Generating images with IC-GAN and using my own dataset
BehzadBozorgtabar opened this issue · 1 comments
Hi,
I followed the instructions to extract the features from my own test images and format the dataset to get the hdf5 files as explained in "Other datasets", and only could generate hdf5 files in the format of [dataset_name][resolution]feats[feature_extractor]_resnet50.hdf5 not the other two.
My question is how to generate images using these generated instances features (in .hdf5 format)? How should generate.images.py be modified for image generation using my own test data? Can you please clarify this part?
Thanks
Hello, what do you mean by "the other two"? Can you please specify which files are you referring to?
You could modify generate_images.py
here to load instead the feature vectors from [dataset_name][resolution]feats[feature_extractor]_resnet50.hdf5
. Some inspiration to do that, here, where root_feats
would be the path to your features in hdf5 format.