Implementations of sklearn.linear_model.SGDClassifier has learning rate argument. Becuase using SGDClassifier we can implement most of the algorithm by passing the loss using GD method.

The implementation of sklearn.linear_model.LogisticsRegressio is more like a LinearRegression by finding optimal weitghts. I cant explain much but This is my understanding.

Implementations of sklearn.linear_model.SGDClassifier has learning rate argument. Becuase using SGDClassifier we can implement most of the algorithm by passing the loss using GD method.

The implementation of sklearn.linear_model.LogisticsRegressio is more like a LinearRegression by finding optimal weitghts. I cant explain much but This is my understanding.