VACE 2.1 reference image to image (one frame)
5.0
0 reviews
Description
VACE performs well in single-frame image generation, with good results when using prompts to guide image editing. Outpainting also yields satisfactory outcomes. However, inpainting tests were unsuccessful, and using ControlNet to preprocess images for guided generation also failed.
https://huggingface.co/QuantStack/Wan2.1_14B_VACE-GGUF
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (3)
- latest (a month ago)
- v20250815-084320
- v20250812-083225
Node Details
Primitive Nodes (12)
Anything Everywhere (1)
CLIPLoaderGGUF (1)
Image Comparer (rgthree) (1)
ModelPatchTorchSettings (1)
ModelSamplingSD3 (1)
PathchSageAttentionKJ (1)
Power Lora Loader (rgthree) (1)
TrimVideoLatent (1)
UnetLoaderGGUF (1)
WanVaceToVideo (1)
WanVideoNAG (1)
easy anythingIndexSwitch (1)
Custom Nodes (13)
ComfyUI
- KSampler (1)
- VAEDecode (1)
- VAELoader (1)
- LoadImage (1)
- EmptyImage (1)
- CLIPTextEncode (2)
- easy imageRemBg (1)
- LayerMask: MaskGrow (1)
- LayerMask: SegmentAnythingUltra V2 (1)
- LayerUtility: ImageBlendAdvance (1)
- ImageAndMaskPreview (1)
- GetImageSize (1)
Model Details
Checkpoints (0)
LoRAs (0)