Enable elastic training for RL

PiperOrigin-RevId: 901000964

Author: abhinavclemson

src/maxtext/trainers/post_train/rl/train_rl.py

@@ -44,7 +44,7 @@
 """
 
 from __future__ import annotations
-from functools import wraps
+from functools import wraps, partial
 from typing import Sequence
 
 import collections
@@ -78,7 +78,7 @@
 from maxtext.trainers.post_train.rl.evaluate_rl import evaluate
 from maxtext.trainers.post_train.rl import utils_rl
 from maxtext.input_pipeline.instruction_data_processing import load_template_from_file
-from maxtext.utils import max_logging, max_utils, maxtext_utils, model_creation_utils
+from maxtext.utils import max_logging, max_utils, maxtext_utils, model_creation_utils, elastic_utils
 
 
 def get_maxtext_model(config, devices=None):
@@ -770,7 +770,19 @@ def main(argv: Sequence[str]) -> None:
 
   max_utils.print_system_information()
   trainer_config, sampler_config, trainer_devices, sampler_devices = setup_configs_and_devices(argv)
-  rl_train(trainer_config, sampler_config, trainer_devices, sampler_devices)
+  
+  if trainer_config.elastic_enabled:
+    max_logging.log("Elastic utils: Elastic training enabled.")
+
+    def elastic_train_wrapper(argv: Sequence[str]) -> None:
+      """Wrapper for elastic training initializes variables and runs the train loop."""
+      t_config, s_config, t_devices, s_devices = setup_configs_and_devices(argv)
+      rl_train(t_config, s_config, t_devices, s_devices)
+
+    train_func = elastic_utils.elastic_retry(trainer_config)(functools.partial(elastic_train_wrapper, argv=argv))
+    train_func()
+  else:
+    rl_train(trainer_config, sampler_config, trainer_devices, sampler_devices)
 
 
 if __name__ == "__main__":