diff --git a/unsloth/models/vision.py b/unsloth/models/vision.py index 4466128a28..f17096af73 100644 --- a/unsloth/models/vision.py +++ b/unsloth/models/vision.py @@ -690,6 +690,8 @@ class FastBaseModel: embeddings = model.get_output_embeddings() if hasattr(embeddings, "training"): embeddings.training = False pass + # Must disable returning hidden states in the case for GRPO + os.environ["UNSLOTH_RETURN_HIDDEN_STATES"] = "0" return model pass