Analysing different layers of CNN for transfer learning

vision
1

I am working on transfer learning. Is there any way to use different subsets of layers of target model and analyse the performance of these layers for smaller dimension and low resolution data compared to data used for target model.