THUDM / ImageReward

[NeurIPS 2023] ImageReward: Learning and Evaluating Human Preferences for Text-to-image Generation

Apache License 2.0

1.15k stars 63 forks source link

fine-tuning: NotImplementedError #100

Open fenglincong opened 1 week ago

fenglincong commented 1 week ago

pip install image-reward

pip install diffusers==0.16.0 accelerate==0.16.0 datasets==2.11.0

from ImageReward import ReFL args = ReFL.parse_args() trainer = ReFL.Trainer("CompVis/stable-diffusion-v1-4", "data/refl_data.json", args=args) trainer.train(args=args)

when fine-tuning the model, I encountered a bug: NotImplementedError: Loading a dataset cached in a LocalFileSystem is not supported.

xujz18 commented 1 week ago

Thanks for your attention! This may because the version of datasets, you may refer to this issue.