Emotional Dialogue Generation Based on Conditional Variational Autoencoder and Dual Emotion Framework

Table 4

Manual evaluation of the generated responses in terms of content and emotion.

Model	Content	Emotion
Seq2Seq [21]	1.258	0.154
ECM [14]	1.286	0.422
CVAE [16]	1.302	0.392
CVAE-MTDA [25]	1.325	0.405
EDGAN [27]	1.314	0.416
CVAE-DE	1.337	0.437