Does the document_embedding benchmark require special S3 permissions to access the dataset?

[QLiangong]

When I run the document_embedding benchmark (daft_main.py), I find that accessing certain S3 object paths like:
s3://eventual-data-test-bucket/digitalcorpora/pdfs/…
results in an AccessDenied error.
Does this mean that I do not have permission to access this bucket?
If so, what do I need to do to obtain the correct permissions so that I can access these objects during the benchmark run?

[kevinzwang]

Hi @QLiangong, we've updated the benchmark to point to a publicly available S3 bucket. Closing this discussion for now, feel free to submit a new one if you have any issues running the benchmark now!