Benji-AnimateDiff-Vid2Anime_FullVersion - Version 20240528

Replace with the new IPAdapter node, basically just re-add it again connect the same way is good to go. :), Because IPA did some update recent months of the custome node.

Benjamin Law

Hi Benji, I know this is an older workflow, the new version doesn't give me the same results (AnimateDiff_IPA_Mask_Realism-Style-3_5.json) as it doesn't use the background replacement image properly or animate it in my attempts. So when I load the old workflow I get an error on the ipadapter section: "IPAdapterAdvanced The size of tensor a (4) must match the size of tensor b (3) at non-singleton dimension 1" I did a check on ipadaptersAdvanced github, and searched forums for the fix. I have had success just deleting and replacing the erroring node in some workflows after the comfyui update, but no luck there, the error remains. I opened both of your workflows to see if I could sus it out and the build appears the same. Maybe it is one of the inputs from another section sending rgbA instead of rgb? I can't seem to locate the solution. It would be great to get this one working for me again. Any help is appreciated.

talon nightshade

The remote host error is normal , we have that lots of time, when processing finish or the GUI in Idel state. For video combine, please aware the Save video option eed to turn On.

Benjamin Law

this is great! I got it all running however, whne it gets to the last group node..(2nd sampleing) it porduce one static image. the Mp4 does save in the output folder in Comfy. not sure why the video combine does not play in the GUI window When lookign at the terminal i see this Prompt executed in 389.88 seconds Exception in callback _ProactorBasePipeTransport._call_connection_lost(None) handle: Traceback (most recent call last): File "asyncio\events.py", line 80, in _run File "asyncio\proactor_events.py", line 165, in _call_connection_lost ConnectionResetError: [WinError 10054] An existing connection was forcibly closed by the remote host any idea why this woudl happen? thanks for your help!

eduardo

Yea, the freebie stuff expired.

Benjamin Law

Hello, I may just be overlooking it but I can't find the link for Lite Version ComfyUI Workflow: https://www.patreon.com/posts/105078161, just brings me back to the full version.

Neuman Robinson

I found the reason and fixed it on your file attached "benji-AnimateDiff-Vid2Anime_FullVersion.json". It run complete and very fast. Thank you for your support!

Chuong Pham

Share full? I think you should pay attention to the video tutorial. I should say a series of Stable Diffusion Animation series of tutorial videos. I understand, most people are lazy to pay attention and spend some time to go through things. They just wanna download and run , i see this typically a zombie. 😉

Benjamin Law

Chuong Pham

Yup

Benjamin Law

I don't want to use "IPA_Character_Only", i can ignore it? and don't fill folder character and background, so i can submit generate "Queue Prompt" it?

Chuong Pham

Do the setting in ControlNet, set it as high as you can for Lineart if you have a close up shot of the face.

Benjamin Law

I'm trying to use this with a music video where someone is singing. The face/mouth tracking isn't looking so great. What can I do to improve the face tracking?

Roger Barlow

Do setting for CN strength, IPA Linear with Concat or Add.

Benjamin Law

@innovate Future, in IPA CHARACTER ONLY, its not taking any character outfit style. only taking background style.

Uday Paila

Please help and share my Youtube video. Thanks

Benjamin Law

A really clear and efficient workflow👍👍👍

Crane

Merry Christmas to me💃🏻

Crane