SVM using cosine kernel - dataset with images of dogs and cats

Question

Hello I am trying to implement SVM by using cosine kernel but I can't undertand how I can do this..

What i thought it was the following, but i think its wrong

svmCosine = cosine_similarity(train_X, train_y)
svmCosine.fit(train_X, train_y)

Could someone help?

Answer 1

I think that the answer to my question is the following

svmCosine= svm.SVC(kernel = cosine_similarity(train_X, train_y))
svmCosine.fit(train_X, train_y)