refactor(rl): add make_algorithm_config function for RLAlgorithmConfig instantiation

src/lerobot/rl/__init__.py

@@ -24,7 +24,10 @@ require_package("grpcio", extra="hilserl", import_name="grpc")
 
 from .algorithms.base import RLAlgorithm as RLAlgorithm
 from .algorithms.configs import RLAlgorithmConfig as RLAlgorithmConfig, TrainingStats as TrainingStats
-from .algorithms.factory import make_algorithm as make_algorithm
+from .algorithms.factory import (
+    make_algorithm as make_algorithm,
+    make_algorithm_config as make_algorithm_config,
+)
 from .algorithms.sac import SACAlgorithm as SACAlgorithm, SACAlgorithmConfig as SACAlgorithmConfig
 from .buffer import ReplayBuffer as ReplayBuffer
 from .data_sources import DataMixer as DataMixer, OnlineOfflineMixer as OnlineOfflineMixer
@@ -35,6 +38,7 @@ __all__ = [
     "RLAlgorithmConfig",
     "TrainingStats",
     "make_algorithm",
+    "make_algorithm_config",
     "SACAlgorithm",
     "SACAlgorithmConfig",
     "RLTrainer",

src/lerobot/rl/algorithms/factory.py

@@ -20,5 +20,28 @@ from lerobot.rl.algorithms.base import RLAlgorithm
 from lerobot.rl.algorithms.configs import RLAlgorithmConfig
 
 
+def make_algorithm_config(algorithm_type: str, **kwargs) -> RLAlgorithmConfig:
+    """Instantiate an :class:`RLAlgorithmConfig` from its registered type name.
+
+    Args:
+        algorithm_type: Registry key of the algorithm (e.g. ``"sac"``).
+        **kwargs: Keyword arguments forwarded to the config class constructor.
+
+    Returns:
+        An instance of the matching ``RLAlgorithmConfig`` subclass.
+
+    Raises:
+        ValueError: If ``algorithm_type`` is not registered.
+    """
+    try:
+        cls = RLAlgorithmConfig.get_choice_class(algorithm_type)
+    except KeyError as err:
+        raise ValueError(
+            f"Algorithm type '{algorithm_type}' is not registered. "
+            f"Available: {list(RLAlgorithmConfig.get_known_choices().keys())}"
+        ) from err
+    return cls(**kwargs)
+
+
 def make_algorithm(cfg: RLAlgorithmConfig, policy: torch.nn.Module) -> RLAlgorithm:
     return cfg.build_algorithm(policy)

src/lerobot/rl/train_rl.py

@@ -21,6 +21,7 @@ from dataclasses import dataclass
 from lerobot.configs.default import DatasetConfig
 from lerobot.configs.train import TrainPipelineConfig
 from lerobot.rl.algorithms.configs import RLAlgorithmConfig
+from lerobot.rl.algorithms.factory import make_algorithm_config
 from lerobot.rl.algorithms.sac import SACAlgorithmConfig  # noqa: F401
 
 
@@ -45,8 +46,7 @@ class TrainRLServerPipelineConfig(TrainPipelineConfig):
         super().validate()
 
         if self.algorithm is None:
-            sac_cls = RLAlgorithmConfig.get_choice_class("sac")
+            self.algorithm = make_algorithm_config("sac")
-            self.algorithm = sac_cls()
 
         # The pipeline owns the policy config; inject it so the algorithm can
         # introspect policy architecture (e.g. ``num_discrete_actions``).