Wan2.1 VACE Inpainting
Editar regiones específicas en videos mientras se preserva el contenido circundante. Excelente para eliminación o reemplazo de objetos.
Acerca de
The Wan2.1 VACE Inpainting workflow is designed to seamlessly edit specific regions in video content while preserving the integrity of the surrounding areas. This is particularly useful for tasks such as object removal or replacement in video sequences. Utilizing the Wan2.1 model, this workflow leverages a combination of nodes including KSampler for sampling, CLIPTextEncode for text encoding, and VAEDecode for decoding video frames. The workflow is structured into several steps, starting from loading models and video files, to creating masks and compositing the final video output.
Technically, the workflow employs the WanVaceToVideo node to convert inpainted frames back into video format, while the TrimVideoLatent and CreateVideo nodes manage the video length and format. The ImageCompositeMasked node is crucial for blending the inpainted regions with the original video, ensuring a natural look.
By using core nodes, the workflow avoids the need for additional custom installations, making it accessible and easy to implement. This workflow is particularly valuable for content creators and video editors who need to make precise edits without affecting the overall quality of the video.
Lo que proporcionas
- Leyendo el flujo de trabajo
Se ejecuta en ComfyUI. Este flujo de trabajo no puede ejecutarse en esta página. Ábrelo en Comfy Cloud con todo cargado o descárgalo para tu propio ComfyUI.
Probar en CloudDentro del flujo de trabajo
Una vista previa de solo lectura del gráfico de nodos. Ábrelo en Comfy Cloud o descárgalo para editar los nodos.
Loading the graph
Open full-size workflow previewArchivos de este flujo de trabajo
- Autor
- ComfyUI
Requiere nodos personalizados
- Leyendo el flujo de trabajo
Usa este flujo de trabajo desde código
Se genera a partir del gráfico y las entradas del flujo de trabajo cuando es posible; de lo contrario, se muestran ejemplos almacenados. Revisa las credenciales y los endpoints antes de ejecutar el código, ya que puede generar cargos.
What you need
- Runs on
- Comfy Cloud
- Clave
- COMFY_API_KEY
Requiere un plan de pago de Cloud y créditos.
# Python 3.10+: pip install comfy-sdk==0.4.0
import os
from comfy_sdk import Comfy
api_key = os.environ.get("COMFY_API_KEY", "").strip()
if not api_key: raise RuntimeError("Set COMFY_API_KEY")
os.environ.setdefault("COMFY_BASE_URL", "https://cloud.comfy.org")
client = Comfy(api_key=api_key)
workflow = client.workflows.from_json({
"3": {
"class_type": "KSampler",
"inputs": {
"cfg": 1,
"denoise": 1,
"latent_image": [
"49",
2
],
"model": [
"48",
0
],
"negative": [
"49",
1
],
"positive": [
"49",
0
],
"sampler_name": "uni_pc",
"scheduler": "simple",
"seed": 584027519362099,
"steps": 4
},
"_meta": {
"title": "KSampler"
}
},
"6": {
"class_type": "CLIPTextEncode",
"inputs": {
"clip": [
"38",
0
],
"text": "A little boy and a dog are hiking in the forest. The puppy runs around the boy cheerfully.\n"
},
"_meta": {
"title": "CLIP Text Encode (Positive Prompt)"
}
},
"7": {
"class_type": "CLIPTextEncode",
"inputs": {
"clip": [
"38",
0
],
"text": "过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走,过曝,"
},
"_meta": {
"title": "CLIP Text Encode (Negative Prompt)"
}
},
"8": {
"class_type": "VAEDecode",
"inputs": {
"samples": [
"58",
0
],
"vae": [
"39",
0
]
},
"_meta": {
"title": "VAE Decode"
}
},
"38": {
"class_type": "CLIPLoader",
"inputs": {
"clip_name": "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
"device": "default",
"type": "wan"
},
"_meta": {
"title": "Load CLIP"
}
},
"39": {
"class_type": "VAELoader",
"inputs": {
"vae_name": "wan_2.1_vae.safetensors"
},
"_meta": {
"title": "Load VAE"
}
},
"48": {
"class_type": "ModelSamplingSD3",
"inputs": {
"model": [
"154",
0
],
"shift": 5
},
"_meta": {
"title": "ModelSamplingSD3"
}
},
"49": {
"class_type": "WanVaceToVideo",
"inputs": {
"batch_size": 1,
"control_masks": [
"130",
0
],
"control_video": [
"208",
0
],
"height": 720,
"length": 81,
"negative": [
"7",
0
],
"positive": [
"6",
0
],
"reference_image": [
"162",
0
],
"strength": 1,
"vae": [
"39",
0
],
"width": 720
},
"_meta": {
"title": "WanVaceToVideo"
}
},
"58": {
"class_type": "TrimVideoLatent",
"inputs": {
"samples": [
"3",
0
],
"trim_amount": [
"49",
3
]
},
"_meta": {
"title": "Trim Video Latent"
}
},
"68": {
"class_type": "CreateVideo",
"inputs": {
"bit_depth": "auto",
"codec": "none",
"color_space": "sRGB",
"fps": [
"210",
2
],
"images": [
"8",
0
]
},
"_meta": {
"title": "Create Video"
}
},
"69": {
"class_type": "SaveVideo",
"inputs": {
"filename_prefix": "video/ComfyUI",
"format": "auto",
"format.codec": "auto",
"video": [
"68",
0
]
},
"_meta": {
"title": "Save Video"
}
},
"111": {
"class_type": "MaskToImage",
"inputs": {
"mask": [
"215",
1
]
},
"_meta": {
"title": "Convert Mask to Image"
}
},
"129": {
"class_type": "RepeatImageBatch",
"inputs": {
"amount": [
"211",
2
],
"image": [
"111",
0
]
},
"_meta": {
"title": "Repeat Image Batch"
}
},
"130": {
"class_type": "ImageToMask",
"inputs": {
"channel": "red",
"image": [
"129",
0
]
},
"_meta": {
"title": "Convert Image to Mask"
}
},
"140": {
"class_type": "UNETLoader",
"inputs": {
"unet_name": "wan2.1_vace_14B_fp16.safetensors",
"weight_dtype": "fp8_e4m3fn_fast"
},
"_meta": {
"title": "Load Diffusion Model"
}
},
"154": {
"class_type": "LoraLoaderModelOnly",
"inputs": {
"lora_name": "Wan21_CausVid_14B_T2V_lora_rank32.safetensors",
"model": [
"140",
0
],
"strength_model": 0.30000000000000004
},
"_meta": {
"title": "Load LoRA"
}
},
"162": {
"class_type": "LoadImage",
"inputs": {
"image": "api_bfl_flux_1_kontext_multiple_images_input_dog.jpg"
},
"_meta": {
"title": "Load Image"
}
},
"208": {
"class_type": "ImageCompositeMasked",
"inputs": {
"destination": [
"213",
0
],
"mask": [
"215",
1
],
"resize_source": True,
"source": [
"216",
0
],
"x": 0,
"y": 0
},
"_meta": {
"title": "Image Composite Masked"
}
},
"209": {
"class_type": "LoadVideo",
"inputs": {
"file": "video_wan_vace_inpainting_input_video.mp4"
},
"_meta": {
"title": "Load Video"
}
},
"210": {
"class_type": "GetVideoComponents",
"inputs": {
"video": [
"209",
0
]
},
"_meta": {
"title": "Get Video Components"
}
},
"211": {
"class_type": "GetImageSize",
"inputs": {
"image": [
"210",
0
]
},
"_meta": {
"title": "Get Image Size"
}
},
"213": {
"class_type": "RebatchImages",
"inputs": {
"batch_size": [
"211",
2
],
"images": [
"210",
0
]
},
"_meta": {
"title": "Rebatch Images"
}
},
"214": {
"class_type": "PreviewImage",
"inputs": {
"images": [
"208",
0
]
},
"_meta": {
"title": "Preview Image"
}
},
"215": {
"class_type": "LoadImage",
"inputs": {
"image": "video_wan_vace_inpainting_1st_frame_and_mask.png"
},
"_meta": {
"title": "Load Image"
}
},
"216": {
"class_type": "MaskToImage",
"inputs": {
"mask": [
"219",
0
]
},
"_meta": {
"title": "Convert Mask to Image"
}
},
"219": {
"class_type": "InvertMask",
"inputs": {
"mask": [
"215",
1
]
},
"_meta": {
"title": "Invert Mask"
}
}
})
# run() also passes api_key to the partner nodes in this workflow.
job = client.run(workflow, api_key=api_key)Preguntas y respuestas
The workflow uses the Wan2.1 model, specifically the VACE-1.3B and VACE-14B models, for video inpainting tasks.
Yes, the VACE-1.3B model supports 480P resolution, while the VACE-14B model supports up to 720P resolution.
No, the workflow is designed using core nodes, so no additional custom nodes are required.
You can use the ImageToMask node to define the region you want to inpaint in the video.