A few fixes in docs

Author: levkk

pgml-docs/docs/user_guides/training/hyperparameter_search.md

@@ -43,7 +43,7 @@ This grid search will train `len(max_depth) * len(n_estimators) * len(learning_r
 
 === "SQL"
 
-    ```sql
+    ```sql linenums="1"
     SELECT * FROM pgml.train(
         'Handwritten Digit Image Classifier', 
         algorithm => 'xgboost', 
@@ -58,7 +58,7 @@ This grid search will train `len(max_depth) * len(n_estimators) * len(learning_r
 
 === "Output"
 
-    ```
+    ```sql linenums="1"
                 project_name            | objective | algorithm_name |  status
     ------------------------------------+-----------+----------------+----------
      Handwritten Digit Image Classifier |           | xgboost        | deployed

pgml-docs/docs/user_guides/training/joint_optimization.md

@@ -1,5 +1,5 @@
 # Joint Optimization
 
-Some algorithms support joint optimization of the objective across multiple outputs, and can improve results compared to using multiple independent models. To leverage multiple outputs in PostgresML, you'll need to substitue the standard usage of `pgml.train` and `pgml.predict` with `pgml.train_multi` and `pgml.predict_multi`. The `_multi` functions are identical, except `train_multi` takes an array of `y_column_names TEXT[]`, and `predict_multi` returns an array of outputs correspondingly.
+Some algorithms support joint optimization of the objective across multiple outputs, and can improve results compared to using multiple independent models. To leverage multiple outputs in PostgresML, you'll need to substitue the standard usage of `pgml.train` and `pgml.predict` with `pgml.train_joint` and `pgml.predict_joint`. The `_joint` functions are identical, except `train_joint` takes an array of `y_column_names TEXT[]`, and `predict_joint` returns an array of outputs correspondingly.
 
-Read more at [scikit-learn](https://scikit-learn.org/stable/modules/classes.html#module-sklearn.multioutput).
+Read more at [scikit-learn](https://scikit-learn.org/stable/modules/classes.html#module-sklearn.multioutput).

pgml-docs/docs/user_guides/training/overview.md

@@ -103,7 +103,7 @@ Now that we've got data, we're ready to train a model using an algorithm. We'll
 
 === "Output"
 
-    ```sql linenumes="1"
+    ```sql linenums="1"
                 project_name            |   objective    | algorithm_name |  status
     ------------------------------------+----------------+----------------+----------
      Handwritten Digit Image Classifier | classification | linear         | deployed