Upload train_grpo.py with huggingface_hub

train_grpo.py

@@ -270,7 +270,7 @@ def main():
     trainer = GRPOTrainer(
         model=model,
         processing_class=tokenizer,
-        config=config,
+        args=config,
         train_dataset=dataset,
         reward_funcs=reward_fn,
     )