Add documentation for warm-start (#255)

* add documentation for warm-start * fix typo * fix typo * Update flaml/tune/tune.py Co-authored-by: Chi Wang <[email protected]> * Update automl.py Co-authored-by: Qingyun Wu <[email protected]> Co-authored-by: Chi Wang <[email protected]>
microsoft · Oct 19, 2021 · 94a81a9 · 94a81a9
1 parent 7809ec1
commit 94a81a9
Show file tree

Hide file tree

Showing 2 changed files with 26 additions and 0 deletions.
diff --git a/flaml/automl.py b/flaml/automl.py
@@ -1402,6 +1402,20 @@ def custom_metric(
                 hyperparamter configurations for the corresponding estimators.
                 The value can be a single hyperparamter configuration dict or a list
                 of hyperparamter configuration dicts.
+                In the following code example, we get starting_points from the
+                automl_experiment and use them in the new_automl_experiment.
+                e.g.,
+                .. code-block:: python
+                    from flaml import AutoML
+                    automl_experiment = AutoML()
+                    X_train, y_train = load_iris(return_X_y=True)
+                    automl_experiment.fit(X_train, y_train)
+                    starting_points = automl_experiment.best_config_per_estimator
+
+                    new_automl_experiment = AutoML()
+                    new_automl_experiment.fit(X_train, y_train,
+                        starting_points=starting_points)
+
             seed: int or None, default=None | The random seed for np.random.
             n_concurrent_trials: [Experimental] int, default=1 | The number of
                 concurrent trials. For n_concurrent_trials > 1, installation of

diff --git a/flaml/tune/tune.py b/flaml/tune/tune.py
@@ -194,6 +194,18 @@ def compute_with_config(config):
             as a list so the optimiser can be told the results without
             needing to re-compute the trial. Must be the same length as
             points_to_evaluate.
+            e.g.,
+            .. code-block:: python
+                points_to_evaluate = [
+                    {"b": .99, "cost_related": {"a": 3}},
+                    {"b": .99, "cost_related": {"a": 2}},
+                ]
+                evaluated_rewards=[3.0, 1.0]
+
+            means that you know the reward for the two configs in
+            points_to_evaluate are 3.0 and 1.0 respectively and want to
+            inform run()
+
         prune_attr: A string of the attribute used for pruning.
             Not necessarily in space.
             When prune_attr is in space, it is a hyperparameter, e.g.,