How to modify the nn:Sequential type parameter

Hi everyone ,
I want modify a nn:Sequential type in my project , In python ,I write the following code and it works:

temp_model = models.vgg19(pretrained=True)
feature_list = temp_model.features
features = list(feature_list)[:33]  # get vgg features

I mean i want use some layers of the pretrained model,How can i implement in libtorch?
I have already write this code:

    vision::models::VGG19 models = vision::models::VGG19(1000, false);
    auto features = models->features;

Thanks a lot!