pytorch-gcn-tox21.ipynb failed CI

Link to the notebook:
https://github.com/aws/amazon-sagemaker-examples/blob/master/sagemaker-python-sdk/dgl_gcn_tox21/pytorch-gcn-tox21.ipynb

Error:

---------------------------------------------------------------------------
Exception encountered at "In [3]":
---------------------------------------------------------------------------
CalledProcessError Traceback (most recent call last)
in
----> 1 get_ipython().run_cell_magic('sh', '', '# For CPU default_docker_name="dgllib/dgl-sagemaker-cpu:dgl_0.4_pytorch_1.2.0_rdkit"\ndefault_docker_name="dgllib/dgl-sagemaker-gpu:dgl_0.4_pytorch_1.2.0_rdkit"\ndocker pull $default_docker_name\n\ndocker_name=sagemaker-dgl-pytorch-gcn-tox21\n\n# For CPU docker build -t $docker_name -f gcn_tox21_cpu.Dockerfile .\ndocker build -t $docker_name -f gcn_tox21_gpu.Dockerfile .\n\naccount=$(aws sts get-caller-identity --query Account --output text)\necho $account\nregion=$(aws configure get region)\n\nfullname="${account}.dkr.ecr.${region}.amazonaws.com/${docker_name}:latest"\n# If the repository doesn\'t exist in ECR, create it.\naws ecr describe-repositories --repos

Start with sagemaker-python-sdk/dgl_gcn_tox21/pytorch-gcn-tox21.ipynb and reproduce the failure in cell In [3]. Inspect the Docker build and AWS CLI commands shown in that cell, then run the notebook's CI check. Done means the notebook completes successfully in CI without the reported CalledProcessError.

Written by the indexing model from the issue text.