Conv1d for sentences classification models

I am new to Pytorch and trying to figure out how to create a CNN model for Sentences Classification problem with input data has (1, 1, nb_examples, max_length_sentences) dimensions and the labels whose dimensions are (nb_examples, nb_categories). How can I setup the Conv1d layers for input data mentioned above and does Pytorch take row by row of the input data to feed into the network to continuously?
Thanks in advance!