Self-Attention-Based Edge Computing Model for Synthesis Image to Text through Next-Generation AI Mechanism

Table 2

Comparison of changes in different hyperparameter indicators.

Method	IS	FID
SA-AttnGAN, λ = 0.1	4.05 ± 0.05	30.25
SA-AttnGAN, λ = 1	4.34 ± 0.02	22.60
SA-AttnGAN, λ = 5	4.52 ± 0.03	14.25
SA-AttnGAN, λ = 10	4.24 ± 0.05	25.56