"Host a Pretrained Model on SageMaker" tutorial fails with "Received server error (500) from primary and could not load the entire response body." about sagemaker-python-sdk HOT 5 CLOSED

I was able to isolate the issue to the model_fn function:

from sagemaker-python-sdk.

Not sure if this is related on not, but the tokenizer.save_pretrained() call is skipping one file locally:

4 files added -->

only 3 files show up locally -->

from sagemaker-python-sdk.

I wrapped model_fn in a try catch and here is the error:

from sagemaker-python-sdk.

Hi, do you have the requirements.txt in code folder? https://github.com/aws/amazon-sagemaker-examples/tree/main/sagemaker-script-mode/pytorch_bert/code
In my test without this file it would have this ModelError

from sagemaker-python-sdk.

Closing this issue since it's not for python sdk. Track the same issue in example-notebooks repo: aws/amazon-sagemaker-examples#4551

from sagemaker-python-sdk.