@Chriqro Wow, thanks for the L tips . great module. thanks .

@juliusmartin sliders blank is an issue with the mixlabs and comfyui scripts nodes. If you have them, reinstall them then update all of comfyui using the manager

@Chriqro Thanks for the workflow - I found the Lora manager but can't see how to load them into the workflow? Sorry still pretty new to this

@Chriqro  Great, thanks didn't see that at first

@Chriqro Awesome! I'll try this.

HELLO BRO

YOU HAVE TELEGRAM, PLEASE SHOW ME HOW TO DO THIS

how to download this workflow

the link is not working

Is this workflow loop video longer than 5 sec or not ?

@zalfonxs09 There is an option to increase the number of seconds but WAN 2.2 itself is limited to 5 seconds (you can render for longer but it starts to have a mind on it's own and won't do what you prompted) you will need a workflow the chains together 5s loops, this one is not from that type.

workflow please

@LoraFlux I am using this workflow this is commented under: Lazy Wan 2.2 I2V workflow

I tried 1440x1800 or 1440x1920 and results were just the subject in focus will blur and almost fade away at times. especially when they are moving, any parts that are moving becomes super blurry and almost fades. for setup ive installed all the nodes, loaded ur workflow, selected lora, set the values to the previous comment and ran it

@Chriqro i'm having the exact same problem, and yes those loras are there in the process section. strength 1.00

Got the same problem. Did not change antything, just added the lightx2v 4step loras. Every moving object just gets grainy and blurry. Any settings I have to adjust?

@Chriqro It is. High on high, low on low. No Loras in the Lora loader, just a clean run. And still the longer it gets, the more blurred it is.

@Chriqro In the beginning of the video the movement is clear, but the image gets more and more blurry the longer the video is. I can still make out the movement, but it's extremely blurry.

@deservingdeath379 Check what Chriqro is suggesting and then check your resolution %, make sure that both the length and width are reduced enough to be under or at 896X1024

@Chriqro My quality was at 25 on a 768 x 768 image, lol. Problem solved! Great workflow, huge thumbs up.

I have the same issue whether I set it to 100% quality or anything else. The results are extremely fuzzy and unusuable. I've tried with 1280x720, 960x960, and other sizes. I've tried 25/50/75/100% on them. No matter what I pick this workflow is, quite literally, not usable outputs. This is on an RTX 4090. Also, the color match node... something is wrong with it. It takes like 70% of the workflow's duration and uses up over 50 GB of RAM and a majority of my 32 thread overclocked Ryzen 3950X, both specs most people do not usually have... and it makes these specs plead for mercy. It uses essentially next to zero GPU / VRAM for color match node. I just don't see how that node is usable for most people unless they're willing to wait literal hours for it to process.

@agent_agent794 for the 896x1024 you cited -- is that a good rule for all outputs on this workflow? I'm trying to understand where you got that from

@yoyomama172909 I think i got this resolution as max res number example either from another workflow or a lora. The few times i tested this workflow with higher res than that one i got fuzy results.

@Chriqro Oh yeah, I'll have to check that later. Thinking about it I don't recall doing that. I think the lora node you used for lora loading threw me and I didn't think about adding the light loras to that loader.

@Chriqro That fixed it. I had forgotten to add the light lora. I did notice another issue though, but I'm not entirely sure if its related to ComfyUI causing a recent bug with memory handling or not, but for some reason I can create a bunch of videos at a given identical resolution and then suddenly 4-5 later it crashes out of memory. While I'm not seeing the same issue with even higher resolutions on other workflows I don't know if it is specifically yours causing the issue or something else. Just mentioning in case others report the same issue.

@Chriqro Thank you I have a better understanding of the workflow now you basically click on the top left circle to expand settings. I have achieved some clean and great results but also where there is way too much blur still trying to figure out what is causing it. Maybe Wan 2.1 Lora's do not work with Wan 2.2. Or certain resolutions do and others not for example 768x768 image gave great results but when the only change I made was to 1024x1024 it would be very blurry but that also can be if it was a bad seed I guess. Thank you for the workflow once you get the hang of it it is very organized and a great asset :)

@Chriqro 100% I scaled the images manually, I tested 512x512, 768x768 1024x1024.

@Chriqro Thank you did you manage to replicate it?

@Chriqro ok thank you.

@Chriqro Hi, i have no idea how to LOAD loras, i start to tex <lora:here i put lora name(i go over file and go rename to copy name) and paste:1> and its give me erorr (((

@Chriqro You have to press Enter to copy it. Mouse click does not work.

@ac2023 This tip works. Mouse over it and then press enter. Extremely bizarre way to do it that most people would not guess. Creator should honestly add either a note to their page or a note above it in the workflow.

@ac2023 adding further to this I just noticed this isn't the intended behavior but a bug of that node. Once you pick the first lora with 'Enter' after mousing over it... the rest can merely be left clicked like you would expect.

@Chriqro omg thanks, im blind lollll

I second this motion. This is a very well organized workflow for the standard 5 seconds, now it needs a faceswap or other method to maintain the face consistency for extended video clips.

@Chriqro Thank you! I'll test it today!

Puppetry works good with just basic Wan I2v. Just doing basic prompts for short term movements, then edit them. I'm not paying a site that does not even try to test the top models. I'm amazed it's till a gamble. Just paying a person in buzz would probably do it? IMO But once they do that, I will pay for a year.

This isn't a lora