Early stopping training using Validation loss as the metric for best model

[Chandula]

I found that the correct parameter is “eval_loss” for the validation loss