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.

Pass device to CLIP model.

comfyanonymousc3e96e637dde
diff --git a/comfy/sd.py b/comfy/sd.py
index fc3551fea..5001d497c 100644
--- a/comfy/sd.py
+++ b/comfy/sd.py
@@ -526,12 +526,13 @@ class CLIP:
def __init__(self, target=None, embedding_directory=None, no_init=False):
if no_init:
return
- params = target.params
+ params = target.params.copy()
clip = target.clip
tokenizer = target.tokenizer
load_device = model_management.text_encoder_device()
offload_device = model_management.text_encoder_offload_device()
+ params['device'] = load_device
self.cond_stage_model = clip(**(params))
#TODO: make sure this doesn't have a quality loss before enabling.
# if model_management.should_use_fp16(load_device):