Skip to content
WorkflowListedExternal
Open source

Wan 2.2 GGUF - T2V I2V and FLF2V workflow for VRAM or RAM limited people — v2 more GGUFed

A collection of 3 workflows forked from 3 other sources and reworked to allow them to do more with less VRAM or RAM. Prevents using the RAM while staying on the VRAM, prevents ComfyUI crashes (you will less likely...

Sourcecivitai.com/models/1893033
TaskMultiCapabilitiesImage to video · Text to videoFamilyWanEstimated VRAMNot yet characterizedBgrade

Runtime profile

Access
External workflow
Status
Listed
Source
Civitai
Updated
Oct 2, 2026

Source description

A collection of 3 workflows forked from 3 other sources and reworked to allow them to do more with less VRAM or RAM. Prevents using the RAM while staying on the VRAM, prevents ComfyUI crashes (you will less likely saturate the RAM), allows you to use bigger models on the same graphic card, with better settings.

The cost? The WAN 2.2 models will have to be loaded each time into memory, which is time consuming, but it is better than nothing. Loading the model from an SSD drive is recommended. Please check your extra_model_paths.yaml file in the main directory of ComfyUI.

I made several typos in the workflow's instructions. It's tedious to update the entire workflow file here (I'd have to submit the Nth " example.png turns into a mighty Phoenix"), so I am posting the corrected version below. Apologies for the inconvenience.

  • Preparation:

Make sure you have all the needed files by enabling everything. If you are not using an SSD to store your files, you're likely wasting a lot of time.

If you run Part 1 and Part 2 separately, in this order, several times, the models will be loaded into memory over and over. It reduces the overall amount of RAM or VRAM needed, but the process is time-consuming. This method is useful when the alternative is a complete workflow failure, such as a crash near the end. I spend an additional 1 minute and 20 seconds per generation with this method, but I also save time by keeping everything in VRAM. For Wan 2.2 FLF2V, this is competitive; for Wan 2.2 T2V, it is much less so.

  • First Part:

Select "Enable Part 1" and disable "Enable Part 2." Apply the usual settings. Run the workflow. It will only process the High CFG part. Once the first part of the workflow is finished, the .latent file will be saved to a cache directory. You will probably need to adapt the save path for this file.

  • Second Part:

Select "Enable Part 2" and disable "Enable Part 1." Make sure the path to the latent file is correct. Run the workflow.

Original workflows:

t2v:

https://civitai.com/models/1835262/wan22-t2v-gguf-lightx2v-lora

I2v:

https://civitai.com/models/1822764/wan-22-i2v-gguf-compact-speed-wf-or-lightning-lora-44-steps

(natively very good)

FLF2V:

https://civitai.com/models/1884402/flf2v-wan-22-video-generation-first-frame-to-last-frame-with-frame-interpolation

(this ís my own model, it is now very far from the original that is linked)

What you need

Hardware and download-size estimates are not available yet.

Requirements

Models

4

  • umt5-xxl-encoder-Q6_K.gguf

    Not resolved

    Text encoder · Unknown

  • wan_2.1_vae.safetensors

    File unverified

    VAE · Unknown

  • Wan2.2-T2V-A14B-HighNoise-Q6_K.gguf

    Not resolved

    Checkpoint · Unknown

  • Wan2.2-T2V-A14B-LowNoise-Q6_K.gguf

    Not resolved

    Checkpoint · Unknown

Node packs

5

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

  • This collection of three Wan 2.2 GGUF workflows is for generating video from text, an image, or first and last frames on systems with limited VRAM or RAM; it produces video and can save a .latent file during a split run.

Flow

  • The collection covers text-to-video, image-to-video, and first-frame-to-last-frame generation.

  • The workflows are reworked to use less VRAM or RAM and allow bigger models on the same graphics card.

  • In a split run, Part 1 processes only the High CFG part and saves a .latent file to a cache directory for Part 2.

Inputs and outputs

  • The generation inputs can be text, an image, or first and last frames; the generated output is video.

  • A split run outputs a .latent file in a cache directory for the next part.

Setup

  • Before running, enable all needed files, check extra_model_paths.yaml in ComfyUI's main directory, and consider storing the model files on an SSD.

    Suggestion · not verified

  • No model-use license is listed.

Dependencies

  • The listed ComfyUI add-ons are ComfyUI-VideoHelperSuite 1.6.1, ComfyUI-KJNodes 1.1.4, and rgthree-comfy 1.0.2507112302.

  • Before setup, verify the package names for comfyui-gguf and was-node-suite-comfyui.

    Suggestion · not verified

  • Have these model files ready: umt5-xxl-encoder-Q6_K.gguf, wan_2.1_vae.safetensors, Wan2.2-T2V-A14B-HighNoise-Q6_K.gguf, and Wan2.2-T2V-A14B-LowNoise-Q6_K.gguf.

    Suggestion · not verified

Configuration

  • To reduce overall RAM or VRAM use, run Part 1 and Part 2 separately, in that order.

    Suggestion · not verified

  • For a split run, enable Part 1 and disable Part 2; then verify or adapt the .latent path, enable Part 2, disable Part 1, and run again.

    Suggestion · not verified

Cautions

  • The WAN 2.2 models load into memory each time, making split runs time-consuming; the workflow instructions also contain typos, so use the corrected instructions when configuring a run.

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiPrimary
civitai.com/models/1893033

Verified Jul 19

Source excerpts

1 excerpt
NoteJul 19civitai.com/models/1893033

Source context: 829 downloads · Type Workflows · Base model Wan Video 2.2 I2V-A14B