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.

ACE VAE works in fp16. (#8055)

comfyanonymous577de83ca9c9
diff --git a/comfy/sd.py b/comfy/sd.py
index ee350d5b5..e98a3aa87 100644
--- a/comfy/sd.py
+++ b/comfy/sd.py
@@ -451,7 +451,7 @@ class VAE:
self.latent_dim = 2
self.process_output = lambda audio: audio
self.process_input = lambda audio: audio
- self.working_dtypes = [torch.bfloat16, torch.float32]
+ self.working_dtypes = [torch.bfloat16, torch.float16, torch.float32]
self.disable_offload = True
self.extra_1d_channel = 16
else: