Fig. 7From: Learning domain-heterogeneous speaker recognition systems with personalized continual federated learningComparison of error rates for C-PFL with different training data input sequencesBack to article page