-p1"],
-    # patches = ["//third_party/tensorflow:tf.patch"],
+    patch_args = ["-p1"],
+    patches = ["//third_party/tensorflow:tf.patch"],
 )
 
 # Inject tensorflow dependencies.

configure/setup.py

@@ -21,15 +21,15 @@
 from setuptools.command.install import install
 from setuptools.dist import Distribution
 
-_VERSION = "1.6.0rc0"
+_VERSION = "1.6.0"
 
 with open("README.md", "r", encoding="utf-8") as fh:
   long_description = fh.read()
 
 REQUIRED_PACKAGES = [
     "numpy",
     "pandas",
-    "tensorflow~=2.14.0rc0",
+    "tensorflow~=2.14.0",
     "six",
     "absl_py",
     "wheel",

documentation/known_issues.md

@@ -44,6 +44,7 @@ The following table shows the compatibility between
 
 tensorflow_decision_forests | tensorflow
 --------------------------- | ---------------
+1.6.0                       | 2.14.0
 1.5.0                       | 2.13.0
 1.3.0 - 1.4.0               | 2.12.0
 1.1.0 - 1.2.0               | 2.11.0

documentation/migration.md

@@ -18,45 +18,7 @@ This doc assumes familiarity with the
 
 <!--ts-->
 
-*   [Migrating from Neural Networks](#migrating-from-neural-networks)
-    *   [Table of Contents](#table-of-contents)
-    *   [Dataset and Features](#dataset-and-features)
-        *   [Validation dataset](#validation-dataset)
-        *   [Dataset I/O](#dataset-io)
-            *   [Train for exactly 1 epoch](#train-for-exactly-1-epoch)
-            *   [Do not shuffle the dataset](#do-not-shuffle-the-dataset)
-            *   [Do not tune the batch size](#do-not-tune-the-batch-size)
-        *   [Large Datasets](#large-datasets)
-            *   [How many examples to use](#how-many-examples-to-use)
-        *   [Feature Normalization / Preprocessing](#feature-normalization--preprocessing)
-            *   [Do not transform data with feature columns](#do-not-transform-data-with-feature-columns)
-            *   [Do not preprocess the features](#do-not-preprocess-the-features)
-            *   [Do not normalize numerical features](#do-not-normalize-numerical-features)
-            *   [Do not encode categorical features (e.g. hashing, one-hot, or
-                embedding)](#do-not-encode-categorical-features-eg-hashing-one-hot-or-embedding)
-            *   [How to handle text features](#how-to-handle-text-features)
-            *   [Do not replace missing features by magic values](#do-not-replace-missing-features-by-magic-values)
-            *   [Handling Images and Time series](#handling-images-and-time-series)
-    *   [Training Pipeline](#training-pipeline)
-        *   [Don't use hardware accelerators e.g. GPU, TPU](#dont-use-hardware-accelerators-eg-gpu-tpu)
-        *   [Don't use checkpointing or mid-training hooks](#dont-use-checkpointing-or-mid-training-hooks)
-        *   [Model Determinism](#model-determinism)
-        *   [Training Configuration](#training-configuration)
-            *   [Specify a task (e.g. classification, ranking) instead of a loss
-                (e.g. binary
-                cross-entropy)](#specify-a-task-eg-classification-ranking-instead-of-a-loss-eg-binary-cross-entropy)
-            *   [Hyper-parameters are semantically stable](#hyper-parameters-are-semantically-stable)
-        *   [Model debugging](#model-debugging)
-            *   [Simple model summary](#simple-model-summary)
-            *   [Training Logs and Tensorboard](#training-logs-and-tensorboard)
-            *   [Feature importance](#feature-importance)
-            *   [Plotting the trees](#plotting-the-trees)
-            *   [Access the tree structure](#access-the-tree-structure)
-            *   [Do not use TensorFlow distribution strategies](#do-not-use-tensorflow-distribution-strategies)
-            *   [Stacking Models](#stacking-models)
-            *   [Migrating from tf.estimator.BoostedTrees
-                {Classifier/Regressor/Estimator}](#migrating-from-tfestimatorboostedtrees-classifierregressorestimator)
-    *   [For Yggdrasil users](#for-yggdrasil-users)
+<!-- Created by https://github.com/ekalinin/github-markdown-toc -->
 
 <!--te-->
 

tensorflow_decision_forests/__init__.py

@@ -51,10 +51,10 @@
 
 """
 
-__version__ = "1.6.0rc0"
+__version__ = "1.6.0"
 __author__ = "Mathieu Guillame-Bert"
 
-compatible_tf_versions = ["2.14.0rc0"]
+compatible_tf_versions = ["2.14.0"]
 __git_version__ = "HEAD"  # Modify for release build.
 
 from tensorflow_decision_forests.tensorflow import check_version

tensorflow_decision_forests/keras/wrappers_pre_generated.py

@@ -43,9 +43,9 @@ class CartModel(core.CoreModel):
   r"""Cart learning algorithm.
 
   A CART (Classification and Regression Trees) a decision tree. The non-leaf
-  nodes contains conditions (also known as splits) while the leaf nodes
-  contains prediction values. The training dataset is divided in two parts. The
-  first is used to grow the tree while the second is used to prune the tree.
+  nodes contains conditions (also known as splits) while the leaf nodes contain
+  prediction values. The training dataset is divided in two parts. The first is
+  used to grow the tree while the second is used to prune the tree.
 
   Usage example:
 
@@ -331,7 +331,7 @@ class CartModel(core.CoreModel):
       `CHI_SQUARED` or `CS`: (p-q)^2/q
         Default: "KULLBACK_LEIBLER".
     validation_ratio: Ratio of the training dataset used to create the
-      validation dataset used to prune the tree. If set to 0, the entire dataset
+      validation dataset for pruning the tree. If set to 0, the entire dataset
       is used for training, and the tree is not pruned. Default: 0.1.
   """
 

tools/test_bazel.sh

@@ -91,12 +91,6 @@ sed -i $ext "s/sha256 = \"${prev_shasum}\",//" WORKSPACE
 TENSORFLOW_BAZELRC="tensorflow_bazelrc"
 curl https://raw.githubusercontent.com/tensorflow/tensorflow/${commit_slug}/.bazelrc -o ${TENSORFLOW_BAZELRC}
 
-# For Tensorflow versions > 2.13, apply a patch to disable hermetic builds.
-if [[ ${TF_MINOR} != "2.13" ]]; then
-  sed -i $ext "s/# patch_args = \[\"-p1\"\],/patch_args = \[\"-p1\"\],/" WORKSPACE
-  sed -i $ext "s/# patches = \[\"\/\/third_party\/tensorflow:tf.patch\"\],/patches = \[\"\/\/third_party\/tensorflow:tf.patch\"\],/" WORKSPACE
-fi
-
 # Bazel common flags. Startup flags are already given through STARTUP_FLAGS
 FLAGS=