Update: WAN 2.2 Smooth Workflow v5.0 - 04/19/2026
"WAN 2.2 Smooth Workflow v5.0" update: The new version now has a First2LastFrame Workflow. Make sure to fully update your Comfyui and the extensions.
At the time of making the Showcase: Stability Matrix v2.15.7 and ComfyUI v.019.3.
Lightx2v Loras used:
HIGH with weight 1.0 -> high_noise_lora_rank64_lightx2v_4step_1022.safetensors
LOW with weight 1.5 - > lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank128_bf16.safetensors
Still uses the same extensions as v4.0 - No need to install more. Make sure ComfyUI, its dependencies and the Extensions are fully updated.
The Showcase is focused on the usage of First2LastFrame since the other workflows are pretty much the same. Added some examples of Loops you can make. I might take a few tries but its possible to make something simple - keep Loops at 5 seconds at most - Wan 2.2 doesn't work very well above that threshold.
Updated Notes: Added mode hyperlinks for some of the Resources and Models Extensions used in the Workflow.
The videos on Showcase used the SmoothMix I2V v.2.
All Extensions for WAN 2.2 Smooth Workflow v5.0:
Links for MMAudio usage:
MMAudio NSFW Model (fine-tuned off the base model)
Put the MMAudio models on ComfyUI/models/mmaudio. Create the folder if it does not exist.
Update: WAN 2.2 Smooth Workflow v4.0 - 03/14/2026
"WAN 2.2 Smooth Workflow v4.0" update: A new version with a new Ksampler method and more functions. Make sure to fully update your Comfyui and the extensions.
NAG: Now uses Normalized Attention Guidance (NAG) in Ksamplers. Check the github page more info. It makes the negative prompt be effective even at CFG 1.0 for improved quality and control.
Adaptive Prompts: You can use Adaptive Prompts when writing your Positive prompt now. Check the github page more info. To check the choices made by the Node just open the Positive Prompt Subgraph.
Updated Notes: Added hyperlinks for all Extensions used in the Workflow. Other notes have been added and updated. There are alot of little notes everywhere. Make sure to read them. ^^
The videos on Showcase used the SmoothMix I2V v.2 and SmoothMix T2v V.3.
All Extensions for WAN 2.2 Smooth Workflow v4.0:
Update: WAN 2.2 Smooth Workflow v3.0 - 19/01/26
"WAN 2.2 Smooth Workflow v3.0" update: Its a big workflow now. Make sure to fully update your Comfyui and the extensions.
All in one: Txt2vid, Img2vid and Audio Options! You can now try adding audio to your videos.
Select your workflow: I tried to make it very simple to change between the Img2vid, Txt2vid and Audio Workflows.
New fps Output: I had to change the video outputs to 25fps to make sure the audio and video are in sync.
Multiple Notes: There are alot of little notes everywhere. Make sure to read them. ^^
Added Extensions:
It uses MMAudio. So install - ComfyUI-MMAudio Nightly
And its requirements:
MMAudio NSFW Model (fine-tuned off the base model)
Puts the MMAudio models on ComfyUI/models/mmaudio. Create the folder if it does not exist.
Small Update - 10/16
Small change made to "WAN 2.2 S. Workflow v2.0" and "WAN 2.2 Txt2Video Workflow v2.0" That should allow for a faster VAE Decoding.
Update: WAN 2.2 Txt2Video Workflow v2.0 - 10/08
"WAN 2.2 Txt2Video Workflow v2.0" update: now its SUPER simple and easy to use.
I made it to use with the Smooth Mix Txt2Vid Wan 2.2 Checkpoint! Make sure to try it! ^^
Here are my recommend video setting when using the workflow with SmoothMix Txt2Vid:
Low/Mid end PCs:
Set Steps to "6" on both KSamplers;
KSampler 1 -> end_at_step = "3";
KSampler 2 -> start_at_end = "3".
High end PCs:
Set Steps to "10" on both KSamplers;
KSampler 1 -> end_at_step = "5";
KSampler 2 -> start_at_end = "5".
Update: WAN 2.2 Smooth Workflow v2.0 - 09/29
"WAN 2.2 S. Workflow v2.0" update: now its SUPER simple and easy to use.
I made it to use with the Smooth Mix Wan 2.2 Checkpoint! Make sure to try it! ^^
All the videos on the showcase for WAN 2.2 S. Workflow v2.0 used Smooth Mix Wan 2.2 Checkpoint.
It does not has Florence anymore. Make sure to install and update all necessary extensions listed below. No new extensions where used.
Update: Adv. Workflow - 08/29
"Adv. WAN 2.2 S. Workflow" update: added the Node "Color Match" before 'Video Result' and 'Last Frame' to help with color consistency.
Update: Txt2Vid Workflow
Added a txt2vid workflow. Thanks for vinque for the suggestion. =)
Update: Adv. Workflow + First2Last Frame
Made a few adjustments to the basic workflow for these new version!
Both use the same tools/weights as the basic workflow so no need to install anything new if you can already use the basic one.
Both versions have "Power Loader Node" set for the High and Low Loras.
More Loras usually means more instructions put on the Positive Prompt so Florence2 is disabled by default.
To make a looping animation on First2Last Workflow make sure to set enough time for the AI to process the loop. Try 5 or 6 seconds.
WAN 2.2 SMOOTH WORKFLOW
PLEASE make sure to read everything.
Making a simple Workflow from Wan is tricky. Wan requires are more complicated set up to work. I'm going into this assuming you have read all tutorials necessary to make Wan 2.2 work on your PC as doing that here is unresonable.
WAB 2.2 is still pretty new and there is still alot to test. All images used on the Showcase use the same set of loras, weights and configurations. The only difference was the image used and the Positive Prompt. I tried using realistic, semirealistic and anime style to show that its possible to animated any style you want by using the configuration on this workflow.
So, before we begin, this Worflow uses the following:
Use the one that fits your own specs. The one used for the showcase was the "wan2.1-i2v-14b-720p-Q8_0" version.
Out of all Loras that increases generation speed this one has given me more consistenty results.
The one used for the showcase was the "Wan21_T2V_14B_lightx2v_cfg_step_distill_lora_rank32" version.
Of course you should test other versions or Loras that speed up generation as well as well.
Showout to @CubeyAI for the Loras. Without these making NSFW results was a pain, and we can't have that can we? Make sure to use the latest version of those Loras.
Used extensions:
==========================================================
Its divided in sections to make it easy to navigate.
PROMPTS
Here is where the magic happens. So lets start here.
There are a few ways to use it:
TYPE NOTHING ON THE POSITIVE
By default Florence will create a prompt for you and a lot of times... thats enough lol. You don't need to type anything on the Positive. For examples of results that use an empty positive prompt check this and this.
I recommend you start by leaving the Positive Prompt empty at first. After you see the results you should begin working on your prompt if you did not like the results.
USE SIMPLE PROMPTS
Don't put a giant wall of text on the Positive Prompt as Florence will do that already. The showcase has Videos with very few prompts, some with just ONE.
DISABLE FLORENCE
If you really can't seem to get what you want disabling Florence might be the way to go.
FLORENCE
Choose your Florence model. Default model is "Florence-2-base-PromptGenv.2.0". Feel free to test other version or disable all of Florence nodes if you think its not helping.
CHECKPOINTS GGUF / SAFETENSOR
Very straight foward. Just choose your Checkpoint: GGUF or Safetensors. Disable the ones you are not going to use. Set to GGUF by default.
SPEEDING TOOLS AND LORAS
Here is where Sage and the Loras are set up. Feel free to change values or disable what you don't what to use.
CLIP / VAE
Still using the same as the wan 2.1
Text encoder: umt5_xxl_fp8_e4m3fn_scaled
VAE: Wan 2.1
IMAGE
Load your Image! :)
VIDEO SIZE
By default its set to 480x720 with Upscale by 2. That results in a 960x1440 video! Default upscale method is "lanczos"!
Feel free to change anything it to fit your own specs.
VIDEO LENGTH
Here you can adjust the video length.
By default is set to 81 or 5 seconds.
VIDEO SETTINGS
All settings here were set after testing. Feel free to change to your liking! :)
VIDEO PREVIEW/RESULT/LAST FRAME
You should be able to see how your video is turning out on the preview. If you don't where the results are going just cancel the run!
The video result is set to be at 30 fps. The video will be saved on Video/(Date_video_was_made).
The last frame of the video will be saved on the same folder as the video. You can use this to extent your videos! Check this post for examples.
====================================
Other recommend extensions:
Description
FAQ
Comments (118)
V3.0 Update. Its an All in one Workflow now. Its big but please don't let it intimidate you. I tried to make sure its as easy to understand as possible. ^^
The showcase shows generations using all workflows in it with audio included. All videos have the workflow embedded on them.
The audio can be very imprecise (and sometimes annoying to get right), but it's fast-generating, so you can try multiple generations just for audio in quick succession.
Have fun!
Dude, did you have problems installing the required nodes? My comfyui cannot install Rifeinterpolation and crfloat to integer.
@rochamiguel663929 No I didn't. I actually forgot to post them as requirements. Dammit. T_T Make sure you are installing the correct extentions: ComfyUI_Comfyroll_CustomNodes and ComfyUI-VFI
EDIT: nevermind i found it in the comfyui-wanvideowrapper extension
i have everything else but the only thing in red is the Normalize Audio Loudness node. what extension is that in?
Hi, I'm new to ComfyUI and I'm having an issue with the MMAudio part of this workflow. Even though I want only nature sounds like a waterfall and birds, the model often adds human voices or speaking sounds into the video. I've already added 'speech, talking, human voice' to the negative prompts, but it doesn't seem to work. Could you please explain why this happens or how I can fix it? Thank you!
Try downloading and switching to the original MMaudio model. The NSFW one is heavily trained on moaning, so voices can't be prompted out.
Need to bring back Florence. I loved that feature made it easier and better when adding to the prompt.
is the RIFEInterpolation node necessary? If it is, it's impossible to clone its repository
What's your problem? I had similar issues and solve it installing using the GIT UI from windows since the "clone git..." comand did not work for me.
@rochamiguel663929 I'm guessing its a problem with the "your-username" part of git clone https://github.com/your-username/ComfyUI-VFI.git, its makes me log into github so i made an account, but no matter what the process would just fail
@captainohyeah1 yeap...Had same problem, here's what I did to solve it. Install GIT for windows, then use GIT UI and there are 2 fields, 1# field for repository adress for example https://github.com/GACLove/ComfyUI-VFI.git and 2# field for destination folder and here's where you put .....\ComfyUI\custom_nodes\ComfyUI-VFI, and that's all. Doing that the GIT UI clone the repository automatically creating the folder and places all the things there. DO the same procedure for other nodes with same problem of GIT log.
@rochamiguel663929 thanks that worked, i just needed some username except mine i guess
@rochamiguel663929 i got same problems```:( BUT , I downloaded zip to /cumstom nodes/ and run "pip install -r requirements.txt"...it works..
Simple, intuitive, efficient, and beautiful.
Help, why can't ComfyUI detect the ComfyUI-MMAudio Nightly node that I manually installed?
WAN 2.2 Smooth Workflow v3.0 работает супер ! Спасибо огромное!!!
Wow great all in one first time i play with sound great
You will find it funny tho that I have an issue my outputs are blurry XD
I tried wan_2.1_vae, wan_2.1_vae_fp8, wan_2.1_vae_fp16 and wan_2.1_vae_fp32,
I tried without lora
I tried rife49 interpolation
all the same
It is time for me to sleep but I will find the solution at some point and i ll get back to you so if someone have the same pbm
great work again !
Hello! I added this to my prompt and it helped a lot with the blur:
"Faces remain sharp and stable with no motion blur, female’s stance steady and grounded, motion settles into a clear final pose, final frame crisp and stable." I used that last part since I wanted to continue the animation based on the last frame.
Wan 2.2 responds to direct commands like that. If you're running into a specific issue ChatGPT can actually provide you with some pointers to counteract the. :)
@Vexy_Kink Thanks it is not like a blurry effect it is more like noise i ll post an example
@Vexy_Kink Ok forget it i found the issue... Forgot the lightx2v lora...
I definitely need more sleep
The workflow is amazingly great my brain needs to go to sleep ;-)
@Vexy_Kink I ll keep your advice to get a steady last frame tho thanks !
@andreeee6327 great that you found the issue. :)
And yes; it will save you a lot of headache if you use first frame animations.
Thank you very much for your amazing workflow !
Hope you will release a SVI pro version ! <3
Getting Funky Eyes please help :(
Im using the WAN 2.2 Smooth Workflow v2.0 I2V with this Lightning Lora lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16 High 3.0 Low 1.5 currently with 10 and 5 steps with a res of 560x940. The problem i have is that Eyes and sometimes faces get distorted with and without additional Loras the further away the character ist the worse it gets. If anyone has an idea how to fix it i would really appreaciate it.
Before i used 6 and 3 steps and 480x720 with the same Lightning Lora and it was even worse.
so why exactly are you using a clip vision? To my knowledge wan doesnt need that and it just adds render time without any benefit from what i read on reddit. What am i missing here?
What happened if we disable it?
@Mekalonika i tested it, nothing changes, the render time was slightly better but only 10 secs. But zero difference in quality so i dont know
@Phac123 Thanks! good to know this.
Hi, is there an option to prevent the initial image from being saved? I couldn’t find it anywhere.
CLIP VISION encoder ? i dont get it , and also my output is blurry and sun
for me its also blurry, did you get it to work?
EDIT: I think i fixed it by reconnecting the load diffusion model nodes
Well I cant add loras...the button wont show up :(
v3.0 I2V is not working for me, is working as T2V
HELP:
it always tells me that "RIFEInterpolation" is still missing even though i have everything installed, and when i try to fix it and install it via cmd and git clone the message plops up shortly when i start comfyUI then dissapears even though the node is still missing.
My Custom Nodes Folder:
comfy-mtb
ComfyUI_Comfyroll_CustomNodes
comfyui-easy-use
comfyui-frame-interpolation
ComfyUI-GGUF
comfyui-kjnodes
comfyui-mmaudio
comfyui-mxtoolkit
comfyui-videohelpersuite
ComfyUI-WanVideoWrapper
rgthree-comfy
wanblockswap
Yeah, same problem. Probably that super helpful Comfy Update but who knows.
@The_Utulek_Complex yeah, that update destroyed quite a few node packs for me. litterly destroyed 2 of my main workflows, needed to fix them over the last 3 days -.-
i fixed it,to install ComfyUI-VFI
@frost_e It worked, u saved my life in a few seconds when I was trying everything out for a WHOLE day, TYSM!
model: GGUF. Everything else in the workflow left unchanged.
Error: KSamplerAdvanced
'dict' object has no attribute 'get_model_object'
I'm also getting this. Did you find a solution?
@earlcochrane I get ModelSamplingSD3 'NoneType' object has no attribute 'clone' error, if I bypass both of these nodes I get KSamplerAdvanced 'NoneType' object has no attribute 'get_model_object' too, seems like this "set for gguf" workflow isn't that great at working with gguf models.
@smithysmittysim364
Getting the same error, even with the smoothMixWAN model, which is ironically GGUF
@earlcochrane I found the solution myself, you have to manually connect the links without any "Get/Set Node"
Is showing as a PTH file not safe. Isn't there a safe version?
What folders do the mmaudio models go into? I'm not finding any good instructions on it...
"Puts the MMAudio models on ComfyUI/models/mmaudio. Create the folder if it does not exist."
When running Smooth Workflow v3.0 AIO and using smoothMixWan2214BI2V_i2vV20High/Low model I get a weird blurring result that is useless. I use the I2V section default. This does not happen if I swap SmoothMix V2 model for the older version. What can cause this?
Have the same issues just blurry shaky images
To solve the blurry problem use the light2xv lora indicated by the author on instructions.
@rochamiguel663929 So are you saying to load f"Wan21_T2V_14B_lightx2v_cfg_step_distill_lora_rank32" into the Power Lora Loader(rgthree) Load it in both high and low? since there is no high/low vesions of that particular lora. The instructions do not mention any of this for the V3.0. Can you give more details?
@stickyburrito547 Heya, if you use "DasiwaWAN22I2V14BLightspeed_synthseductionHigh/LowV9.safetensors" the results are fantastic. I just wanted to make it work using the defaults. But as I mentioned the results are awful with WF default models. This model is top notch in all NSFW WF's. Probably the best I ever seen, follows prompt perfectly.
@essence25 Please tell me again, what? I have the same question, put that in both low/high or what?
@KoujiAI Jsut use the SynthSeduction v9 set(comes in high and low) instead of the smoothMixWan22.
@essence25 Hey, i managed to undrestand how this sh works in the meantime, thanks for reaching out :)
Blurry output with image to video - please help!
Common problem. To solve the blurry problem you have to use the light2xv loras indicated by the author. In fact, the links are in the description.
Had the same problem - you need to load lightx2v loras in the lora nodes
Having an issue when dragging the workflow JSON file into comfyui, I get an endless progress wheel. Anyone else deal with this issue?
Yes, you need to unzip the file first, place the .json file somewhere, and drag that. It bugs out if the file is dragged from the zip itself.
- Hope it helps
@Spuuko553 You're the goat. Thank you
I'm getting a totally black output. Cannot figure it out. I've had no issues with WAN2.1 but first time trying WAN2.2. Help?
the same as me...
What can I do to prevent faces from becoming incredibly sloppy within a second or two?
I am getting an error in ComfyUI only when generating audio with MM-Audio. The interface starts normally and image and video workflows work fine, but exactly when the audio generation begins, the following PyTorch error appears:
free_upper_bound + pytorch_used_bytes[device] <= device_total – INTERNAL ASSERT FAILED at CUDAMallocAsyncAllocator.cppBased on the behavior, this does not look like a normal out-of-VRAM problem (there is no “CUDA out of memory” message). It appears to be an issue with PyTorch’s CUDA asynchronous memory allocator, which is known to fail in scenarios where large models perform many large memory allocations during inference, such as MM-Audio.
The error only happens when the audio generation step starts. The same ComfyUI setup and workflows continue to work correctly for image and video generation.
As a workaround, starting ComfyUI with the disable-cuda-malloc option prevents the crash, but it makes audio generation noticeably slower (6 minutes to video 5 seconds ). This suggests that the problem is not caused by the workflow itself, but by the way PyTorch manages GPU memory during MM-Audio inference.
What do you mean workflow is available in showcase vids? When i download it and drag n drop it, it doesnt find the workflow.
Faces dont seem to be retained at all, not sure why
Are you using any Lora's? Some Lora's are trained on heavy facial datasets, which can cause facial warping. I've used all of the workflow here, and I've not encountered said issue, so it'll be a setup issue on your end. Read through the description again to confirm you're using the correct models and so on.
I need help, my results look terrible. I'm just doing simple cumshot generations and I get white sludge and grainy/noisy quality. Im using the exact lightning lora and Q6K smooth v2. Ive tried with and without the nsfw encoder and nsfw clip vision, still looks bad eitehr way. The only significant change I made to this workflow is I got rid of the resolution picker group and replaced it with a node that lets me just say how many megapixels I want (I set it to ~0.5mp) so that way it scales my picture without having to worry about aspect ratio. That megapixel node works with Wan 2.1 FusionX just fine.
Can someone explain to me like I'm the absolute moron I am how to get this working? I've never used Comfy, and I feel like trying to jump into this is like trying to solve calculus when I haven't learned arithmetic. I downloaded the high and low models. I grabbed a VAE, but it might be the wrong one. I grabbed a text encoder, I think, but it's not seeing it. It says it's missing nodes, I install them, restart, still missing the same nodes.
Does anyone have a step by step, model by model (preferably with links) somewhere?
It was daunting for me at first but now it's fairly easy. Just make sure the req'd downloads correspond with the correct folder because usually that's why it states you are missing nodes. Make sure you have comfyui manager to install missing nodes. Just look over all of the node boxes in comfyui and you'll get used to it. Generally, each of them are all part of a working whole--just that each component has its own configurations. I would also suggest just taking the time to watch a YT video or read Reddit. If a dummy like me can get it, I'm sure you can.
Also use some sort of AI for any errors. They generally provide an accurate way to find success for issues.
@mistermista1 Thanks! I ended up getting up and running on a different workflow that just focuses on img2vid. The stuff it was saying it installed, but then was still missing I downloaded manually and added. I think there's a large pack of custom nodes that recently depreciated. The other work flow also automatically downloaded the VAE and text encoder it wanted which was nice.
This particular workflow is still missing two nodes. Something for GGUF (But I have the safetensor models so I think I can ignore those warnings) and something for... oh, not looking at it, but I think it was REInterpolation or something like that. So this workflow was still crashing. I'll probably come back to it eventually but happily for the moment I'm up and running.
@DarkfireAI Comrade by the Bread how did you managed? I'm stuck in this GGFU loader, a Rife interpolation and a CR float to integer
@MrBread The workflow we're commenting on isn't working for me for the same reasons. If you're using safetensor models I -think- you can disable the GGUF bits to get past those. When it kicks up the error go to the NodesMap on the left, search for the node number it gives you, and shut the eye icon. I assume this disables it, but honestly I have no idea what I'm doing.
Disabling those it still fails, which is, I assume the Rife bit and float integer. Not sure what to do with those bits.
Instead I'm using this workflow which only does img2img, but that cuts the fat and it has built in extensions to do longer clips. https://civitai.com/models/2337814/smooth-svi-pro-workflow
same model and initial parameters, why my video is blur?my gpu is 5090
same here! someone pls answer, all settings are default!
Running with a 2060 here and had a smooth, crisp video. I may have an answer for you though as I was able to recreate a blurry video. I've had a similar problem with their v2.0 workflow that would also give me random OOM even though I would only change the image to another of the same size. Are you clicking on that big, blank, green box that says, "VIDEO Width x Height?" If so, that is your problem. There is an invisible graph there that changes the resolution. Right click it, go to "properties" and set the resolution from there.
Tip: If you want to avoid accidentally clicking it in the future, just collapse it as you can still right click to change the properties.
@NuroticStalker hi that wasn't it, for me it's very much visible. question - are both Ksamplers supposed to have it on 6? Also is shift necessary? i dont really know what is i (started comfy few days ago).
@WItchesAI I had to up my step count when I wasnt using a lightning model or i'd get blurry vids or vids that blurry at the end, I noticed almost no blurrying at like 20-30 steps without lightning lora
@johhnnymann1 Thank you! After adding in the lightning loras, all the blurriness is gone! https://huggingface.co/lightx2v/Wan2.2-Distill-Loras/tree/main
You can't delete the text "fianal prtomt preview" from https://civitai.com/models/1847730?modelVersionId=2091039. Is this done on purpose to make us suffer?
Thanks! Checking it out rn. It seems the comfyui moded to transformers > 5 and huggingface > 0.36 and it broke the mmaudio dependency. I currently use this fix: TypeError: BigVGAN._from_pretrained() missing 2 required keyword-only arguments: 'proxies' and 'resume_download' · Issue #63 · kijai/ComfyUI-MMAudio
The resulting videos seem good.
yes, i try to raise this issue, the author mark it fixed. but i dont know how to fix it. Do i just downlooad the nodes.py and replace it for now?
@damoclesong ahh, maybe it's a lewd fix fixing the back compatibility. They left tat bug for a reason I suppose. Sadly, each time I update comfy, I'd have to apply that fix manually (redacting the file, not just replacing it). Btw there were numerous issues with exactly this error.
@shigatsu_renaki what lines are u redacting in file?
Tryin to fix that too, but always have errors
@aiAnaista I replaced the nodes.py file in the Qwen nodes folder. Don't forget to backup the old version.
I ran the latest workflow without adjusting any parameters and it took me an hour to generate the video, which is too strange. I used 4080
Its surprisingly neat and cool your latest workflow. Really apreciate it dude.
BTW any tips for the prompts in the audio box? Do I need to put the same one on the image? Or just describe generally the type of sound? Can you give some pointers?
GJ again keep it up ;)
I can't seem to make it work, it's very hard for me to get the RIFEInterpolation node functionning, it's always marked with a red cross despite many attempts at getting it from git repos etc after failing to install it with the custom node manager, if someone would be so kind as to help me I would rly appreciate it
Apperently it comes from here instead
https://github.com/GACLove/ComfyUI-VFI
@Spyrite165 fatal: repository 'https://github.com/your-username/ComfyUI-VFI.git/' not found when i try to install it via cd etc.
@OkBar1458824 it's git clone https://github.com/GACLove/ComfyUI-VFI
@bargez got it. Thanks
Doesn't work for me. I have to login into gihub, confirm, the I get the error.
remote: Repository not found.
I also can't install a node called "CR Float To Integer". No idea where to get it. :(
@QuietSpark the node should be from https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes
@bargez Thanks for the quick reply. That solved the "CR Float To Integer" issue.
RIFEInterpolation is still missing, though.
@QuietSpark idk about that github error, tried with another pc and the git clone worked fine, download the whole repository as .zip i guess
@bargez didn't know, this was an option. :) Worked like a charm. Thanks!
What node is saving the input image? trying to keep my video folder only videos and I cant trace where that one is happening.
MMAudioFeatureUtilsLoader
An error happened while trying to locate the files on the Hub and we cannot find the appropriate snapshot folder for the specified revision on the local disk. Please check your internet connection and try again.
step #5 in PNG Multi PNG Extend says: "Enable the toggle at the bottom of this section". There is no toggle. If I activate the "Final Step, activate at the end of creating the video", it throws "'NoneType' object has no attribute 'shape'" from Save Image/Save Latent inside the Multi PNG extend section. What am I doing wrong? All the steps up up to there work as expected.
Sadly unusable until BigVGAN._from_pretrained() missing 2 required keyword-only arguments: 'proxies' and 'resume_download' is fixed.
me too
This is caused by mm-audio. Try to update mm-audio first. For me this didn't pull the most recent version from github. I had to download this file:
https://github.com/kijai/ComfyUI-MMAudio/blob/main/mmaudio/ext/bigvgan_v2/bigvgan.py
and replace the existing in "ComfyUI\custom_nodes\comfyui-mmaudio\mmaudio\ext\bigvgan_v2"
Hmmm I think I found a glitch maybe?
On the resolution, I noticed that the only value that it accepts is the width.
Example: 608X800 or 608x912 or 608x1000 it outputs the same fixed resolution of 608x768.
Only when I increase the width does the height changes giving me another output res...
Is it just me? Does this happen because of my original pic res?
I am installing mmadio custom node though manager, but the workflow keeps telling me i need to install it even though i have it in folder. I am using 3.0v AIO
EDIT: I did it, just run the "python_embeded\python.exe -m pip install -r ComfyUI\custom_nodes\ComfyUI-MMAudio\requirements.txt" in my comfy ui root folder (folder with .bat files you use for launch and python_embedded folder), this command is for windows portable tho, i found it using a google search and it is listed on github page so if u have linux/smth else look there
I had the same thing with the node comfryroll. And now i have another problem with the missing RIFEinterpolation node from comfy-vfi, wich can't be downloaded anymore from git...
@OkBar1458824 I have the same problem with RIFEinterpolation it is suposed to be in ComfyUI-Frame-Interpolation package but doesn't seem to work
I'm having an issue with the video, it looks way to weird and off like movements are to awkward. idk what's going wrong but its the same with the v2 workflow as well there the movements are to fast looking.
I'm using WAN2.2 S. Workflow v2.0 and everything is running normally, but the generated video is always completely black. I don't know why.
Disable SageAttention, it works for me
Check the VAE bro.
can i run it on AMZ rx 9070? i have portable comfyui with rocm torch 7.2
Hello, I'm stuck on ModelSamplingSD3 with the error code 'NoneType' object has no attribute 'clone'. How can I fix it?
I cant get this to work, i try changing as little as possible, with no loras and still i get a "water-effect" like video and nothing else with a blurry version of the original picture. Anyone has any input on this?