FisherKing-WAN2.2-[GGUF-14B]-T2V-ReferenceWorkflow-v1.0 [Low VRAM Compatible]
세부 정보
파일 다운로드 (1)
모델 설명
Workflow Goal
Provide a clean, educational reference implementation for WAN 2.2 Text-to-Video generation.
This workflow focuses on simplicity, reproducibility, and education rather than including every available feature. It provides a tested baseline for generating high-quality cinematic videos directly from text prompts while remaining easy to understand, modify, and extend.
Treat this workflow as a starting point and customize the prompt, LoRA stack, and generation settings for your preferred artistic direction.
Version 1.0
Purpose
✓ Educational
✓ Reference Workflow
✓ Easy to Understand
✓ Easy to Extend
✓ Low VRAM Friendly
Workflow Pipeline
Text Prompt
↓
Prompt Engineering
↓
High Noise Sampling
↓
Low Noise Sampling
↓
VAE Decode
↓
(Optional) RIFE Frame Interpolation
↓
Final Video Output
Verified Settings
The following settings were used to validate this workflow and are recommended as the baseline configuration.
Sampling
✓ CFG : 1.0
✓ Steps : 6 (3 High Noise + 3 Low Noise)
✓ Shift : 10
✓ Sampler : Euler
✓ Scheduler : Simple
Video
✓ Frames : 81
✓ Output FPS : 16 FPS
✓ Final FPS (RIFE Enabled) : 32 FPS
✓ Recommended Resolution : 640 × 360 (16:9)
Required LoRA Configuration
This workflow uses the same Lightx2V LoRA during both sampling stages.
LoRA
lightx2v_t2v_14b_cfg_step_distill_v2_lora_rank32_bf16
High Noise
Strength : 2.0
Low Noise
Strength : 1.0
These strengths were used to validate the workflow and are recommended as the baseline configuration.
Optional Post Processing
The workflow includes an optional RIFE Frame Interpolation stage.
When enabled:
Input : 16 FPS
Output : 32 FPS
Produces smoother motion while preserving the original video duration.
Disable this stage if you prefer faster processing or do not have the required RIFE model installed.
Hardware & Resolution Notes
Validated using:
✓ NVIDIA RTX 2080 (8 GB VRAM)
✓ 64 GB System RAM
✓ ComfyUI v0.27+
Although optimized for 8 GB VRAM, WAN 2.2 Text-to-Video remains computationally intensive.
System RAM is equally important for handling intermediate tensors and memory paging during video generation.
Expected behavior:
8 GB VRAM + 16 GB RAM
Possible out-of-memory errors
Slower generation
8 GB VRAM + 32 GB RAM
Better stability
Performance depends on available system memory
8 GB VRAM + 64 GB RAM (or more)
Recommended configuration
Matches the environment used to validate this workflow
Design Philosophy
This workflow intentionally avoids unnecessary complexity.
The objective is to provide a stable, reproducible reference implementation that users can understand, learn from, and extend for their own creative projects.
Features
✓ Clean reference implementation
✓ Prompt Engineering ready
✓ High / Low Noise sampling pipeline
✓ Modular LoRA configuration
✓ Optional RIFE interpolation (16 → 32 FPS)
✓ Chrono Save integration for CivitAI
✓ Low VRAM focused workflow design