bytetrack-training failing on 2.1 Convert SageMaker Ground Truth annotation into MOT Challenge annotation

[J-Korn]

I am trying to run the bytetrack-training notebook following the "Implement a multi-object tracking solution on a custom dataset with Amazon SageMaker" guide found here https://aws.amazon.com/blogs/machine-learning/implement-a-multi-object-tracking-solution-on-a-custom-dataset-with-amazon-sagemaker/ and it fails on the Create a SageMaker Processing job to convert annotation step of 2.1 Convert SageMaker Ground Truth annotation into MOT Challenge annotation
I have already ran data-preparation notebook and have manually labeled the video frames, ending up with the s3://sagemaker-studio-595698581270-lybz9w5v4l/Test/sample-data-gt/dataset-training/manifests/output/output.manifest file, which looks like this:

{"source-ref":"s3://sagemaker-studio-595698581270-lybz9w5v4l/Test/sample-data/sample_0.mp4/sample_0.mp4-sequence-1690194891.json","dataset-training-ref":"s3://sagemaker-studio-595698581270-lybz9w5v4l/Test/sample-data-gt/dataset-training/annotations/consolidated-annotation/output/0/SeqLabel.json","dataset-training-ref-metadata":{"class-map":{"0":"Person","1":"Car"},"job-name":"labeling-job/dataset-training","human-annotated":"yes","creation-date":"2023-07-24T11:35:27.090000","type":"groundtruth/video-object-tracking"}}
etc...

This is different than the image shown at the end of the data-preparation notebook that shows the manifest having a mot-bytetrack2-ref field and I suspect this is why the bytetrack-training notebook fails with error FileNotFoundError: [Errno 2] No such file or directory: '/opt/ml/processing/gt/manifests/output/output.manifest'