Hi Pierre,

Using the partial_fit function, past learning of the model will be restored and the same model will be trained with the new batch of data. You can first call partial_fit function with 100,000 instances, then for re-train the new-batch of 10,000 instances using partial_fit.

Please refer to: https://datascience.stackexchange.com/questions/68599/incremental-learning-with-sklearn-warm-start-partial-fit-fit