I think your comment filter is blocking my post - so here it is again without urls = Thank you, very interesting video. Quick tip - since you're on SDXL the control nets are less steady than 1.5, the adapter should be hotshotxl instead of animatediff v3. In addition, you should use the domain adapter at 0.5 strengh (simple lora only loader) as part of a stack that includes a fluid-friendly LORA . Just set the context window to 8 and overlap at 3 for hotshotxl.
Hi. I have been experimenting with various controlnets and 3d scenes, for various purposes, would love to learn more, get it more consistent between generations, where do get the links or more info?
Wow, this is very exciting! Am I right that you didn't give any materials to the chocolates in Houdini, it's done in ComfyUI? How consistent is this workflow, I mean if you have a longer scene how well does it retain the details? For example, if you have a clothing simulation and let's say it will be a leather jacket with stitches and wear, how well can the scene maintain that? Thanks in advance! By the way, hats off to you!
![[UNCUT] เก็บ "สจ.โต้ง" เอี่ยวมาเฟีย รู้ตัวคนบงการ ก่อนสั่งลั่นไก I คนดังนั่งเคลียร์ I 13 ธ.ค.67](http://i.ytimg.com/vi/o-J_p62D40w/mqdefault.jpg)
This is amazing, thank you for sharing your knowledge, Ardy!
cheers!!
Ardy thank you man! Your work give me hope in future of vfx!! You are great man! thank you bro
Thank you very much, glad to know I can help a tiny bit.
@a Ardy please check your json file, I can't load it, I have load all custom nodes, and now I have no errors, but also no nodes in workspace
sorry , all the best, everything works, nodes are very far in workspace, I have seen it))) thank you great work
lol, thanks ardy. bit better than slumming with houdini and redshift all those years ago :D
agree lol
Thanks mate, this is amazing.
for some reason i follow every step here and my result is just colorful noise...
download my JSON file and go from there
Following with extreme interest. Thank you
hi , great work realy keep it up , by the way i cant find the workflow file on your page , can u help?
NICE
wow thats great
I wonder how long will take untill we get an AI render engine for 3d applications
I think your comment filter is blocking my post - so here it is again without urls = Thank you, very interesting video. Quick tip - since you're on SDXL the control nets are less steady than 1.5, the adapter should be hotshotxl instead of animatediff v3. In addition, you should use the domain adapter at 0.5 strengh (simple lora only loader) as part of a stack that includes a fluid-friendly LORA . Just set the context window to 8 and overlap at 3 for hotshotxl.
Hi. I have been experimenting with various controlnets and 3d scenes, for various purposes, would love to learn more, get it more consistent between generations, where do get the links or more info?
I am not on SDXL,
Wow, this is very exciting! Am I right that you didn't give any materials to the chocolates in Houdini, it's done in ComfyUI? How consistent is this workflow, I mean if you have a longer scene how well does it retain the details? For example, if you have a clothing simulation and let's say it will be a leather jacket with stitches and wear, how well can the scene maintain that? Thanks in advance! By the way, hats off to you!
that might be a little difficult, depends on the complexity of your renders tho, worth testing it out!