Please provide the subset of COCO dataset that you have used

facebookresearch / stable_signature

Official implementation of the paper "The Stable Signature Rooting Watermarks in Latent Diffusion Models"

Other

385 stars 48 forks source link

Please provide the subset of COCO dataset that you have used #23

Closed neuralnaresh closed 2 months ago

neuralnaresh commented 5 months ago

Dear Authors,

I am trying to reproduce the results mentioned in the Stable signature paper and I request you to kindly provide the subset of the coco dataset that you have used.

Hoping to hear from you soon.

Best Regards.

pierrefdz commented 5 months ago

Hi, we use the 2014 splits:

for the training, we use the train subset
for evaluation, we use the validation subset

Don't hesitate if you have further questions

ChelsyHu commented 4 months ago

You used the all images in COCO2014 training and validation dataset for training the extractor in stable signature, right?

pierrefdz commented 4 months ago

Hi! Yes, we used the whole dataset for training, the validation dataset is only used for FID (since the extraction is done on images generated from COCO prompts)