From 0de2fef1ebecb74bc33e0e4c1deef02dea6c1839 Mon Sep 17 00:00:00 2001 From: Daniel Han Date: Wed, 5 Feb 2025 06:28:54 -0800 Subject: [PATCH] Update rl.py --- unsloth/models/rl.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/unsloth/models/rl.py b/unsloth/models/rl.py index c4a835ed93..932a29f78a 100644 --- a/unsloth/models/rl.py +++ b/unsloth/models/rl.py @@ -39,12 +39,12 @@ def PatchRL(FastLanguageModel): @contextmanager def unsloth_unwrap_model_for_generation(model, *args, **kwargs): # Must use for_inference to allow inference in Unsloth - print("$$$$$$$$$$$$$$$$$$$$$$$") with unwrap_model_for_generation(model, *args, **kwargs) as unwrapped_model: FastLanguageModel.for_inference(unwrapped_model) yield unwrapped_model # Return back to training mode FastLanguageModel.for_training (model) + yield model pass import trl.trainer