Software Engineering Courses (SECourses)•3y ago

yes really good. for distant shot inpainting face fix. or you can add much more distant shots to try

yes really good. for distant shot inpainting face fix. or you can add much more distant shots to try to get good results in your training set

EEnlil Wolf Just wanted to stop & restart it because I wanted to upload classification files...

Furkan Gözükara SECoursesOP•3/9/23, 10:44 AM

solution is easy

Furkan Gözükara SECoursesOP•3/9/23, 10:44 AM

save config

Furkan Gözükara SECoursesOP•3/9/23, 10:44 AM

and load config

Furkan Gözükara SECoursesOP•3/9/23, 10:44 AM

also dont forget to put back slash to beginning of folder path

Furkan Gözükara SECoursesOP•3/9/23, 10:44 AM

in runpod

Furkan Gözükara SECoursesOP•3/9/23, 10:44 AM

e.g. /workspace/folder

SSibun hello guys Ive use dreambooth colab to train a model, after convert weights to c...

Furkan Gözükara SECoursesOP•3/9/23, 10:46 AM

this is probably related to google colab web ui script you are using

Furkan Gözükara SECoursesOP•3/9/23, 10:46 AM

if you can afford runpod works as smooth as pc

Ggatsby Hey everyone 🙂 when using dreambooth and generating class images, is it better ...

Furkan Gözükara SECoursesOP•3/9/23, 10:46 AM

actually not necessary

FFurkan Gözükara SECourses yes really good. for distant shot inpainting face fix. or you can add much more ...

P1pSl1p•3/9/23, 10:49 AM

I will try that next. Need more images for the dataset. I've learned so much here. Thank you for providing this space.

Furkan Gözükara SECoursesOP•3/9/23, 11:06 AM

FFurkan Gözükara SECourses if you can afford runpod works as smooth as pc

Sibun•3/9/23, 12:00 PM

i think my laptop cant run stable diffusion, 1650 suck

Ma7adi805•3/9/23, 12:08 PM

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB (GPU 0; 4.00 GiB total capacity; 3.36 GiB already allocated; 0 bytes free; 3.44 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB (GPU 0; 4.00 GiB total capacity; 3.36 GiB already allocated; 0 bytes free; 3.44 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

how can I fix this

SSibun i think my laptop cant run stable diffusion, 1650 suck 🥺

Furkan Gözükara SECoursesOP•3/9/23, 12:35 PM

it runs on gtx 1050 ti

Furkan Gözükara SECoursesOP•3/9/23, 12:35 PM

use --lowvram

Furkan Gözükara SECoursesOP•3/9/23, 12:35 PM

but you cant train

MMa7adi805 ``` torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB...

Furkan Gözükara SECoursesOP•3/9/23, 12:35 PM

you cant do training with that

Furkan Gözükara SECoursesOP•3/9/23, 12:36 PM

but you can generate images

Furkan Gözükara SECoursesOP•3/9/23, 12:36 PM

--medvram

Furkan Gözükara SECoursesOP•3/9/23, 12:36 PM

if not works

Furkan Gözükara SECoursesOP•3/9/23, 12:36 PM

--lowvram

FFurkan Gözükara SECourses but you can generate images

Ma7adi805•3/9/23, 1:01 PM

I got that output when it was generating img2img

FFurkan Gözükara SECourses --medvram

Ma7adi805•3/9/23, 1:01 PM

will try

Ma7adi805•3/9/23, 1:08 PM

it worked thanks

Ma7adi805•3/9/23, 1:10 PM

I have a question, when I upload my face in img2img, it doesnt show my face, and even if there was a face it barely looks like me

MMa7adi805 I have a question, when I upload my face in img2img, it doesnt show my face, and...

Furkan Gözükara SECoursesOP•3/9/23, 1:45 PM

it doesnt work that way

Furkan Gözükara SECoursesOP•3/9/23, 1:45 PM

you have to teach your face

Ma7adi805•3/9/23, 1:49 PM

oh ok, another question might be my last, in the terminal it says

To create a public link, set `share=True` in `launch()`.

To create a public link, set `share=True` in `launch()`.

I want to do that how can i do it I cannot find launch()

Thermidor•3/9/23, 3:20 PM

In the file webui.sh where it says args type --share then save this file and run again

TThermidor In the file webui.sh where it says args type --share then save this file and run...

Furkan Gözükara SECoursesOP•3/9/23, 3:44 PM

correct

Ruairi Robinson•3/9/23, 4:10 PM

getting this:

>git pull
Already up to date.

but it hasnt been updated in over a week, i think its broken

RRuairi Robinson getting this: >git pull Already up to date. but it hasnt been updated in over...

Ma7adi805•3/9/23, 4:11 PM

I think you need to save the new one that is on git

RRuairi Robinson getting this: >git pull Already up to date. but it hasnt been updated in over...

Furkan Gözükara SECoursesOP•3/9/23, 4:35 PM

can you show screenshot

Furkan Gözükara SECoursesOP•3/9/23, 4:35 PM

it has been updated many times this week

Furkan Gözükara SECoursesOP•3/9/23, 4:35 PM

controlnet extension

Ruairi Robinson•3/9/23, 4:35 PM

i think i figured it out, deleted dreambooth extension folder and it let me update everything

bush•3/9/23, 5:38 PM

Getting Loss=0.00 during training, what does it mean?

Haru•3/9/23, 5:39 PM

hi, i need help finding were is override setings ?

PP1pSl1p yeah, I'm starting to see that. slowly.lol

Steel•3/9/23, 5:39 PM

if you want better faces and inpainting, train a model of the person where you only have closeup crops of the face. super tight, then use that model for inpaint. You can also use both loras at the same time, like 0.4 strenght on each

Haru•3/9/23, 5:39 PM