TF! C4W1 bad compute metrcis

Course Q&A TensorFlow Developer Professional Sequences, Time Series and Prediction
1

i am getting some strange output from the comput ementrics near the end. please excude my bad typeing as my vsion is real poor. thanks

also. i ended up using the numner 361 in one of the calculations. why does that number workk? (assuming that it is correct).

converting the metrics to numpy floats improves the “metrics are numpy numeric types: True” (was false) , but i still get the same error.

Compute evaluation metrics

mse, mae = compute_metrics(series_valid, diff_moving_avg_plus_past)

print(f"mse: {mse:.2f}, mae: {mae:.2f} for moving average plus past forecast")

mse: 95.75, mae: 7.56 for moving average plus past forecast

Expected Output:

mse: 8.50, mae: 2.33 for moving average plus past forecast
2

Please see this

3

that post was about the last cell, my problem is before that at the third to the last cell.

i used the number 310 to get the past series. can anyone explain where that number came from? thanks

4

Thanks for clarifying.

One thing to keep in mind is that the shapes of the validation series and the generated series has to match. If this doesn’t help, please click my name and message your notebook as an attachment.

5

past_series has to be created correctly. Please see this hint:
Now let's bring back the trend and seasonality by adding the past values from t – 365:

6

i was able to finish the course and the specialization. thank you for your assistance…