Pass in "input" to the invoke method.

Keyword arg should be "input" instead of "input_data" during the invoke method.

Author: amymsft

articles/machine-learning/how-to-use-batch-endpoint-sdk-v2.md

@@ -193,7 +193,7 @@ Using the `MLClient` created earlier, we'll get a handle to the endpoint. The en
     # invoke the endpoint for batch scoring job
     job = ml_client.batch_endpoints.invoke(
         endpoint_name=batch_endpoint_name,
-        input_data=input,
+        input=input,
         deployment_name="non-mlflow-deployment",  # name is required as default deployment is not set
         params_override=[{"mini_batch_size": "20"}, {"compute.instance_count": "4"}],
     )
@@ -222,4 +222,4 @@ ml_client.batch_endpoints.begin_delete(name=batch_endpoint_name)
 
 ## Next steps
 
-If you encounter problems using batch endpoints, see [Troubleshooting batch endpoints](how-to-troubleshoot-batch-endpoints.md).
+If you encounter problems using batch endpoints, see [Troubleshooting batch endpoints](how-to-troubleshoot-batch-endpoints.md).