You can find a selection of articles, blog posts, models, and other resources about deep learning on this page.
A comprehensive overview of deep learning architectures can be found at: Towards Data Science. It covers all the important architectures, such as:
Batch sizes over 32 are not useful to increase performance, "…the best performance has been consistently obtained for mini-batch sizes between m=2 and m=32 , which contrasts with recent work advocating the use of mini-batch sizes in the thousands."
Deconvolution/Transposed convolution can lead to checkerboard patterns especially in strong colors because of kernels overlapping and maybe because of gradient artifacts.