Wan 2.2 14B Fun Inp
Generate videos from start and end frames using Wan 2.2 Fun Inp.
概要
This ComfyUI workflow, titled 'Wan 2.2 14B Fun Inp', is designed to generate videos by interpolating between start and end frames using the Wan 2.2 model. The workflow leverages a series of nodes including CLIPLoader, VAELoader, and ModelSamplingSD3 to load and process the input images, while the WanFunInpaintToVideo node plays a crucial role in creating the video by applying inpainting techniques.
The workflow is particularly useful for creating seamless transitions in videos, making it ideal for artistic video projects or animations where smooth morphing between frames is desired. The use of advanced nodes like KSamplerAdvanced and VAEDecode ensures high-quality video output, while the inclusion of LoraLoaderModelOnly allows for flexible model adjustments.
入力するもの
- ワークフローを読み込み中
ComfyUI で実行できます。 このワークフローはこのページでは実行できません。すべてが読み込まれた状態でComfy Cloudで開くか、ご自身のComfyUI用にダウンロードしてください。
Cloudで試すワークフローの内部
ノードグラフの読み取り専用プレビューです。ノードを編集するには、Comfy Cloudで開くかダウンロードしてください。
Loading the graph
Open full-size workflow preview- 作成者
- ComfyUI
カスタムノードが必要です
- ワークフローを読み込み中
コードからこのワークフローを使用する
可能な場合はワークフローグラフと入力から生成されます。それ以外の場合は保存された例が表示されます。料金が発生する場合があるため、コードを実行する前に認証情報とエンドポイントを確認してください。
What you need
- Runs on
- Comfy Cloud
- キー
- COMFY_API_KEY
有料の Cloud プランとクレジットが必要です。
# 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({
"90": {
"class_type": "CLIPLoader",
"inputs": {
"clip_name": "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
"device": "default",
"type": "wan"
},
"_meta": {
"title": "Load CLIP"
}
},
"91": {
"class_type": "CLIPTextEncode",
"inputs": {
"clip": [
"90",
0
],
"text": "色调艳丽，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"
},
"_meta": {
"title": "CLIP Text Encode (Negative Prompt)"
}
},
"92": {
"class_type": "VAELoader",
"inputs": {
"vae_name": "wan_2.1_vae.safetensors"
},
"_meta": {
"title": "Load VAE"
}
},
"93": {
"class_type": "ModelSamplingSD3",
"inputs": {
"model": [
"116",
0
],
"shift": 8
},
"_meta": {
"title": "ModelSamplingSD3"
}
},
"94": {
"class_type": "ModelSamplingSD3",
"inputs": {
"model": [
"117",
0
],
"shift": 8
},
"_meta": {
"title": "ModelSamplingSD3"
}
},
"95": {
"class_type": "KSamplerAdvanced",
"inputs": {
"add_noise": "disable",
"cfg": 1,
"end_at_step": 4,
"latent_image": [
"96",
0
],
"model": [
"94",
0
],
"negative": [
"111",
1
],
"noise_seed": 0,
"positive": [
"111",
0
],
"return_with_leftover_noise": "disable",
"sampler_name": "euler",
"scheduler": "simple",
"start_at_step": 2,
"steps": 4
},
"_meta": {
"title": "KSampler (Advanced)"
}
},
"96": {
"class_type": "KSamplerAdvanced",
"inputs": {
"add_noise": "enable",
"cfg": 1,
"end_at_step": 2,
"latent_image": [
"111",
2
],
"model": [
"93",
0
],
"negative": [
"111",
1
],
"noise_seed": 221824495232956,
"positive": [
"111",
0
],
"return_with_leftover_noise": "enable",
"sampler_name": "euler",
"scheduler": "simple",
"start_at_step": 0,
"steps": 4
},
"_meta": {
"title": "KSampler (Advanced)"
}
},
"97": {
"class_type": "VAEDecode",
"inputs": {
"samples": [
"95",
0
],
"vae": [
"92",
0
]
},
"_meta": {
"title": "VAE Decode"
}
},
"99": {
"class_type": "CLIPTextEncode",
"inputs": {
"clip": [
"90",
0
],
"text": "A dreamy scene where a little cat is sleeping. Zoom in, and the cat opens its eyes, looks up, and blinks. In Q-style, with ice crystals."
},
"_meta": {
"title": "CLIP Text Encode (Positive Prompt)"
}
},
"100": {
"class_type": "CreateVideo",
"inputs": {
"bit_depth": "auto",
"codec": "none",
"color_space": "sRGB",
"fps": 16,
"images": [
"97",
0
]
},
"_meta": {
"title": "Create Video"
}
},
"101": {
"class_type": "UNETLoader",
"inputs": {
"unet_name": "wan2.2_fun_inpaint_high_noise_14B_fp8_scaled.safetensors",
"weight_dtype": "default"
},
"_meta": {
"title": "Load Diffusion Model"
}
},
"102": {
"class_type": "UNETLoader",
"inputs": {
"unet_name": "wan2.2_fun_inpaint_low_noise_14B_fp8_scaled.safetensors",
"weight_dtype": "default"
},
"_meta": {
"title": "Load Diffusion Model"
}
},
"110": {
"class_type": "LoadImage",
"inputs": {
"image": "video_wan2_2_14B_fun_inpaint_start_image.png"
},
"_meta": {
"title": "Load Image"
}
},
"111": {
"class_type": "WanFunInpaintToVideo",
"inputs": {
"batch_size": 1,
"end_image": [
"112",
0
],
"height": 640,
"length": 81,
"negative": [
"91",
0
],
"positive": [
"99",
0
],
"start_image": [
"110",
0
],
"vae": [
"92",
0
],
"width": 640
},
"_meta": {
"title": "WanFunInpaintToVideo"
}
},
"112": {
"class_type": "LoadImage",
"inputs": {
"image": "video_wan2_2_14B_fun_inpaint_end_image.png"
},
"_meta": {
"title": "Load Image"
}
},
"116": {
"class_type": "LoraLoaderModelOnly",
"inputs": {
"lora_name": "wan2.2_i2v_lightx2v_4steps_lora_v1_high_noise.safetensors",
"model": [
"101",
0
],
"strength_model": 1
},
"_meta": {
"title": "Load LoRA"
}
},
"117": {
"class_type": "LoraLoaderModelOnly",
"inputs": {
"lora_name": "wan2.2_i2v_lightx2v_4steps_lora_v1_low_noise.safetensors",
"model": [
"102",
0
],
"strength_model": 1
},
"_meta": {
"title": "Load LoRA"
}
},
"158": {
"class_type": "SaveVideo",
"inputs": {
"filename_prefix": "video/ComfyUI",
"format": "auto",
"format.codec": "auto",
"video": [
"100",
0
]
},
"_meta": {
"title": "Save Video"
}
}
})
# run() also passes api_key to the partner nodes in this workflow.
job = client.run(workflow, api_key=api_key)よくある質問
The workflow uses the Wan2.2 model along with nodes like CLIPLoader and VAELoader for image processing and video generation.
Yes, the workflow includes nodes for configuring video size and length to suit your VRAM capacity and project needs.
Ensure that your input frames are of high quality and use the KSamplerAdvanced and VAEDecode nodes for optimal video processing.
Yes, you can input various prompts in the CLIPTextEncode node to guide the video style and appearance according to your preferences.