parquet_to_recordio_protobuf.ipynb failed CI

Link to the notebook:
https://github.com/aws/amazon-sagemaker-examples/blob/master/advanced_functionality/parquet_to_recordio_protobuf/parquet_to_recordio_protobuf.ipynb

Error:

---------------------------------------------------------------------------
Exception encountered at "In [4]":
---------------------------------------------------------------------------
ModuleNotFoundError Traceback (most recent call last)
in
----> 1 from fastparquet import write
2 from fastparquet import ParquetFile
3
4 def save_as_parquet_file(dataset, filename, label_col):
5 X = dataset[0]

/opt/conda/lib/python3.7/site-packages/fastparquet/__init__.py in
3
4 from .thrift_structures import parquet_thrift
----> 5 from .core import read_thrift
6 from .writer import write
7 from . import core, schema, converted_types, api

/opt/conda/lib/python3.7/site-packages/fastparquet/core.py in
7 from thrift.protocol.TCompactProtocol import TCompactProtocol
8
----> 9 from . import encoding
10 from .

[...]

ModuleNotFoundError: No module named 'numba.experimental'

Open advanced_functionality/parquet_to_recordio_protobuf/parquet_to_recordio_protobuf.ipynb and inspect the failing In [4] cell, beginning with its fastparquet and numba imports. Run the notebook or its CI check to reproduce the ModuleNotFoundError; done means the notebook completes without this import failure.

Written by the indexing model from the issue text.