Currently the Neural Network Predictor defaults to AdamW optimizer with L1Loss, however we would like to be able to customize this.
Currently the Neural Network Predictor defaults to AdamW optimizer with L1Loss, however we would like to be able to customize this.