I’m encountering an issue with the grading of the Sequences, Time Series, and Prediction (TensorFlow Course 4, Week 1) programming assignment: Working with Generated Time Series.
Despite completing the tasks, all my results are being evaluated as 0/10, and I receive the following error:
Error: “There was a problem compiling the code from your notebook, please check that you saved before submitting. Details: name ‘series_valid’ is not defined.”
I’ve confirmed that series_valid is defined in the notebook. I’ve tried the following steps to resolve the issue:
Updated the notebook and started over
Followed the debugging steps from a similar post on the forum
However, the notebook still scores 0 for all tasks. Has anyone else faced this issue, or does anyone have suggestions for how to resolve it?
No, I didn’t use anywhere else and also adding to my query later, when I tried, I was getting grading error and also instead of above-mentioned error, now I am getting this one There was a problem compiling the code from your notebook; please check that you saved before submitting. Details: ‘NoneType’ object is not iterable
Yeah, I have already followed this but am still getting the same issue. This is the last lab remaining on which I am stuck and all other weeks after this are already done; only week 1 is there due to this lab.
grade function diff moving average.
For code line
Perform the correct slicing
You are suppose to mention the slicing using the SPLIT TIME AND WINDOW_SIZE instead of using len function of series valid as using len function of series would prevent the end range at which correct slicing would be required.
Grade function diff moving average plus past
For code line
Slice the whole SERIES to get the past values.
You want to get the value from the previous period for each forecasted value
past_series = @REPLACE EQUALS SERIES[None:None] the None again wants you to replace the values previous period.
Read this explanation again
Now let’s bring back the trend and seasonality by adding the past values from t – 365. For each value you want to forecast, you will be adding the exact same point, but from the previous cycle in the original time series.
t is your SPLIT TIME, the past value is t-365 and the previous cycle would be again 365.
this could be exactly how you mention for naive forecast except for the past value cycle here being 365.
These both issue were mentioned in the link comment I previously shared with you.