Skip to content

shahabty/PSPNet-Pytorch

Repository files navigation

PSPNet-Pytorch

This Repo contains an implemetation of "Pyramid Scene Parsing Network" in Pytorch. Pretrained weights are converted from Official Caffe Repo. The performance is: [acc 0.95196], [acc_cls 0.83369], [mean_iu 0.75547], [fwavacc 0.91172] on Validation set of CityScapes Dataset.

Installation

  1. Install Anaconda3 from Here
  2. Create a Conda Environment:
username@PC:~$ conda env create -f environment.yml
  1. Download Cityscapes Dataset from Here
  2. Download Caffe Pretrained from Here and put it in Caffe-PSPNet folder
  3. Run the code:
username@PC:~$ python main.py

Qualitative

Note

Weight conversion from Caffe to Pytorch is modified from pytorch-semseg.
Preprocessing and loss function is modified from pytorch-semantic-segmentation.
Since the differences are significant, I decided not to add a branch to any of the the above project.

About

Implemetation of Pyramid Scene Parsing Network in Pytorch

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages