Why the same data set, the same parameters, the final attack success rate and the number of queries are different? - Githubissues

MadryLab / blackbox-bandits

Code for "Prior Convictions: Black-Box Adversarial Attacks with Bandits and Priors"

https://arxiv.org/abs/1807.07978

MIT License

61 stars 14 forks source link

Why the same data set, the same parameters, the final attack success rate and the number of queries are different? #10

Open Kitzzaaa opened 2 years ago

Kitzzaaa commented 2 years ago

Why the same data set, the same parameters, the final attack success rate and the number of queries are different?