Normalizing data for better Neural Network performance

Neural Networks and neural network based architecturres are powerful models that can deal with abstract problems but they are known for taking a long time to train. In this video, we learn how normalization can speed up neural network training.

❓To get the most out of the course, don't forget to answer the end of module questions:

👉 You can find the answers here:

RESOURCES:

COURSES:

[b.o.p.]

Very good explanation I like how you even used a graph to explain it.

[derekcaramella]

How do you generate the visualization, your explanation makes sense, but I am not sure what axes represent

[simplehacker_]

How would you go about normalising a set of enum values? Let's say I have a colour parameter with three choices, (Red, Green, Blue). Would you assign each of these a number (1, 2, 3) and then normalise between 0-1, so that the three colours are normalised to 0, 0.5, and 1?

[lakeguy]

how do you choose between scaling your inputs between 0 & 1 or normalizing with a mean of 0 and std dev of 1? So you scale or normalize your inputs, do you then use batch normalization? Always? sometimes? why? (another good video, thank you!)

[cryptobob]

hi when we say that out neural network is well tuned?if we used rmse for training data and validation data...which number is good for them...?