XuandongZhao / WatermarkAttacker

Invisible Image Watermarks Are Provably Removable Using Generative AI

Home Page:https://arxiv.org/abs/2306.01953

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

No module diffusers

nhandang-ai4ia opened this issue · comments

Thank you for the awesome project. I have tried following the demo notebook and got the error in importing the module diffusers. I have tried downgrade the version to 0.18.2 but cannot import the ReSDPipeline.
I then check the code and see the ReSDPipeline seems to be the custom pipeline. Can you guide me further on how to solve this?
Thank you

You need to install the modified diffusers in our repo.
Did you try pip install -r requirements.txt and pip install -e .?

Yes i did, the import only failed at diffusers

Can you uninstall the diffusers and install it again?

I downgrade it to version 0.18.2 then i can import the diffusers, but i cannot imort the custom ReSDPipeline. Install the pip requirements get me the version: 0.20.0.dev0.
I run on google colab

Seems like i manage to know why the error. It simply because of the system path. When i change from diffusers to src.diffusers, it now able to start import the module successfully. I still need to track down the path more.
Thank you for the support

After changing all diffusers to src.diffusers for all module, i am now able to successfully import all. I will now test if i can reproduce the results from the demo notebook.

I did manage to successfully run the demo notebook after changing the path as mentioned above.
For RIVAGAN, i got the following error ValueError: This ORT build has ['AzureExecutionProvider', 'CPUExecutionProvider'] enabled. Since ORT 1.9, you are required to explicitly set the providers parameter when instantiating InferenceSession. For example, onnxruntime.InferenceSession(..., providers=['AzureExecutionProvider', 'CPUExecutionProvider'], ...). Do you know how to fix this?
If i want to show the image after adding watermark and after removing the watermark, which line of codes i need to run or modify, can you guide me?
Thank you

You may need to check whether you can successfully run https://github.com/ShieldMnt/invisible-watermark

For images after adding the watermark and after removing the watermark, you need to load the images in the examples folder.

I successfully run this awesome project except for the RivaGan, thank you for the great support. I will close this issue