diff --git a/unsloth/models/rl.py b/unsloth/models/rl.py index 350c8f9d41..b98545fcd2 100755 --- a/unsloth/models/rl.py +++ b/unsloth/models/rl.py @@ -155,6 +155,7 @@ def _patch_resume_from_checkpoint_memory(trainer_class): _unsloth_train_with_resume_guard._unsloth_resume_guard = True trainer_class.train = _unsloth_train_with_resume_guard + def PatchRL(FastLanguageModel): try: from trl.models.utils import unwrap_model_for_generation