Course 3 week 1 exercise 6 - NameError: unbound axis name: Embedding_9088_256

My Lab Id: nvsuhhouimys

I’m getting following error when I try to run below code in ’ Exercise 6 - train_model’

=========
‘training_loop = train_model(model, train_task, [eval_task], 100, output_dir_expand)’

Error stack trace is:
File […]/site-packages/jax/core.py, line 1681, in axis_frame
f’unbound axis name: {axis_name}. The following axis names (e.g. defined ’

NameError: unbound axis name: Embedding_9088_256. The following axis names (e.g. defined by pmap) are available to collective operations:

Any idea why?

Thank You!

Hi @Ratnakar_Palle,

You have an issue in your Ex 5. Your implementation of the mean_layer is not as expected.

Best,
Mubsi

Thank you Mubsi for the edit. Now, I see it is passing.

UNQ_C5 (UNIQUE CELL IDENTIFIER, DO NOT EDIT)

====

fyi, that cell was marked ‘do not edit’, it came with assignment.

Hi @Ratnakar_Palle,

This comment is not about the code cell as a whole, it is regarding that particular line only “UNQ_C5 (UNIQUE CELL IDENTIFIER, DO NOT EDIT)”.

Basically, the autograder looks for the tags such as this, UNQ_C5, when it looks for the cells it needs to grade. If you remove this line, the autograder breaks. Hence, as a precaution it is mention in the assignments to not remove these tags.

Best,
Mubsi