stand-alone rl (#42)

Author: hjh0119

cookbook/legacy/grpo/lora.py

@@ -315,6 +315,7 @@ def create_dataset():
 
 
 def train():
+    raise NotImplementedError("Not implemented")
     nproc_per_node, actor_ranks, ref_ranks = parse_device_config()
 
     device_groups = create_device_groups(actor_ranks, ref_ranks)