My stuff

  • My Workflows

  • Liked Workflows

  • Following Workflows

Go to OpenArt main site
Upload workflow

Wan-Animate Infinite Workflow

5.0

0 reviews
1
844
339
0
Description
  • Origin: This workflow is from this website; I made simple modifications because it couldn't generate long videos.
  • Automated "Fire-and-Forget" Mode: For projects with a simple background, you can enable a hands-free workflow. Set Lap to 1 and the run mode to instant. This allows the process to run unattended in the background; it is designed to automatically stop by producing an error once there are no more frames to process.
  • Background Consistency: The automated method can create poor transitions between video segments. For better consistency, use a high-quality frame from the end of the previous clip as the starting input for the next one.
  • Audio Sync: Audio will be out of sync due to framerate mismatches and must be fixed in post-production. To sync the audio within the workflow, you would need to either convert the input video's framerate or make complex modifications.
  • Regenerating a Clip: Output files are named in the format wanvideo_Lap_0000x.mp4. If you are unhappy with a specific segment (e.g., the 3rd clip), set the Lap value to 3 and change the video sampler's seed to regenerate just that part.
  • Initial Value: The Lap count should be set to 1 at the start of every new project.
  • Customization: This is a quick-test workflow. You should customize the resolution (width and height) and the base model to suit your needs.
  • Final Assembly: Once you have all the clips you're satisfied with, combine them and add the final audio in a video editing software.

1. 工作流来自这此网站,我发现不能生成长视频,所以简单修改一下

2. 如果背景简单,将左上角Lap改为1,将Run改为instant,就可以后台运行了,它最后会因为帧数不够报错停止运行

3. 上面的方法生成的视频背景衔接效果会比较差,对背景一致性要求高的可以截取前一个视频中的高质量图片作为输入图片

4. 由于输入视频帧的不匹配会导致声音的不匹配,声音需要后期重新配,想同步声音需要转换输入视频帧数,或者修改工作流(要改的地方比较多,这边不好描述)

5. 输出视频的命名规则是 wanvideo+Lap_0000x.mp4,如果对第Lap个视频不满意,修改对应的Lap值,然后修改video sampler的seed重新生成

6. 每个新项目视频的Lap初始值改为1

7. 这是快速测试工作流,自行修改宽高以及底膜

8. 最后对满意的片段在剪辑软件中按数字顺序排列配音合成

Discussion

(No comments yet)

Loading...

Author

3
2.5K
20
6.6K

No reviews yet

  • - latest (23 days ago)

Primitive Nodes (57)

BlockifyMask (1)

DownloadAndLoadSAM2Model (1)

DrawMaskOnImage (1)

FaceMaskFromPoseKeypoints (1)

GetNode (18)

ImageConcatMulti (1)

ImageCropByMaskAndResize (1)

ImageResizeKJv2 (1)

MarkdownNote (1)

Note (2)

PointsEditor (1)

Reroute (1)

Sam2Segmentation (1)

SetNode (13)

WanVideoAnimateEmbeds (1)

WanVideoBlockSwap (1)

WanVideoClipVisionEncode (1)

WanVideoContextOptions (1)

WanVideoDecode (1)

WanVideoLoraSelectMulti (1)

WanVideoModelLoader (1)

WanVideoSampler (1)

WanVideoSetBlockSwap (1)

WanVideoSetLoRAs (1)

WanVideoTextEncodeCached (1)

WanVideoTorchCompileSettings (1)

WanVideoVAELoader (1)

Custom Nodes (24)

Comfyroll Studio

  • - CR Integer To String (1)

  • - CR Seed (1)

ComfyUI

  • - GrowMask (1)

  • - CLIPVisionLoader (1)

  • - LoadImage (1)

  • - PreviewImage (1)

  • - easy showAnything (3)

  • - easy string (1)

  • - SimpleMath+ (2)

  • - DWPreprocessor (1)

  • - PixelPerfectResolution (1)

  • - VHS_VideoCombine (4)

  • - VHS_LoadVideo (1)

  • - GetImageSizeAndCount (1)

  • - INTConstant (3)

  • - JoinStrings (1)

Checkpoints (0)

LoRAs (2)

WanAnimate_relight_lora_fp16.safetensors

lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors