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.

Fix mistake in last commit. (#8496)

comfyanonymous8a4ff747bd91
diff --git a/comfy/ldm/flux/model.py b/comfy/ldm/flux/model.py
index 09dd2482c..846703d52 100644
--- a/comfy/ldm/flux/model.py
+++ b/comfy/ldm/flux/model.py
@@ -159,6 +159,9 @@ class Flux(nn.Module):
if add is not None:
img += add
+ if img.dtype == torch.float16:
+ img = torch.nan_to_num(img, nan=0.0, posinf=65504, neginf=-65504)
+
img = torch.cat((txt, img), 1)
for i, block in enumerate(self.single_blocks):
@@ -187,9 +190,6 @@ class Flux(nn.Module):
if add is not None:
img[:, txt.shape[1] :, ...] += add
- if img.dtype == torch.float16:
- img = torch.nan_to_num(img, nan=0.0, posinf=65504, neginf=-65504)
-
img = img[:, txt.shape[1] :, ...]
img = self.final_layer(img, vec) # (N, T, patch_size ** 2 * out_channels)