Learning Rate - Is a higher learning rate better or worse?

A high learning rate can cause the network to bypass the minimum of the loss function and diverge, leading to unstable and inaccurate predictions. Conversely, a low learning rate might cause the network to converge slowly or get trapped in a local minimum, resulting in subpar and sluggish performance.