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.

Make pyav loading code handle tRNS PNG. (#13607)

comfyanonymousc7a517c2f9d1
diff --git a/comfy_api/latest/_input_impl/video_types.py b/comfy_api/latest/_input_impl/video_types.py
index b2daa3d7d..6ed41bba8 100644
--- a/comfy_api/latest/_input_impl/video_types.py
+++ b/comfy_api/latest/_input_impl/video_types.py
@@ -284,7 +284,7 @@ class VideoFromFile(VideoInput):
if not checked_alpha:
for comp in frame.format.components:
- if comp.is_alpha:
+ if comp.is_alpha or frame.format.name == "pal8":
alphas = []
image_format = 'gbrapf32le'
break