Prepare release of TF-DF 1.9.0rc0 and YDF 1.9.0rc0

PiperOrigin-RevId: 610665881

CHANGELOG.md

@@ -1,11 +1,11 @@
 # Changelog
 
-## HEAD
+## 1.9.0rc0 - 2024-02-26
 
 ### Feature
 
--   Add "parallel_trials" parameter in the hyper-parameter tuner to control the
-    number of trials to run in parallel.
+-   Add "parallel_trials" parameter in the hyper-parameter tuner to control the number of trials to run in parallel.
+-   Add support for custom losses.
 
 ## 1.8.0 - 2023-11-17
 

documentation/public/docs/hyperparameters.md

@@ -24,7 +24,6 @@ learner: "RANDOM_FOREST"
   num_trees: 1000
 }
 ```
-
 ## GRADIENT_BOOSTED_TREES
 
 A [Gradient Boosted Trees](https://statweb.stanford.edu/~jhf/ftp/trebst.pdf)
@@ -295,7 +294,8 @@ reasonable time.
 -   **Type:** Integer **Default:** 6 **Possible values:** min:-1
 
 -   Maximum depth of the tree. `max_depth=1` means that all trees will be roots.
-    Negative values are ignored.
+    `max_depth=-1` means that tree depth is not restricted by this parameter.
+    Values <= -2 will be ignored.
 
 #### [max_num_nodes](https://github.com/google/yggdrasil-decision-forests/blob/main/yggdrasil_decision_forests/learner/decision_tree/decision_tree.proto)
 
@@ -698,7 +698,8 @@ reasonable time.
 -   **Type:** Integer **Default:** 16 **Possible values:** min:-1
 
 -   Maximum depth of the tree. `max_depth=1` means that all trees will be roots.
-    Negative values are ignored.
+    `max_depth=-1` means that tree depth is not restricted by this parameter.
+    Values <= -2 will be ignored.
 
 #### [max_num_nodes](https://github.com/google/yggdrasil-decision-forests/blob/main/yggdrasil_decision_forests/learner/decision_tree/decision_tree.proto)
 
@@ -983,7 +984,8 @@ The hyper-parameter protobuffers are used with the C++ and CLI APIs.
 -   **Type:** Integer **Default:** 16 **Possible values:** min:-1
 
 -   Maximum depth of the tree. `max_depth=1` means that all trees will be roots.
-    Negative values are ignored.
+    `max_depth=-1` means that tree depth is not restricted by this parameter.
+    Values <= -2 will be ignored.
 
 #### [max_num_nodes](https://github.com/google/yggdrasil-decision-forests/blob/main/yggdrasil_decision_forests/learner/decision_tree/decision_tree.proto)
 
@@ -1165,7 +1167,8 @@ The hyper-parameter protobuffers are used with the C++ and CLI APIs.
 -   **Type:** Integer **Default:** 6 **Possible values:** min:-1
 
 -   Maximum depth of the tree. `max_depth=1` means that all trees will be roots.
-    Negative values are ignored.
+    `max_depth=-1` means that tree depth is not restricted by this parameter.
+    Values <= -2 will be ignored.
 
 #### [max_unique_values_for_discretized_numerical](https://github.com/google/yggdrasil-decision-forests/blob/main/yggdrasil_decision_forests/learner/distributed_gradient_boosted_trees/distributed_gradient_boosted_trees.proto)
 

yggdrasil_decision_forests/learner/distributed_gradient_boosted_trees/distributed_gradient_boosted_trees.proto

@@ -21,7 +21,6 @@ import "yggdrasil_decision_forests/learner/abstract_learner.proto";
 import "yggdrasil_decision_forests/learner/distributed_decision_tree/dataset_cache/dataset_cache.proto";
 import "yggdrasil_decision_forests/learner/distributed_decision_tree/load_balancer/load_balancer.proto";
 import "yggdrasil_decision_forests/learner/gradient_boosted_trees/gradient_boosted_trees.proto";
-import "yggdrasil_decision_forests/model/gradient_boosted_trees/gradient_boosted_trees.proto";
 
 // Training configuration for the Distributed Gradient Boosted Trees algorithm.
 message DistributedGradientBoostedTreesTrainingConfig {

yggdrasil_decision_forests/learner/distributed_gradient_boosted_trees/worker.proto

@@ -20,7 +20,6 @@ package yggdrasil_decision_forests.model.distributed_gradient_boosted_trees.prot
 import "yggdrasil_decision_forests/dataset/data_spec.proto";
 import "yggdrasil_decision_forests/learner/abstract_learner.proto";
 import "yggdrasil_decision_forests/learner/decision_tree/decision_tree.proto";
-import "yggdrasil_decision_forests/learner/distributed_decision_tree/dataset_cache/dataset_cache.proto";
 import "yggdrasil_decision_forests/learner/distributed_decision_tree/load_balancer/load_balancer.proto";
 import "yggdrasil_decision_forests/learner/distributed_decision_tree/training.proto";
 import "yggdrasil_decision_forests/model/decision_tree/decision_tree.proto";

yggdrasil_decision_forests/model/abstract_model.proto

@@ -17,7 +17,6 @@ syntax = "proto2";
 
 package yggdrasil_decision_forests.model.proto;
 
-import "yggdrasil_decision_forests/dataset/data_spec.proto";
 import "yggdrasil_decision_forests/dataset/weight.proto";
 import "yggdrasil_decision_forests/model/hyperparameter.proto";