This is intended to allow functionality like https://www.tensorflow.org/api_docs/python/tf/train/ExponentialMovingAverage
class EMACallback[source]
EMACallback(decay=0.9999,use_ema_weights:bool=True) ::Callback
Model Exponential Moving Average. Empirically it has been found that using the moving average of the trained parameters of a deep network is better than using its trained parameters directly.
If use_ema_weights, then the ema parameters of the network is set after training end.