June 10, Update: stmfnet
Just to save folks a headache the file is available for download as a model now.
June 06, Update: V4.0
Just a minor improvement, using a double interpolation of SFTMVFI results in higher fidelity of dynamic or fast motion scenes. Video is manually loaded to reduce time when generating high quantities of test runs. Now you can select which videos you want to upscale, will copy the name of the original input video.
Included the context framing node from Kijai's wrapper. Usually leave this off will exponentially increase generation times if on; however, can let you generate much longer videos beyond what your vram will allow you.
Interpolation requires higher VRAM if scaling if greater than 1.0. Changed the scaling factor to float for more control.
Updated nodes if there is any issues feel free to let me know.
https://drive.google.com/file/d/1s5JJdt5X69AO2E2uuaes17aPwlWIQagG/view
Place it in the folder: Custom_nodes\ComfyUI-Frame-Interpolation\ckpts\stmfnet
April 14, Update: V3.3
Updated Nodes.
March 27, Update: V3.2
Small changes to the Torch Compile and Tea Cache for more dynamic outputs using my Lora's.
Workflow is suitable for a 24GB card using the 14B models and a 12GB card using the 1.3B model.
Updates:
***Error within the naming of outputs corrected.
Uploaded a very simple and straightforward Wan I2V workflow. The workflow includes an upscaler with frame interpolation to bring your videos to a cool and smooth 60FPS. No BS just a simple addition. Original WanI2V is based on the workflow example of the amazing work from Kijai linked here: https://github.com/kijai/ComfyUI-WanVideoWrapper/tree/main/example_workflows.
Additionally, I've left the annotations within.
Prev Updates:
Compatible with 12GB memory cards. Tested using a 3060 RTX 12GB using this workflow can comfortably go up to 85 frames. 97 Frames is pushing the absolute limits will take about an hour to generate.
Due to demand this workflow is a small improvement to the Super Simple HV2V for those who don't do postprocessing outside of Comfyui.
Designed for use with Hunyuan Video and trained Lora models, it can upscale videos to true 60 FPS. It combines detailed notes on denoising strength, step optimization, aspect ratio adjustments, frame interpolation, and upscaling to create high-quality, smooth videos.
Order matters:
Generate → Upscale → Interpolate. Upscaling first preserves detail for interpolation.
Adjust parameters based on your Lora’s behavior and hardware limits.
Description
Updated Interpolation to use STFMNET
FAQ
Comments (17)
Hey there - was checking out your workflow and it threw an error related to SFTMVFI. Basically, it couldn't download the model from the repos.. tried 3 different ones I believe, but no luck. You know a fix for that by chance?
I'll add it in the download along with instructions on where to put it in your comfyui directory shortly. Thanks I think I ran into this same issue and just downloaded it manually.
Sweet! Thank you
@Bleedy https://drive.google.com/file/d/1s5JJdt5X69AO2E2uuaes17aPwlWIQagG/view
Place it in the folder. You might need to create the new folder. Custom_nodes\ComfyUI-Frame-Interpolation\ckpts\stmfnet
@SingularUnity perfect! thanks for sharing the wf and the extra file. Got it working no problem now!
@Bleedy No worries, let me know if you have any other questions and keep me posted on how your gens turn out. Always interested in seeing what folks do with the stuff I put together.
@SingularUnity I normally just use Kijai's wanwrapper, but was looking for an upscaler option and that's when your wf caught my attention. It is very similar to the wanwrapper, so definitely intuitive and easy to use. Thanks again!
@Bleedy Thank you for the kind words. That's the goal, one of my pet peaves is people overcomplicating WF or changing the colors and layouts that ultimately make it difficult to understand. Really you don't need an all in one workflow to make a polished final product. I have about a dozen workflows I sift through depending on my needs, one for promos, one for image gens, one for videos, etc. Most of them are only a dozen or so nodes. Simply.
On a side note what you're thoughts on my posting local training articles? do you suspect there is an appetite for it here or are folks mostly doing their training on site now?
@SingularUnity I honestly don't really have a good idea on that subject. I have often wanted to get into training myself, but I am always cranking out new content instead. Most of my own stuff is custom but based on a merge of a merge of a merge of a merge, lol... but haven't really focused time on LoRa training yet.
CAN SOMEONE TELL WHAT`S THE DIFFERENCE BETTWEEN THIS AND VACE?
Imagine, if you will, VACE as a flamboyant video‑to‑video top‑hat‑wearing gentleman who first dunks your source clip into a bubbling fondue of “guided image‑to‑video” sauce, twirls his monocle, and then pirouettes back out as a freshly bedazzled motion picture.
Now this workflow is like a humble, shoeless plain i2v. It simply takes your lonely stills, whispers “animate, you silly jpegs,” and like magic the moving pictures happen, hold the extra garnish. No dapper, no monocle, just straight up i2v.
VACE = video 2 video (+ fancy i2v seasoning).
This workflow = image 2 video (toast‑dry, zero sprinkles).
@SingularUnity thanks for so vivid explanation!
But would VACE model quality be better as output, compared to plain WAN 2.1 14B I2V model?
@tinderizza in terms of quality the difference is not huge, the main issue with VACE is that it changes the character features drastically, it treats your starting image as a reference (create something like this image) instead of animating the image as it is
The google drive link is not working "Sorry, unable to open the file at this time.
Please check the address and try again."
I just put it on this page as an update.
@SingularUnity u tha goat