Skip to content

Latest commit

 

History

History

GANalyze_instructions

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
mmnet.py
mmnet.py
 
 

README.md

Preparation:

Clone GANalyze via:

git clone /~https://github.com/LoreGoetschalckx/GANalyze.git
cd GANalyze

Download pre-trained generators:

cd pytorch; sh download_pretrained.sh

Add mmnet.py and pre-trained MachineMem/HumanMem predictors (available at this link) to GANalyze/pytorch/assessors/.

Training:

For MachineMem GANalyze:

python train_pytorch.py --assessor mmnet

For HumanMem GANalyze: Simply change machinemem_predictor to humanmem_predictor in mmnet.py (line 29).

python train_pytorch.py --assessor mmnet

Testing:

Run:

python test_pytorch.py --checkpoint_dir [path to your checkpoint folder]

Modify your checkpoint path to swap between HumanMem and MachineMem. Make sure the checkpoint is consistent with the weights you are using in mmnet.py (machine or human).

Cite:

If you use something from GANalyze, you may cite:

@inproceedings{goetschalckx2019ganalyze,
  title={Ganalyze: Toward visual definitions of cognitive image properties},
  author={Goetschalckx, Lore and Andonian, Alex and Oliva, Aude and Isola, Phillip},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={5744--5753},
  year={2019}
}