A question about training different datasets without initialization

[fjjjj1229]

How to make the model continuously train on two datasets. Specifically, how to continue training on the second data without initialization after training on the first one.

Thanks

[chaojiewang94]

for bayesian generative models, there is a var named is_initial_local to control the initialization of local params, and u can set this var as False when u want to retrain the model on the second dataset, thx

def train(self, data:np.ndarray, num_epochs: int=1, is_train: bool = True, is_initial_local: bool=True):