ott.tools.gaussian_mixture.fit_gmm_pair.get_fit_model_em_fn# ott.tools.gaussian_mixture.fit_gmm_pair.get_fit_model_em_fn(weight_transport, learning_rate=0.001, jit=True)[source]# Get a function that performs penalized EM. We precompile and precompute a few quantities that we put into a closure. Parameters: weight_transport (float) – weight for the transportation loss in the total loss learning_rate (float) – learning rate to use for the Adam optimizer jit (bool) – if True, precompile key methods Returns: A function that performs generalized, penalized EM.
