Question about loading an existing model to continue training

[deepsidhpura777]

In the documentation, it is mentioned that we can load an existing model by providing a model_filename parameter. If I call the crf.fit on a new training set, will it continue training or override the previously learnt parameters? I have multiple different data sets each with around 100,000 sequences, so I really need to make sure the feature of continuing training exists since my data sets don't fit the RAM. Thanks !!