not that it matter the sweet spot is still around epoch 200 and lora strength 0.95-1.0

X

xlads•9/16/24, 5:33 PM

I was driving down in my car today and couldn’t help but notice all the billboard ads and how ai is revolutionizing that entire market

KKoudkunstje (Roelof)not that it matter the sweet spot is still around epoch 200 and lora strength 0....

F

Furkan Gözükara SECourses•9/16/24, 5:39 PM

nice

Xxlads I was driving down in my car today and couldn’t help but notice all the billboar...

F

Furkan Gözükara SECourses•9/16/24, 5:39 PM

yep i also see so many ads are ai made

FFurkan Gözükara SECourses ok try gpu id 1

K

Kallamamran•9/16/24, 6:25 PM

Same problem

K

Kallamamran•9/16/24, 6:28 PM

K

Kallamamran•9/16/24, 6:28 PM

I don't get it

L

liorftw•9/16/24, 6:32 PM

Hey I used the supir from the videos tutorial and somehow it takes really long time for 50 steps with rtx 4090 any suggestion why?

FFurkan Gözükara SECourses ok then set accelerator

K

Kallamamran•9/16/24, 6:34 PM

So I ran the manual accelerate config and set GPU:1 there. Now it works

K

Kallamamran•9/16/24, 6:34 PM

Takes forever though

L

liorftw•9/16/24, 6:40 PM

im getting this error with supir:( some one know why is that?

Lliorftw im getting this error with supir:( some one know why is that?

D

dupopalot•9/16/24, 6:58 PM

Out of memory. How much VRAM and RAM do you have?

L

liorftw•9/16/24, 7:02 PM

oh ok i fixed it thanks it seems that too many things was using my vram

D

Dylan Normandin•9/16/24, 7:36 PM

I am trying to follow the massed compute FLUX LORA tutorial. I downloaded FLUX models, added them to stableswarmUI folders correctly. But now I get error which says "ERROR: self-start comfy-UI-0 on port 7823 failed. "

L

liorftw•9/16/24, 8:40 PM

what does it do? should i enable it?

L

liorftw•9/16/24, 8:40 PM

in the supir upscaler

KKallamamran So I ran the manual accelerate config and set GPU:1 there. Now it works

F

Furkan Gözükara SECourses•9/16/24, 9:21 PM

great

Lliorftw Hey I used the supir from the videos tutorial and somehow it takes really long t...

F

Furkan Gözükara SECourses•9/16/24, 9:21 PM

if not using shared vram

F

Furkan Gözükara SECourses•9/16/24, 9:22 PM

it is its speed

Lliorftw im getting this error with supir:( some one know why is that?

F

Furkan Gözükara SECourses•9/16/24, 9:22 PM

start as tiled

F

Furkan Gözükara SECourses•9/16/24, 9:22 PM

and fp8

Lliorftw what does it do? should i enable it?

F

Furkan Gözükara SECourses•9/16/24, 9:22 PM

not much worthy to play with

F

Furkan Gözükara SECourses•9/16/24, 9:23 PM

we already have 2 configs you can test them

Z

Zono50•9/16/24, 9:26 PM

@Dr. Furkan Gözükara using SUPIR and getting these errors, any suggestions? decoderF

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 128
    xFormers wasn't build with CUDA support
    attn_bias type is <class 'NoneType'>
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 128
    xFormers wasn't build with CUDA support
    attn_bias type is <class 'NoneType'>
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 128
    xFormers wasn't build with CUDA support
    attn_bias type is <class 'NoneType'>
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 128
    xFormers wasn't build with CUDA support
    attn_bias type is <class 'NoneType'>
    operator wasn't built - see

python -m xformers.info

 for more info

 for more info

flshattF@0.0.0

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 256
    xFormers wasn't build with CUDA support
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 256
    xFormers wasn't build with CUDA support
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 256
    xFormers wasn't build with CUDA support
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 256
    xFormers wasn't build with CUDA support
    operator wasn't built - see

python -m xformers.info

 for more info

 for more info

cutlassF

 is not supported because:
    xFormers wasn't build with CUDA support
    operator wasn't built - see

 is not supported because:
    xFormers wasn't build with CUDA support
    operator wasn't built - see

 is not supported because:
    xFormers wasn't build with CUDA support
    operator wasn't built - see

 is not supported because:
    xFormers wasn't build with CUDA support
    operator wasn't built - see

python -m xformers.info

 for more info

 for more info

smallkF

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 32
    xFormers wasn't build with CUDA support
    dtype=torch.bfloat16 (supported: {torch.float32})
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 32
    xFormers wasn't build with CUDA support
    dtype=torch.bfloat16 (supported: {torch.float32})
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 32
    xFormers wasn't build with CUDA support
    dtype=torch.bfloat16 (supported: {torch.float32})
    operator wasn't built - see

 is not supported because:
    max(query.shape[-1] != value.shape[-1]) > 32
    xFormers wasn't build with CUDA support
    dtype=torch.bfloat16 (supported: {torch.float32})
    operator wasn't built - see

python -m xformers.info` for more info
unsupported embed per head: 512

ZZono50 @Dr. Furkan Gözükara using SUPIR and getting these errors, any suggestions? dec...

F

Furkan Gözükara SECourses•9/16/24, 9:27 PM

did you do a fresh install of latest version_

F

Furkan Gözükara SECourses•9/16/24, 9:27 PM

this is xformers incompability

F

Furkan Gözükara SECourses•9/16/24, 9:27 PM

my installer shouldnt have this issue

Z

Zono50•9/16/24, 9:29 PM

i'm using arch on my local system

ZZono50 i'm using arch on my local system

F

Furkan Gözükara SECourses•9/16/24, 9:31 PM

xformers only for nvidia gpus afaik

S

slipkn67•9/16/24, 9:32 PM

how do I add realistic scars, do I need to create a LORA for that?

Sslipkn67 how do I add realistic scars, do I need to create a LORA for that?

F

Furkan Gözükara SECourses•9/16/24, 9:44 PM

you can search for better prompting too

F

Furkan Gözükara SECourses•9/16/24, 9:44 PM

but lora may work better

F

Furkan Gözükara SECourses•9/16/24, 9:46 PM

th

F

Furkan Gözükara SECourses•9/16/24, 9:47 PM

this config failed so bad

F

Furkan Gözükara SECourses•9/16/24, 9:48 PM

man i need 16 gpus having machine :d

S

slipkn67•9/16/24, 11:08 PM

why you want to be bald bro?

FFurkan Gözükara SECourses you can search for better prompting too

S

slipkn67•9/16/24, 11:10 PM

I tried facial scar, knife cut scar etc, it draws a bloody mess lol, but I want a cool scar like I'm some ancient warrior

Sslipkn67 I tried facial scar, knife cut scar etc, it draws a bloody mess lol, but I want ...

F

Furkan Gözükara SECourses•9/16/24, 11:21 PM

probably you need lora

Sslipkn67 why you want to be bald bro?

F

Furkan Gözükara SECourses•9/16/24, 11:22 PM

testing

Z

Zuck•9/16/24, 11:46 PM

@Dr. Furkan Gözükara Have you tried training a LORA for an object? I'm finding when i go to generate the image using the LORA i have to describe the object in detail for it to retain most of the objects details

ZZuck @Dr. Furkan Gözükara Have you tried training a LORA for an object? I'm finding w...

F

Furkan Gözükara SECourses•9/16/24, 11:49 PM

i only tried a style yet

F

Furkan Gözükara SECourses•9/16/24, 11:49 PM

it worked very well

F

Furkan Gözükara SECourses•9/16/24, 11:49 PM

did you caption object during training?

Z

Zuck•9/16/24, 11:51 PM

Ya i captioned really well, assigned trigger word, and described everything around the object i was training but didn't describe the objects details.

I wonder if i have to increase the LORA's strength when generating images...

Z

Zuck•9/16/24, 11:52 PM

to be fair ive been using ai-toolkit, going to watch your kohya video, does kohya have a more up to date training method?

Z

Zuck•9/16/24, 11:54 PM

I've done it succesfully with SD, but flux is a whole other beast

F

Furkan Gözükara SECourses•9/16/24, 11:55 PM

red color blue beard almost made it fine tuning amazing