Skip to content

Trying to get large receptive fields in a computationally efficient way with meaningful effective receptive fields!

Notifications You must be signed in to change notification settings

fabioseel/receptive-fields

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Receptive Fields

Elaborate various mechanisms to visualize receptive fields. In particular, this project aims at inspecting how meaningful respective fields can be retrieved from small-scale CNNs when using:

  • stride
  • dilation
  • separable convolutions
  • ...

Installation

Needed packages:

  • pytorch
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
  • matplotlib
conda install matplotlib
  • tqdm
conda install tqdm
  • ipykernel (conda)
  • torchinfo (pip)

About

Trying to get large receptive fields in a computationally efficient way with meaningful effective receptive fields!

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages