C1W4 Exercise 6

Course Q&A Practical Data Science Analyze Datasets and Train ML Models using AutoML
1

I’m stumped. The entire notebook runs, model works, all other exercises are given full points in grading, but Exercise 6 grades as wrong. Have this:

data_channels = {
### BEGIN SOLUTION - DO NOT delete this comment for grading purposes
‘train’: train_s3_uri, # Replace None
‘validation’: validation_s3_uri # Replace None
### END SOLUTION - DO NOT delete this comment for grading purposes
}

What am I missing???

As an aside, I have to say that the lab organization in some ways is quite confusing. the data_channels dict is discussed, but the next two exercises want simple uris (not from the data_channel dict, and the data_channel_dict is then defined LATER – why introduce it and then wait til two exercises later to define it?? My inclination was to define it immediately and then use the data_channels[‘train’] and data_channels[‘validation’] in the next two exercises. That would of course work in practice, but fails the grader. This sort of ‘you must do it in the way I’m thinking, even if my thinking is round-about’ is frustrating and has learners who fully understand what’s needed wasting time trying to satisfy the auto-grader. :frowning:

Thanks.

2

but the next two exercises want simple uris (not from the data_channel dict, and the data_channel_dict is then defined LATER

That’s where you are wrong and you don’t get the marks for this exercise.

estimator.fit() requires input and accepts both train and validation. Those inputs HAVE to be sagemaker.inputs.TrainingInput() and they are referred to as some_dictionary[“train”] and some_dictionary[“validation”].

1 Like
3

Got it, thanks! Appreciate the tip.