Training loop grader failure

My submission passed with 90% but I’d like to figure out the failure. It only says:

There was a problem grading your submission. Check stderr for more details.

I obviously don’t have access to stderr - is there a way to run these locally? UNQ_C8 seemed pretty trivial to implement so not sure where this is going wrong. All of the unit tests pass.

Thanks - Jim

Stderr is the output you see under “Show grader output”. I have not done NLP, but there are lots of ways to pass unit tests with code that is not general. One common way is to reference the global variable that happens to be passed as an argument to your function, rather than referencing it through the formal parameter.

Oh no, self referential error message! :smiley:

FYI it was because I was using the global train/eval generators. Sigh. Fixed and 100%.

Glad to hear you found the solution! Thanks for following up with the explanation for future reference.