Error in batch_inference_with_file_data function in pipeline_with_parallel_nodes example

### Operating System

MacOS

### Version Information

Python Version: python=3.7.6
azure-ai-ml package version: 1.5.0

### Steps to reproduce

I am running the [pipeline_with_parallel_nodes.ipynb](https://github.com/Azure/azureml-examples/blob/main/sdk/python/jobs/pipelines/1g_pipeline_with_parallel_nodes/pipeline_with_parallel_nodes.ipynb) example. However, I encountered an error at the step batch_inference_with_file_data.

I have already checked if the output parameters are correctly set, do we need to explicitly define the data store? Can someone help me identify the issue and suggest a solution?

Thank you.

### Expected behavior

Compute: [GPU - 4 x NVIDIA Tesla K80](https://www.nvidia.com/en-gb/data-center/tesla-k80)

### Actual behavior

Error: azureml_common.parallel_run.exception.ParallelTaskException: ParallelTaskException error. Exit code: 42. Message: Cannot find account key or SAS token in the specified datastore, and credential passthrough is not enabled.

I used "az login" before running the notebook in Jupyter Notebook

### Addition information

_No response_

Start by running pipeline_with_parallel_nodes.ipynb and reproducing the batch_inference_with_file_data failure with the reported Python and azure-ai-ml versions. Inspect how the notebook configures its datastore credentials and authentication; done means the step completes without the ParallelTaskException about a missing account key or SAS token.

Written by the indexing model from the issue text.