Hi,
i have a few questions about the normalization in Example 01: train a network for regression.
a. Is it correct that only the training data is normalized in this example?, i would expect that the validation- & test-data should be normalized as well?
b. how is the normalization performed exactly, do you normalize the noisy & noiseless (training) data separately (calculating and applying mean & std of both sets separately), or do you calculate a single mean & std-value that is applied to both?
many thanks in advance,
best regards,
Rolf
Hi,
i have a few questions about the normalization in Example 01: train a network for regression.
a. Is it correct that only the training data is normalized in this example?, i would expect that the validation- & test-data should be normalized as well?
b. how is the normalization performed exactly, do you normalize the noisy & noiseless (training) data separately (calculating and applying mean & std of both sets separately), or do you calculate a single mean & std-value that is applied to both?
many thanks in advance,
best regards,
Rolf