TADW: Traceable and Anti-detection Dynamic Watermarking of Deep Neural Networks

Table 9

Testing accuracy on clean models and watermarked models under different .

Model
	TextCNN (%)	TextRNN (%)	BERT (%)	TextCNN (%)	TextRNN (%)	BERT (%)
Clean	93.88	93.30	94.25	93.88	93.30	94.25
SN-10-1	93.46	92.80	93.83	93.46	92.83	93.91
SN-10-2	93.47	92.97	93.83	93.42	92.80	93.82
SN-10-5	93.59	92.86	94.00	93.39	92.88	93.78
SN-10-10	93.39	93.03	93.75	93.39	92.80	93.80