Skip to content
master
Switch branches/tags
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 

README.md

DSCLRCN-PyTorch

PyTorch implementation of https://arxiv.org/abs/1610.01708

  • The main file used to run training and testing is 'DSCLRCN_exp.ipynb'. Please have a look at it

  • All the pretrained models used to build the architecture of the network are available as PyTorch loadable models under the following GDrive link: https://drive.google.com/open?id=1aQLyS_kz1Uuj6wP9fl_qX0gpNm2n6LQ5 Please download them and copy them into their corresponding folders

  • The dataset should be arranged in the directory 'Dataset/Transformed' as three dictionaries ('train_datadict', 'val_datadict' and test_datadict') each containing the images and fixation maps resized to 92x128; in addition to a mean image computed from the training dataset for normalization. Please have a look at 'util/data_utils.py' file

  • The original sources of the models: PlacesCNN: http://places.csail.mit.edu/downloadCNN.html Segmentation ResNet50 Encoder: https://github.com/hangzhaomit/semantic-segmentation-pytorch VGG16 (available with PyTorch 'torchvision' library)

About

PyTorch implementation of https://arxiv.org/abs/1610.01708

Resources

Releases

No releases published

Packages

No packages published