@nomewang ,thank you very much for sharing your code! However, I have a small question about the code.
In the code, The definition of OCSVM says that
self.detect_fuser = linear_model.SGDOneClassSVM(random_state=42, nu=args.ocsvm_nu, max_iter=args.ocsvm_maxiter)
self.seg_fuser = linear_model.SGDOneClassSVM(random_state=42, nu=args.ocsvm_nu, max_iter=args.ocsvm_maxiter)
However, in the paper says that "For DLF, we use two linear OCSVMs with SGD optimizers, the learning rate is set as 1×10−4 and train 1000 steps for each class."
So I wanna know how to set training rate in this code about OCSVM. Can you answer the question about it?
Thank you a lot again!
@nomewang ,thank you very much for sharing your code! However, I have a small question about the code.
In the code, The definition of OCSVM says that
self.detect_fuser = linear_model.SGDOneClassSVM(random_state=42, nu=args.ocsvm_nu, max_iter=args.ocsvm_maxiter)self.seg_fuser = linear_model.SGDOneClassSVM(random_state=42, nu=args.ocsvm_nu, max_iter=args.ocsvm_maxiter)However, in the paper says that "For DLF, we use two linear OCSVMs with SGD optimizers, the learning rate is set as 1×10−4 and train 1000 steps for each class."
So I wanna know how to set training rate in this code about OCSVM. Can you answer the question about it?
Thank you a lot again!