Skip to content

Allows DDPM Scheduler to have Zero Variance on Generation Runs #13141

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
rustadadam wants to merge 3 commits into
base: main
Choose a base branch
from
+6 −2
Open

Allows DDPM Scheduler to have Zero Variance on Generation Runs #13141

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
bc226f1
First update to adjust the diffuser to have no step noise
rustadadam Feb 10, 2026
ad7bcf8
updates
rustadadam Feb 10, 2026
4c2d419
Merge branch 'main' into main
rustadadam Feb 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 6 additions & 2 deletions src/diffusers/schedulers/scheduling_ddpm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,8 +152,9 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
The beta schedule, a mapping from a beta range to a sequence of betas for stepping the model.
trained_betas (`np.ndarray`, *optional*):
An array of betas to pass directly to the constructor without using `beta_start` and `beta_end`.
variance_type (`"fixed_small"`, `"fixed_small_log"`, `"fixed_large"`, `"fixed_large_log"`, `"learned"`, or `"learned_range"`, defaults to `"fixed_small"`):
Clip the variance when adding noise to the denoised sample.
variance_type (`"fixed_small"`, `"fixed_small_log"`, `"fixed_large"`, `"fixed_large_log"`, `"learned"`, `"learned_range"`, or `"zeros"`, defaults to `"fixed_small"`):
The type of variance to compute. If `None`, uses the variance type specified in the scheduler
configuration.
clip_sample (`bool`, defaults to `True`):
Clip the predicted sample for numerical stability.
clip_sample_range (`float`, defaults to `1.0`):
Expand Down Expand Up @@ -198,6 +199,7 @@ def __init__(
"fixed_large_log",
"learned",
"learned_range",
"zeros",
] = "fixed_small",
clip_sample: bool = True,
prediction_type: Literal["epsilon", "sample", "v_prediction"] = "epsilon",
Expand Down Expand Up @@ -553,6 +555,8 @@ def step(
elif self.variance_type == "learned_range":
variance = self._get_variance(t, predicted_variance=predicted_variance)
variance = torch.exp(0.5 * variance) * variance_noise
elif self.variance_type == "zeros":
variance = torch.zeros_like(pred_prev_sample)
else:
variance = (self._get_variance(t, predicted_variance=predicted_variance) ** 0.5) * variance_noise

Expand Down