--- license: mit tags: - stable-diffusion - stable-diffusion-diffusers inference: false --- # Watermarked (strong) VAE for SDXL You can use this watermarked VAE for your existing SDXL image generation pipelines to make sure generated images are watermarked. ## Usage ```py from diffusers.models import AutoencoderKL from diffusers import StableDiffusionXLPipeline model = "stabilityai/sdxl-turbo" vae = AutoencoderKL.from_pretrained("imatag/stable-signature-bzh-sdxl-vae-strong") pipe = StableDiffusionXLPipeline.from_pretrained(model, vae=vae) ``` For more information, please have a look at [the official demo](https://huggingface.co/spaces/imatag/stable-signature-bzh) ## Purpose This model was finetuned from the original StableDiffusion-XL autoencoder with the additional objective of including a strong invisible watermark, following the procedure of [StableSignature](https://ai.meta.com/blog/stable-signature-watermarking-generative-ai/). This watermark is detectable by [IMATAG](https://www.imatag.com/)'s demo BZH decoder, available via API. Compared to other watermarking solutions such as [invisible-watermark](https://github.com/ShieldMnt/invisible-watermark): - the watermark is quite robust to many unintentional editing transforms such as cropping, resizing, changing contrast or brightness, and JPEG compression - the watermark has zero overhead in terms of computational power - the watermark is embedded in the model weights and not as a post-processing of the non-watermarked image - the probability of detecting the watermark by chance on non-watermarked images may be computed analytically and limited to a specific threshold ## Detecting the watermark For security reasons we cannot release the weights of the detector. Instead, you may check for the presence of a watermark with the [detect_api.py](https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/detect_api.py) script like this: ```shell python detect_api.py test.png ``` ### Visual _Visualization of watermark impact on 512x512 images from the COCO2017 validation dataset prompts with SDXL-turbo._


512x512: original (left), watermarked (middle), difference (right)

### Robustness The watermark is robust to most editorial changes to an image, such as cropping, changing brightness or contrast, resizing or JPEG compression:


from left to right: original, watermarked, difference, 'crop 50% + brigthen 50% + jpeg 80%' attack, 'downscale 2x + crop 50% + brigthen 50% + jpeg 50%' attack

The Receiver Operating Characteristic curve shows its performance under no modification and when modified by the two scenarios illustrated above:

More details available in our [announcement](https://www.imatag.com/blog/unlocking-the-future-of-content-authentication-imatags-breakthrough-in-ai-generated-image-watermarking) and our lab's [blog post](https://imatag-lab.medium.com/stable-signature-meets-bzh-53ad0ba13691). For watermarked models with a different key, support for payload, other perceptual compromises, robustness to other attacks, or faster detection, please [contact IMATAG](https://pages.imatag.com/contact-us-imatag).