Closed marciofreire closed 2 weeks ago
Hey @marciofreire Could you provide a reproducible script for this bug? We recently changed the parameter signature for the consistency across models. If you're using the latest version of KerasHub, you should be able to instantiate SD3 like this:
text_to_image = StableDiffusion3TextToImage.from_preset(
"stable_diffusion_3_medium",
image_shape=(height, width, 3),
dtype="float16",
)
Hey James, thanks a lot for getting back to me!
I was working with the SD3 example from this link: "https://github.com/keras-team/keras-io/blob/master/guides/keras_hub/stable_diffusion_3_in_keras_hub.py".
I noticed you updated the SD3 example to account for the new parameter signature.
Best, Márcio
Great!! Since the issue was resolved, I will close the issue.
Describe the bug
To Reproduce
Expected behavior
Additional context
Would you like to help us fix it? I wish I had this knowledge.