Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

This isn't used anywhere.

comfyanonymous94e4fe39d868
diff --git a/comfy/ldm/models/diffusion/ddim.py b/comfy/ldm/models/diffusion/ddim.py
index 139c8e01e..befab0075 100644
--- a/comfy/ldm/models/diffusion/ddim.py
+++ b/comfy/ldm/models/diffusion/ddim.py
@@ -33,7 +33,6 @@ class DDIMSampler(object):
assert alphas_cumprod.shape[0] == self.ddpm_num_timesteps, 'alphas have to be defined for each timestep'
to_torch = lambda x: x.clone().detach().to(torch.float32).to(self.device)
- self.register_buffer('betas', to_torch(self.model.betas))
self.register_buffer('alphas_cumprod', to_torch(alphas_cumprod))
self.register_buffer('alphas_cumprod_prev', to_torch(self.model.alphas_cumprod_prev))
@@ -195,7 +194,7 @@ class DDIMSampler(object):
temperature=1., noise_dropout=0., score_corrector=None, corrector_kwargs=None,
unconditional_guidance_scale=1., unconditional_conditioning=None, dynamic_threshold=None,
ucg_schedule=None, denoise_function=None, extra_args=None, to_zero=True, end_step=None, disable_pbar=False):
- device = self.model.betas.device
+ device = self.model.alphas_cumprod.device
b = shape[0]
if x_T is None:
img = torch.randn(shape, device=device)