ComfyUI.org

Unlock the Power of Aging Timelapse Videos with Wan2.1 Model Workflow

CN
ComfyUI.org
2025-05-30 07:02:29

1. Workflow Overview

mbaggsu6icj5bi8k6k4425d6b86dcafaef1a040e615f7ffdfd45a0b289bd89416491ac0edd82746655.gif

This workflow leverages Wan2.1 Model to generate "Aging Timelapse" effect videos, featuring:

  • Image-to-Video Conversion: Transform input images into dynamic aging process videos

  • Time-Lapse Effects: Simulate wrinkles, hair whitening, and posture changes

  • Frame Interpolation: Boost smoothness via GIMM-VFI (16fps→32fps)

  • Multi-Stage Control: Combines T5 text encoding, CLIP vision encoding, and LoRA fine-tuning

2. Core Models

Model Name

Function

Path

Wan2_1-I2V-14B-480P_fp8_e4m3fn.safetensors

Main video generation model

ComfyUI/models/wan_video/

Aging Timelapse (Wan2.1 I2V LoRA)_v1.0

Aging effect adapter

ComfyUI/models/loras/

umt5-xxl-enc-fp8_e4m3fn.safetensors

Multilingual text encoder

ComfyUI/models/wan_text_encoder/

gimmvfi_r_arb_lpips_fp32.safetensors

Frame interpolation model

Auto-download to ComfyUI/models/gimm_vfi/

3. Key Components

Node Name

Function

Installation

WanVideoModelLoader

Loads Wan2.1 video model

Install ComfyUI-WanVideoWrapper

WanVideoTextEncode

Processes aging effect prompts

Same as above

GIMMVFI_interpolate

Frame interpolation (2x)

Install ComfyUI-GIMM-VFI

WanVideoTeaCache

VRAM optimization

Built-in with ComfyUI-WanVideoWrapper

VHS_VideoCombine

Video rendering & export

Install ComfyUI-VideoHelperSuite

4. Workflow Structure

Group 1: Image-to-Video (Aging Effects)

  • Inputs:

    • Source image (e.g., ComfyUI_05329_.png)

    • Aging prompt (e.g., "fine wrinkles, gradual hair whitening")

  • Process:

    1. Extract visual features via WanVideoImageClipEncode

    2. Encode text descriptions with umt5

    3. Enhance details using LoRA

  • Output: 512x768 latent video

Group 2: Interpolation & Export

  • Process:

    1. Upsample 16fps to 32fps with GIMMVFI

    2. Render MP4 via VHS_VideoCombine (H.264)

  • Params:

    • CRF=19 (high quality)

    • Frame rate: 32fps

5. Inputs & Outputs

  • Required Inputs:

    • 1024x1536 portrait image (PNG)

    • Aging description text (Chinese/English)

    • Seed value (default: random)

  • Final Output:

    • MP4 video (e.g., wanvideo_00007.mp4)

    • Saved to: ComfyUI/output/

6. Notes

  • ⚠️ VRAM Requirement: Minimum 16GB (24GB+ recommended)

  • 💡 Model Downloads:

    • Frame interpolation model (~4GB) auto-downloads on first run

    • Wan2.1 models require manual placement if missing

  • 🔧 Tuning Tips:

    • Adjust 0.26 in WanVideoTeaCache for speed/quality trade-off

    • Modify interpolation multiplier in GIMMVFI (default: 2x)

previous
Create Breathtaking Silhouette Art: A ComfyUI Workflow Tutorial
next
Mastering Video-to-Video Translation: A Deep Dive into Wan2.1 VACE Model and ComfyUI

Recommend