Loading...
> >

This is DAZ3D

SasquatchIsCoolSasquatchIsCool Posts: 262
edited July 28 in The Commons

This is when DAZ3d meets AI.  Been practicing a lot with creating a system to connect the power of both.  I think I found it.   Anyone want training on this?  In fact can you find the items I used from DAZ3d? 

Post edited by SasquatchIsCool on

Comments

  • shootybearshootybear Posts: 147

    I'm interested. :)

  • nonesuch00nonesuch00 Posts: 18,722

    I think a set of DAZ 3D renders, ComfyUI, Flux, Flux Kontext, controlnets, with some of those workflows at CivitAI can work really well together.

  • FSMCDesignsFSMCDesigns Posts: 12,843
    edited July 28

    Nice. Too bad the hair waas screwed up around her ears. What animation app with you use. I have been animating DS renders with WAN 2.1 and framepack, sooooooo much faster than using keyframes and DS.

    Just did this in a couple of minutes with a DS render using WAN 2.1

    Post edited by FSMCDesigns on
  • SnowSultanSnowSultan Posts: 3,773

    FSMC, are you using a local Comfy install to run WAN 2.1? I'm curious to know if you can run a 3D animation through an animation AI model with a Lora and make it 2D, but almost every Lora I see for animation is p0rn. 

  • TheMysteryIsThePointTheMysteryIsThePoint Posts: 3,220

    SasquatchIsCool said:

    This is when DAZ3d meets AI.  Been practicing a lot with creating a system to connect the power of both.  I think I found it.   Anyone want training on this?  In fact can you find the items I used from DAZ3d? 

    I'm interested as well.

  • WendyLuvsCatzWendyLuvsCatz Posts: 40,046
    edited July 28

    TheMysteryIsThePoint said:

    SasquatchIsCool said:

    This is when DAZ3d meets AI.  Been practicing a lot with creating a system to connect the power of both.  I think I found it.   Anyone want training on this?  In fact can you find the items I used from DAZ3d? 

    I'm interested as well.

    the OP always starts new threads not revealing much

    you can use Wan2.1 to do this

    I use Pinokio but since you are a coder, conda and the GitHub repository would likely be more your thing 

    https://github.com/Wan-Video/Wan2.1

     

    I don't want to hijack threads or share non DAZ related stuff but this is an example of what one can do

    I used a LoRa trained on my Brother in Fooocus for the images and the Cakeify LoRa in Wan for this

     

    Post edited by WendyLuvsCatz on
  • FSMCDesignsFSMCDesigns Posts: 12,843

    SnowSultan said:

    FSMC, are you using a local Comfy install to run WAN 2.1? I'm curious to know if you can run a 3D animation through an animation AI model with a Lora and make it 2D, but almost every Lora I see for animation is p0rn. 

    I am using NIM online since my GPU can't handle the local install very well. For non nudity and longer vids I use Krea since they offer WAN 2.1 also. I don't think either has the option for any LoRAs.

  • WonderlandWonderland Posts: 7,133

    This is simple to do now in Midjourney 

  • SnowSultanSnowSultan Posts: 3,773

    I am using NIM online since my GPU can't handle the local install very well. For non nudity and longer vids I use Krea since they offer WAN 2.1 also.

    Thank you very much.

     

     This is simple to do now in Midjourney 

     Last time I checked, Midjourney will censor anything more risque than a woman in a t-shirt and shorts. 90% of AI videos wouldn't exisit if Midjourney was the only method to generate them.  ;)

  • TheMysteryIsThePointTheMysteryIsThePoint Posts: 3,220

    WendyLuvsCatz said:

    the OP always starts new threads not revealing much

    you can use Wan2.1 to do this

    I use Pinokio but since you are a coder, conda and the GitHub repository would likely be more your thing 

    https://github.com/Wan-Video/Wan2.1

     

    I don't want to hijack threads or share non DAZ related stuff but this is an example of what one can do

    I used a LoRa trained on my Brother in Fooocus for the images and the Cakeify LoRa in Wan for this

     

    Thanks for the insight, Wendy! I'll check it out when I get home.

  • windli3356windli3356 Posts: 290

    I spent quiet a bit effort and money to experiement on it and didn't like it, for one, nearly all AI models I used will change the face of my character (not for better) and doing all kind unwanted thing, picture to video was fine, but too much censorship on 100% legal and none NSFW stuff. 

  • SasquatchIsCoolSasquatchIsCool Posts: 262

    windli3356 said:

    I spent quiet a bit effort and money to experiement on it and didn't like it, for one, nearly all AI models I used will change the face of my character (not for better) and doing all kind unwanted thing, picture to video was fine, but too much censorship on 100% legal and none NSFW stuff. 

    If you want to do like simple video and keep things consistent  I'd say go to Haliou off of wavespeed.  It allows a little more saucier stuff.

  • NetherFalconNetherFalcon Posts: 885
    edited August 8

    Veo 3 can also do some amazing things.  A secondary option would be Hailou.

    https://hailuoai.video/share/ai-video/p31EywrDyJ4J?source-scene=shared&source-media=shared_link

    Here's a video from Hailuo, animated from one of my renders.  They're good if you aren't concerned about audio.  Quick generation and follow the prompts decently enough, especially for the price and usage of the tokens. 

    Post edited by NetherFalcon on
Sign In or Register to comment.