We created three NN models for an image classification task. The following table shows the loss metrics after training each model for around 1000 epochs (similar to what you did in Assignment#8). NN models Model 1 Model 2 Model 3 Training Loss Test Loss 0.20 0.25 0.23 0.22 0.22 0.20 Which model should we choose to have better predictions on new data points?