Where to get a synset.txt file after training Sagemaker Image Classfification

Hello,

Here's what I am trying to do: I built two models using Sagemaker's built-in Image Classification algorithm. I want to deploy these two models in the same docker container, also on Sagemaker.

[this](https://docs.aws.amazon.com/sagemaker/latest/dg/multi-model-endpoints.html) entry in the manual tells me that to do so, I need to build my own docker container, as currently none of the pre-built containers support mutli-model image classification.

I then went to this tutorial: [this](https://github.com/awslabs/amazon-sagemaker-examples/blob/8c62b998e0a9b7f4886a932450e7606643a5c944/advanced_functionality/multi_model_bring_your_own/multi_model_endpoint_bring_your_own.ipynb) to build that container. Here's where I get stuck: Apparently I need something called a synset.txt file, which contains label for the training data?

I trained my models using Sagemaker's Augmented manifest format, as detailed [here](https://docs.aws.amazon.com/sagemaker/latest/dg/augmented-manifest.html). So neither my input, nor the output of Sagemaker contains such a file. Any suggestions of where I could get it? Details on what should be in that file would be welcome too. I understand I need to provide labeled datasets, similar to what the Augmented Manifest does, but I have been unable to find any clear explanation of this type of MxNet-File.

Start with the linked SageMaker multi-model endpoints documentation, augmented manifest documentation, and multi_model_endpoint_bring_your_own.ipynb. Trace where the notebook expects synset.txt and compare that requirement with the augmented manifest workflow. Done means the required file format and a documented way to obtain or construct it are clear.

Written by the indexing model from the issue text.