Software Engineering Courses (SECourses)•7mo ago

https://www.youtube.com/watch?v=vKU1iYDwkEI&ab_channel=uisato

Measuræ v1.2 / Audioreactive Generative Geometries - [TouchDesigner]

Updated system for audioreactively generative geometries, intervened with various AI techniques.

More experiments, project files, and tutorials, through: https://www.uisato.art/

#touchdesigner #aiart #audioreactive

___
0:00 - Stained Glass → Occultist Drawing → Map
0:13 - Experimental Notation → Pencil Drawing → Old Map
0:26 - Glass ...

Uuisato https://www.youtube.com/watch?v=vKU1iYDwkEI&ab_channel=uisato

Furkan Gözükara SECourses•5/29/25, 3:48 PM

cool

Martins•5/30/25, 11:00 AM

I just did fine-tuning and am now getting 'AssertionError: You do not have CLIP state dict!' Does anybody know what could be the cause?

MMartins I just did fine-tuning and am now getting 'AssertionError: You do not have CLIP ...

Furkan Gözükara SECourses•5/30/25, 12:13 PM

do you have clip downloaded accurately?

Furkan Gözükara SECourses•5/30/25, 12:13 PM

fine tuned model is only unet

Martins•5/30/25, 11:39 PM

i am not sure. I think i will start over and see how it goes.

Original message was deleted

Furkan Gözükara SECourses•5/31/25, 12:42 PM

did you use our onetrainer config?

Furkan Gözükara SECourses•5/31/25, 12:42 PM

you can use kohya extract lora feature

Furkan Gözükara SECourses•5/31/25, 12:42 PM

i recommend dreambooth and extract lora

Furkan Gözükara SECourses•5/31/25, 12:42 PM

so dreambooth also bad quality?

Furkan Gözükara SECourses•5/31/25, 12:42 PM

ok that is good

Furkan Gözükara SECourses•5/31/25, 12:42 PM

so all you need to do is extract lora and use on same base model

Original message was deleted

Furkan Gözükara SECourses•5/31/25, 12:43 PM

of course you can

Furkan Gözükara SECourses•5/31/25, 12:43 PM

onetrainer generated model file same as kohya

Furkan Gözükara SECourses•5/31/25, 12:43 PM

you have your safetensors right?

Furkan Gözükara SECourses•5/31/25, 12:43 PM

like 6 7 gb

Original message was deleted

Furkan Gözükara SECourses•5/31/25, 12:44 PM

yep

Furkan Gözükara SECourses•5/31/25, 12:44 PM

because onetrainer has EMA

Furkan Gözükara SECourses•5/31/25, 12:45 PM

it improves

AiInfluence•5/31/25, 12:45 PM

so doctor

AiInfluence•5/31/25, 12:45 PM

is it better to always tirain on dreambnooth

AiInfluence•5/31/25, 12:45 PM

if i want lora?

AiInfluence•5/31/25, 12:45 PM

insted of lora tab

AiInfluence•5/31/25, 12:45 PM

is it bnetter quality

AiInfluence•5/31/25, 12:46 PM

do the annoying lines go away?

AAiInfluence is it better to always tirain on dreambnooth

Furkan Gözükara SECourses•5/31/25, 12:46 PM

for sdxl yes

Furkan Gözükara SECourses•5/31/25, 12:46 PM

for flux i think native lora training better

Furkan Gözükara SECourses•5/31/25, 12:46 PM

extraction not working great

Furkan Gözükara SECourses•5/31/25, 12:46 PM

probably extraction logic is broken not accurate

Furkan Gözükara SECourses•5/31/25, 12:46 PM

for flux

FFurkan Gözükara SECourses probably extraction logic is broken not accurate

AiInfluence•5/31/25, 1:21 PM

then why you say you reccomend it

Original message was deleted

AiInfluence•5/31/25, 1:21 PM

yeah only problem is the line artifacts that come up

AiInfluence•5/31/25, 1:22 PM

swarm

AiInfluence•5/31/25, 1:22 PM

i put 2k res

AiInfluence•5/31/25, 1:22 PM

the lines only appear when using a lora

AiInfluence•5/31/25, 1:25 PM

yeah

AiInfluence•5/31/25, 1:25 PM

i found supir

AiInfluence•5/31/25, 1:25 PM

so it gets the job done

AiInfluence•5/31/25, 1:25 PM

but its incconvienti you know

AiInfluence•5/31/25, 1:25 PM

db fixes those artifacts

AiInfluence•5/31/25, 1:25 PM

just amazing

AiInfluence•5/31/25, 1:26 PM

i never tried to upscale in swarm

AiInfluence•5/31/25, 1:26 PM

can you guide me?

AiInfluence•5/31/25, 1:26 PM

sure

AAiInfluence then why you say you reccomend it

Furkan Gözükara SECourses•5/31/25, 1:51 PM

i dont recommend for flux anymore

Furkan Gözükara SECourses•5/31/25, 1:51 PM

after i did some more testing and usage

Furkan Gözükara SECourses•5/31/25, 1:52 PM

but it has 1 advantage

Furkan Gözükara SECourses•5/31/25, 1:52 PM

if you upscale to 2x, extract lora works better than lora since lora produces lines

Elli_Oaken•6/2/25, 12:28 PM

Hi guys! I've been training FLUX LoRAs using Kohya on Windows with 4 RTX GPUs. About a month ago, everything worked fine, but now I keep getting crashes related to torch.distributed.get_world_size() when using Accelerate.

Has something changed recently in PyTorch or Accelerate that breaks multi-GPU training on Windows?
Is there any recommended workaround for this?

EElli_Oaken Hi guys! I've been training FLUX LoRAs using Kohya on Windows with 4 RTX GPUs. A...

Furkan Gözükara SECourses•6/2/25, 2:52 PM

yes you need like

Furkan Gözükara SECourses•6/2/25, 2:52 PM

SXM or NVLINK machines

https://www.youtube.com/watch?v=vKU1iYDwkEI&ab_channel=uisato

Similar Threads

Similar Threads

Similar Threads