Retrain a H2ORandomForestEstimator using different data whilst maintaining hyper parameters

Keywords´╝Ü random-forest h2o


I have saved an optimised H2ORandomForestEstimator model using h2o.save_model() (python API). I now want to load this model and re-train it with different variants of my data, whilst maintaining the optimised hyperparams (e.g. ntrees, max_depth).

When I do this by calling train() on the loaded models however all the hyperparams appear to be reset to their defaults. What is the recommended way to achieve this?