Autopilot experiment fails to read the data.

Question

I have a CSV file, on S3, in the same region, but when I'm trying to run an autopilot experiment, I'll get the following error:

"Unable to fetch column headers.
Your dataset file must be a .csv or .parquet file and currently, be in your S3 bucket."

Can someone please help me with how I should fix this issue?

Thank you

Answer

This has to do with your CSV file. can you download and open your CSV to check if it has headers? whats the delimiter? is that delimiter allowed by autopilot/sagemaker? are you trying it via the API or console? Do you get the same error while trying from both ways? looking at these areas will help you narrow the problem space. it would also help to share some screenshots of the error/console to help debug further.

Autopilot experiment fails to read the data.

相關內容