- A learning rate scheduler: https://github.com/ultralytics/yolov3/issues/238
- Löffel, E. P. (2019). Improved Concept Embeddings for Learning Prerequisite Chains.

Examples of burn-in: Positive examples

- Markov chain Monte Carlo (MCMC): http://users.stat.umn.edu/~geyer/mcmc/burn.html

Examples of burn-in: Negative examples

University of Michigan - Ann Arbor

Burn-in is an initial phase before normally running your model. The epochs of burn-in is number of epochs to ramp learning rate from either 0 or 1 to your final learning rate. It is helpful to find good embeddings in some cases.

Learn Before

Related

Learn After