Why do we need to move image to torch.float32 in FluxPipeline?

[Podidiving]

👋

I dont quite understand the logic behind moving init_image to torch.float32 format

diffusers/src/diffusers/pipelines/flux/pipeline_flux_img2img.py

Line 829 in 026507c

init_image = init_image.to(dtype=torch.float32)

because next the variable is used in prepare_latents method where it is moved to another dtype. Is this a legacy from stable diffusion vae model or is there some other idea behind it?