This MR introduces the following changes
continue_trainingoption for all models. With that, you can continue a broken training. For example, if your Job was killed due to time constrains, set this option to
Trueand the latest model is loaded with weights, optimiser state and epoch number. This is then used for continuing the training.