chambersu
patreon
17 fan art model sketches, 3D AI test, workflow/Asset update
🕑 Added 2025-02-18 17:53:25 +0000 UTC20250218Post_Preview.zip
I don't know why unable to upload preview images on Patreon tonight,
so I posted a download link above,
(late at night now, I want to post them and take a rest soon ~
It is hard to showcase and explain them easily for me,
Welcome to ask directly if I miss anything, THX~
Discord Channel Here:
https://discord.gg/XAfS57Rrj6
Email chambersu1996@gmail.com
Or Patreon here~
I should focus on asset polishment (for both original design and fan art...) but I found 3D AI improved a lot so I changed my plan this month,
I tried to train real people's Flux LoRAs,
to catch unique face features from them, for anime character design,
( for disliking anime characters sharing similar faces),
including the data sets and training param here,
(Photos are downloaded from Instagram,
Ana (@theanastasiah) · Instagram 照片和视频
Kanna Oyama (@kanna0123) · Instagram 照片和视频
WINTER (@imwinter) · Instagram 照片和视频
My laptop cannot train Flux LoRAs locally,
so I use cloud services like CivitAI and LiblibAI, I can only share the params instead of output and upload the training preset files directly...~
Download:
20250218Post_FluxFaceLoRAs.zipTrellis 3D AI grid models test samples,
TRELLIS - a Hugging Face Space by JeffreyXiang
which is open-source, and can be used locally for simple modes.Tripo 3D AI characters/mech/monster model test samples,
Tripo AI - Create Your First 3D Model with Text and Image in Seconds
which is good at one-image-to-model sketching.Face builder samples, a really cool addon for converting portrait photos to head models,
FaceBuilder for Blender | KeenTools
which is good at modeling real faces and anime faces with reasonable 3D structure.
Download:
20250218Post_TrellisTripoFaceBuilder.zipMale Shadowmancer, Humanoid monster (non-exclusive models for a dark Victorian-style game)
MaleShadowmancers_v03_t03_Packed.blendI used Rodin 3D Ai for my character workflow finally,
Rodin - Free AI 3D Model Generator from Images | Hyper3D
which is good at converting multi-view images to model sketches,
Also updated CKPT/LoRA/workflow a little bit,
drag the pictures to ComfyUI to load the workflows and take a look~
'NovaWaiHassaku07' is a merged IL CKPT,
'NovaWaiHassaku07' = 0.5A+0.25B+0.25C,
or 0.25A+0.25B+0.5C, you can merge and test which one is more suitable in 10 minutes ~(IL CKPTs are updated so fast now...~
A
Nova Anime XL - IL v3.0 Happy New Year | Illustrious Checkpoint | Civitai
B
WAI-NSFW-illustrious-SDXL - v9.0 | Illustrious Checkpoint | Civitai
C
Hassaku XL (Illustrious) - v2 | Illustrious Checkpoint | Civitai
Download:
20250218Post_Rodin17FanArtWorkflowLoRA.zip
I found Auto Rig is great for rigs and animations quickly too, I will post rigged samples later~
I will upload the second post for the last batch's 5 polished high-poly shaded characters (Revna/Sophie/Elara/Misaka Mikoto), to make sure able to detail these sketches as high-poly shaded models, (clean-retopo mesh and tex projection/bake),
and this batch's 17 fan art character sketches above, I think they are famous so I didn't list them here, can Google by their names~
then March for male base mesh/fan art characters, procedural oil painting tex/shader, and some fx test, some tiny things I will post too if they work fine(won't detail if not...~
will detail auto rig, video mocap, and retarget even CFX in the future(we have tested for some projects already), also grids from environment to build levels,
maybe open some commission slot on Patreon when the fan art workflow is ready~
Feel free to contact me by leaving a comment or mailing to chambersu1996@gmail.com, if any questions or anything here interests you~
Also Discord Channel Here:
https://discord.gg/XAfS57Rrj6
will reply ASAP ~
