Fix documentation build

PiperOrigin-RevId: 702282558
google · Dec 3, 2024 · 4147904 · 4147904
1 parent cfd4275
commit 4147904
Show file tree

Hide file tree

Showing 5 changed files with 10 additions and 20 deletions.
diff --git a/documentation/public/docs/guide_feature_semantics.md b/documentation/public/docs/guide_feature_semantics.md
@@ -11,7 +11,7 @@ example, a feature represented by 64-bit integers might have a numerical
 semantic or a categorical semantic.
 
 In basic cases, YDF detects the semantics of a feature automatically, so you
-only need to check them after the training (e.g. use [model.describe()](/py_api/GenericModel/#ydf.GenericModel.describe)). If YDF
+only need to check them after the training (e.g. use [model.describe()](py_api/GenericModel/#ydf.GenericModel.describe)). If YDF
 does not detect the correct semantic, you can manually override it. Using the
 wrong semantic negatively impacts the training speed and quality of a model.
 Also, YDF is not able to consume all the types of features. In such cases,
@@ -22,7 +22,7 @@ select them, and gives recommendations on how to feed different types of
 features into the model.
 
 This guide assumes basic familiarity with YDF, e.g. the
-[Getting Started](/tutorial/getting_started)
+[Getting Started](tutorial/getting_started)
 tutorial.
 
 ## Introduction: How to specify feature semantics
@@ -248,7 +248,7 @@ YDF “unrolls” each entry of the vector to an individual feature. These
 features are named `text.0_of_512`, `text.1_of_512`, etc. Note that all
 vectors must have the exact same size - if the vectors have different sizes,
 consider the CATEGORICAL_SET semantic. See
-[here](/tutorial/multidimensional_feature)
+[here](tutorial/multidimensional_feature)
 for a more detailed example.
 
 ```python
@@ -297,7 +297,7 @@ numerical unix time generally does not work well.
 
 Time series datasets require advanced feature preprocessing for good model
 quality. Check the
-[special guide](/tutorial/time_sequences)
+[special guide](tutorial/time_sequences)
 in the YDF documentation for more information.
 
 !!! warning
@@ -464,7 +464,7 @@ are considered, and how the algorithm finds the best split.
     thresholds (e.g., "age >= 30").
 
     For even more powerful models,
-    [enable oblique splits](/guide_how_to_improve_model/#use-oblique-trees).
+    [enable oblique splits](guide_how_to_improve_model/#use-oblique-trees).
     This allows YDF to learn splits that combine multiple numerical features
     (e.g., "0.3 \* age + 0.7 \* income >= 50"). This is particularly helpful for
     smaller datasets but requires more training time.

diff --git a/documentation/public/docs/index.md b/documentation/public/docs/index.md
@@ -172,7 +172,7 @@ model.save("/tmp/my_model")
 
 -   Compose YDF models with Neural Network models in
     [TensorFlow, Keras](tutorial/compose_with_tf.ipynb), and
-    [JAX](compose_with_jax.ipynb).
+    [JAX](tutorial/compose_with_jax.ipynb).
 -   [Distributed training](tutorial/distributed_training.ipynb) over billions of
     examples and hundreds of machines.
 -   Use cutting-edge learning algorithm such as oblique splits, honest trees,

diff --git a/documentation/public/docs/py_api/evaluate.md b/documentation/public/docs/py_api/evaluate.md
diff --git a/documentation/public/docs/py_api/index.md b/documentation/public/docs/py_api/index.md
@@ -50,16 +50,8 @@ A **Feature Selector** finds the optimal set of input features for the model.
 -   [BackwardSelectionFeatureSelector](BackwardSelectionFeatureSelector.md)
 -   [FeatureSelectorLogs](FeatureSelectorLogs.md)
 
-## Prediction Evaluation
-
-**Prediction Evaluation** allows evaluating arbitrary predictions with YDF's
-tools. 
-
--   [Prediction evaluation](evaluate.md)
-
 ## Other
 
--   [verbose](utilities.md#ydf.verbose): Control the amount of logging.
 -   [load_model](utilities.md#ydf.load_model): Load a model from disk.
 -   [Feature](utilities.md#ydf.Feature): Input feature specific hyper-parameters
     e.g. semantic, constraints.
@@ -68,6 +60,8 @@ tools.
     classification.
 -   [Semantic](utilities.md#ydf.Semantic): How an input feature in interpreted
     e.g. numerical, categorical.
+-   [evaluate_predictions](utilities.md#ydf.evaluate_predictions): Evaluates predictions of YDF and non-YDF models.
+-   [verbose](utilities.md#ydf.verbose): Control the amount of logging.
 -   [start_worker](utilities.md#ydf.start_worker): Start a worker for
     distributed training.
 -   [strict](utilities.md#ydf.strict): Show more logs.

diff --git a/documentation/public/docs/py_api/utilities.md b/documentation/public/docs/py_api/utilities.md
@@ -16,6 +16,8 @@
 
 ::: ydf.Semantic
 
+::: ydf.evaluate_predictions
+
 ::: ydf.start_worker
 
 ::: ydf.strict