V0.5.6 (#128)

* recover ConcurrencyLimiter

* cost attribute

* update notebooks

Co-authored-by: Chi Wang <wang.chi@microsoft.com>
Co-authored-by: Qingyun Wu <qiw@microsoft.com>

flaml/automl.py

@@ -981,9 +981,15 @@ class AutoML:
         self._retrained_config = {}
         est_retrain_time = next_trial_time = 0
         best_config_sig = None
+        # use ConcurrencyLimiter to limit the amount of concurrency when
+        # using a search algorithm
         better = True  # whether we find a better model in one trial
         if self._ensemble:
             self.best_model = {}
+        try:
+            from ray.tune.suggest import ConcurrencyLimiter
+        except ImportError:
+            from .searcher.suggestion import ConcurrencyLimiter
         if self._hpo_method in ('cfo', 'grid'):
             from flaml import CFO as SearchAlgo
         elif 'optuna' == self._hpo_method:
@@ -1056,11 +1062,13 @@ class AutoML:
                         metric='val_loss', mode='min', space=search_space,
                         points_to_evaluate=points_to_evaluate,
                     )
-                search_state.search_alg = algo
+                search_state.search_alg = ConcurrencyLimiter(algo,
+                                                             max_concurrent=1)
             else:
                 search_space = None
                 if self._hpo_method in ('bs', 'cfo'):
                     search_state.search_alg.set_search_properties(
+                        metric=None, mode=None,
                         config={
                             'metric_target': self._state.best_loss,
                         },

flaml/searcher/blendsearch.py

@@ -133,8 +133,8 @@ class BlendSearch(Searcher):
         else:
             self._gs = None
         self._ls = self.LocalSearch(
-            init_config, metric, mode, cat_hp_cost, space,
-            prune_attr, min_resource, max_resource, reduction_factor, seed)
+            init_config, metric, mode, cat_hp_cost, space, prune_attr,
+            min_resource, max_resource, reduction_factor, self.cost_attr, seed)
         self._init_search()
 
     def set_search_properties(self,

flaml/searcher/flow2.py

@@ -26,7 +26,6 @@ class FLOW2(Searcher):
 
     STEPSIZE = 0.1
     STEP_LOWER_BOUND = 0.0001
-    cost_attr = 'time_total_s'
 
     def __init__(self,
                  init_config: dict,
@@ -38,6 +37,7 @@ class FLOW2(Searcher):
                  min_resource: Optional[float] = None,
                  max_resource: Optional[float] = None,
                  resource_multiple_factor: Optional[float] = 4,
+                 cost_attr: Optional[str] = 'time_total_s',
                  seed: Optional[int] = 20):
         '''Constructor
 
@@ -73,6 +73,7 @@ class FLOW2(Searcher):
                 prune_attr; only valid if prune_attr is not in space.
             resource_multiple_factor: A float of the multiplicative factor
                 used for increasing resource.
+            cost_attr: A string of the attribute used for cost.
             seed: An integer of the random seed.
         '''
         if mode:
@@ -98,6 +99,7 @@ class FLOW2(Searcher):
         self.prune_attr = prune_attr
         self.min_resource = min_resource
         self.resource_multiple_factor = resource_multiple_factor or 4
+        self.cost_attr = cost_attr
         self.max_resource = max_resource
         self._resource = None
         self._step_lb = np.Inf
@@ -292,7 +294,7 @@ class FLOW2(Searcher):
             init_config, self.metric, self.mode, self._cat_hp_cost,
             unflatten_dict(self.space), self.prune_attr,
             self.min_resource, self.max_resource,
-            self.resource_multiple_factor, self._seed + 1)
+            self.resource_multiple_factor, self.cost_attr, self._seed + 1)
         flow2.best_obj = obj * self.metric_op  # minimize internally
         flow2.cost_incumbent = cost
         self._seed += 1
@@ -534,7 +536,7 @@ class FLOW2(Searcher):
             config[self.prune_attr] = self._resource
             self._direction_tried = None
             self._configs[trial_id] = (config, self.step)
-            return config
+            return unflatten_dict(config)
         self._num_allowed4incumbent -= 1
         move = self.incumbent.copy()
         if self._direction_tried is not None:
@@ -553,7 +555,7 @@ class FLOW2(Searcher):
         self._proposed_by[trial_id] = self.incumbent
         self._configs[trial_id] = (config, self.step)
         self._num_proposedby_incumbent += 1
-        best_config = flatten_dict(self.best_config)
+        best_config = self.best_config
         if self._init_phase:
             if self._direction_tried is None:
                 if self._same:

flaml/version.py

@@ -1 +1 @@
-__version__ = "0.5.5"
+__version__ = "0.5.6"