Software Engineering Courses (SECourses)•3y ago

check if the settings in the printer work for the type of paper you used, and which tray you used fo

check if the settings in the printer work for the type of paper you used, and which tray you used for the paper, sometimes printers give errors when using different type of paper or paper format and using different trays... but i don't know what could be making do lines on the paper except the printer

AAlaBeat check if the settings in the printer work for the type of paper you used, and wh...

Trimac•3/26/23, 2:17 PM

Im not sure what to do wd tray but i saw in the rear tray in settings page. Btw, this is cannon g1010.

Trimac•3/26/23, 2:17 PM

Sorry for my english

AAlaBeat i found out, i'm installing 0.0.17.dev484, hope it's good and doesn't brick sd

Furkan Gözükara SECourses•3/26/23, 2:22 PM

activate venv

Furkan Gözükara SECourses•3/26/23, 2:23 PM

pip uninstall xformers

Furkan Gözükara SECourses•3/26/23, 2:24 PM

there are several dev versions

Furkan Gözükara SECourses•3/26/23, 2:24 PM

0.0.17.dev480, 0.0.17.dev481, 0.0.17.dev482, 0.0.17.dev483, 0.0.17.dev484, 0.0.17rc481, 0.0.17rc482

Furkan Gözükara SECourses•3/26/23, 2:24 PM

you can try latest and see if works

Furkan Gözükara SECourses•3/26/23, 2:24 PM

rc means release candidate

TTrimac Hey guys how are you? I have a little project printing some Ai images and create...

Furkan Gözükara SECourses•3/26/23, 2:25 PM

they look cool

FFurkan Gözükara SECourses follow this tutorial it works great https://youtu.be/oI0GEnrySS0

mathieu•3/26/23, 2:32 PM

i get the same error with today's uptdate..
using : git checkout and an other commit fix the " -f --port 3000" issue
but creat other compatibility problems ...

Mmathieu i get the same error with today's uptdate.. using : git checkout and an other co...

Furkan Gözükara SECourses•3/26/23, 2:47 PM

automatic1111 pushed so many updates without testing

Furkan Gözükara SECourses•3/26/23, 2:47 PM

it causes issues :/

NNate this is how the whole lot looks. Any help would be greatly appreciated. So far...

mathieu•3/26/23, 2:47 PM

look like it work now

i've used : git checkout a9eef1fbb1dcdce4f0eb0b8e0f79dcd4c96713e1
after : git pull

and use : pip install -r requirements.txt
in dreambooth AND Webui folder befor restarting the pod

FFurkan Gözükara SECourses 0.0.17.dev480, 0.0.17.dev481, 0.0.17.dev482, 0.0.17.dev483, 0.0.17.dev484, 0.0.1...

AlaBeatOP•3/26/23, 3:09 PM

these want to install torch 2 and i don't want it because it creates issues with my sd, happened before when i tried to put torch 2

FFurkan Gözükara SECourses welcome. sorry for late reply

genialgenteel•3/26/23, 3:16 PM

Hi, thanks for responding! It never gets past that point, it freezes there and never actually starts training. It even cuts off the text right the way I typed it and never actually says what the values were not passed to. It just says "were not passed to" and it's blank after that.

I'm trying a different method now by preparing a dataset to try training a LoRA instead to see if that works better. It's taking a while since I have a 1000+ pictures to caption, but we'll see if it works this time.

AAlaBeat these want to install torch 2 and i don't want it because it creates issues with...

Furkan Gözükara SECourses•3/26/23, 3:18 PM

i see

Furkan Gözükara SECourses•3/26/23, 3:28 PM

i have written a layout for the video it is 2000 words lol

Furkan Gözükara SECourses•3/26/23, 3:28 PM

now i have to record it

sainthell•3/26/23, 3:35 PM

hello.. what does it mean "fatal:reference is not a tree" when i'm trying to change dreambooth version of a1111. i got this commit from the github page and doesn't seem to be working. but it seem like whatever i try i get the same result

Zono50•3/26/23, 4:27 PM

i'f i'm training a model of myself and i'm already seeing myself at 10 epochs, is that bad?

felixsanz•3/26/23, 4:55 PM

@Furkan Gözükara SECourses do you use v1-5-pruned.safetensor or v1-5-pruned-emaonly.safetensor for training? and why?

felixsanz•3/26/23, 4:56 PM

i still don't understand ema vs nonema vs fullema... everybody says one thing and then the opposite..

Haashim•3/26/23, 6:00 PM

@Furkan Gözükara SECourses my bad I did not make it to my workstation yet today

Haashim•3/26/23, 6:00 PM

Did you get the deflicker ?

Stevens•3/26/23, 6:11 PM

I am trying to do it now but I have not used the program yet so I just have this.

HHaashim @Furkan Gözükara SECourses my bad I did not make it to my workstation yet today

Furkan Gözükara SECourses•3/26/23, 6:35 PM

i tried a demo

Furkan Gözükara SECourses•3/26/23, 6:35 PM

and deflicker makes huge diff

Furkan Gözükara SECourses•3/26/23, 6:35 PM

i tested 2x demo deflicker from some other company

Furkan Gözükara SECourses•3/26/23, 6:35 PM

so 5x would make significant difference

Original message was deleted

Furkan Gözükara SECourses•3/26/23, 6:35 PM

your training set have blurry images?

Ffelixsanz @Furkan Gözükara SECourses do you use v1-5-pruned.safetensor or v1-5-pruned-emao...

Furkan Gözükara SECourses•3/26/23, 6:35 PM

v1-5-pruned.safetensor better for training

SStevens I am trying to do it now but I have not used the program yet so I just have this...

Furkan Gözükara SECourses•3/26/23, 6:36 PM

ye you add 5 and it makes huge improvement

FFurkan Gözükara SECourses your training set have blurry images?

Szutykos Macska•3/26/23, 6:36 PM

No, i mean the pictures when i generate, when doing the sampling steps

Stevens•3/26/23, 6:36 PM

I think I am running it - my computer is starting to smoke!

stavsap•3/26/23, 6:37 PM

can anyone please share his loss graph over the steps in dreambooth?

stavsap•3/26/23, 6:37 PM

do you manage to get a stable loss decrease?

FFurkan Gözükara SECourses v1-5-pruned.safetensor better for training

Szutykos Macska•3/26/23, 6:38 PM

but if i use nothing, but only the triggerword for myself, i get better resoults, if i use another one like a cap or something things go wild, if the picture is generated for a bit further my faces starts to mutate

Stevens•3/26/23, 6:38 PM

We might need a Junior Varsity team for SD and I will def be in it