DPMSolverSinglestepScheduler lacks self.sigmas during initialization

[LWprogramming]

Describe the bug

DPMSolverSinglestepScheduler not initialized with sigmas, but required for add_noise. only set after setting timesteps.

Reproduction

import torch
from diffusers import DPMSolverSinglestepScheduler

foo = DPMSolverSinglestepScheduler()
placeholder_latents = torch.randn(1, 64, 64)
placeholder_noise = torch.randn(1, 64, 64)
print(hasattr(foo, "sigmas")) # False
foo.add_noise(placeholder_latents, placeholder_noise, 10) # crash

Logs

No response

System Info

• diffusers version: 0.22.2
• Platform: Linux-5.19.17-coreweave-x86_64-with-glibc2.29
• Python version: 3.8.10
• PyTorch version (GPU?): 2.1.0a0+fe05266 (True)
• Huggingface_hub version: 0.14.1
• Transformers version: 4.30.0.dev0
• Accelerate version: 0.19.0
• xFormers version: not installed
• Using GPU in script?: yes
• Using distributed or parallel set-up in script?: yes

Who can help?

@patrickvonplaten

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[LWprogramming]

did this end up getting changed, wontfix, or something else?

[yiyixuxu]

hi @LWprogramming
we did fix it here #6006
sorry I forgot to link it to this issue