[rllib/tune] Cache get_preprocessor() calls, default max_failur… (#6211)

python/ray/tune/tune.py

@@ -74,7 +74,7 @@ def run(run_or_experiment,
         checkpoint_score_attr=None,
         global_checkpoint_period=10,
         export_formats=None,
-        max_failures=3,
+        max_failures=0,
         restore=None,
         search_alg=None,
         scheduler=None,

rllib/models/model.py

@@ -232,6 +232,10 @@ def restore_original_dimensions(obs, obs_space, tensorlib=tf):
         return obs
 
 
+# Cache of preprocessors, for if the user is calling unpack obs often.
+_cache = {}
+
+
 def _unpack_obs(obs, space, tensorlib=tf):
     """Unpack a flattened Dict or Tuple observation array/tensor.
 
@@ -243,7 +247,13 @@ def _unpack_obs(obs, space, tensorlib=tf):
 
     if (isinstance(space, gym.spaces.Dict)
             or isinstance(space, gym.spaces.Tuple)):
-        prep = get_preprocessor(space)(space)
+        if id(space) in _cache:
+            prep = _cache[id(space)]
+        else:
+            prep = get_preprocessor(space)(space)
+            # Make an attempt to cache the result, if enough space left.
+            if len(_cache) < 999:
+                _cache[id(space)] = prep
         if len(obs.shape) != 2 or obs.shape[1] != prep.shape[0]:
             raise ValueError(
                 "Expected flattened obs shape of [None, {}], got {}".format(