How does ReduceLROnPlateau work?
easyAnswer
- Track a validation metric; if it doesn't improve for 'patience' epochs, multiply LR by a factor (e.g., 0.5).
- Adaptive to the actual loss curve — doesn't require guessing the schedule in advance.
- Good default for prototyping.
- Downside: introduces validation feedback into training, and its aggressiveness (patience, factor) matters a lot.
Check yourself — multiple choice
- Reduces LR every step
- Reduces LR when val metric plateaus for N epochs — adaptive, feedback-driven
- Same as cosine
- Never used in practice
ReduceLROnPlateau: cut LR when val metric stagnates for patience epochs.
#schedules#learning-rate
Practise Deep Learning
214 interview questions in this topic.