Hello, is anybody going to implement "Don't Decay the Learning Rate, Increase the Batch Size" : https://arxiv.org/abs/1711.00489 from Google in Keras ? If not, is there a way to put custom batch sizes in a single epoch ?
instead of using model.fit, use model.train_on_batch so that you can control the batch size.
Most helpful comment
instead of using
model.fit, usemodel.train_on_batchso that you can control the batch size.