DDPMScheduler claims to have set_alpha_to_one in combination with steps_offset

[bghira]

When inspecting the configuration for SD/SDXL we noticed that set_alpha_to_one is mentioned in the scheduler's init docstring but it is nowhere to be found. It seems that DDPM either lacks this functionality, or does not need it.

[yiyixuxu]

Hi @bghira

It wasn't implemented for DDPM so I think it's a mistake in docstring.

Do you think we need to add this?

[bghira]

i'm not sure that it needs it, there is a further discussion on this issue now in #6068

[sayakpaul]

#6931

[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.

[sayakpaul]

#7128

[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.

[jbbqqf]

Hi! Triaging older issues this week — I think this one has shipped and can be closed.

• PR #7128 ("Change step_offset scheduler docstrings", merged 2024-03-14) reworked the affected scheduler docstrings.
• On current main, scheduling_ddpm.py L170-L180 only documents steps_offset; the misleading set_alpha_to_one mention is gone.
• The __init__ signature at scheduling_ddpm.py L209 confirms DDPMScheduler does not expose set_alpha_to_one, matching the corrected docs.

Would you mind closing this out? If a different scheduler's docstring still has the same problem, point me at it and I'll dig further.

(Disclosure: I drafted this comment with help from Claude Code while triaging stale issues; the references above were verified manually against the current main branch.)

[bghira]

@jbbqqf you behave like a spam bot to be honest, please stop