[Google_Bootcamp_Day10]

Tuning Process

Hyperparameters

Try random values: Don’t use a grid

• grid : try all, then choose the best (not works well in deep learning)
• random : choose random points, try all, then choose the best (zoom in again, then sample more densely)

Appropriate scale for hyperparameters (not uniformly random)

• Use log scale

[Source] https://www.coursera.org/learn/deep-neural-network