Software Engineering Courses (SECourses)•3y ago

I stuck in 195 step out of 550 for 15 min. but resource usage shows something is working. what shoul

I stuck in 195 step out of 550 for 15 min. but resource usage shows something is working. what should I do?

Should I restart the pod?

Screen_Shot_2023-03-09_at_5.00.46_PM.png

Screen_Shot_2023-03-09_at_5.01.05_PM.png

vjavad1443OP•3/10/23, 12:07 AM

and after restar how can I resume the process and avoid generating images from the start ?

Vvjavad1443 I stuck in 195 step out of 550 for 15 min. but resource usage shows something is...

Furkan Gözükara SECourses•3/10/23, 12:08 AM

yes you can resume

Furkan Gözükara SECourses•3/10/23, 12:08 AM

just click load

Furkan Gözükara SECourses•3/10/23, 12:17 AM

I made a new video about how to generate nice text with controlnet. which one of these looks best to use in thumbnail do you think? you can download entire folder to check on your computer as well. thank you for suggestion : https://drive.google.com/drive/folders/1c3npVB7kJ3meZDsjOnUnhHhXilJEmiWb?usp=share_link

FFurkan Gözükara SECourses I made a new video about how to generate nice text with controlnet. which one o...

dave5555•3/10/23, 2:09 AM

cool, i like the ones with shadows best , the ones that interplay with the background etc ......... can u control fonts ?

dave5555•3/10/23, 2:20 AM

is there an extension that lets you provide a batch folder for control net? for example have a batch folder to feed open pose , while keeping the img2img guide image as normal

dave5555•3/10/23, 2:20 AM

i saw a thread in github where they were waiting for merge approval from auto

dave5555•3/10/23, 2:20 AM

i feel like i saw something somewhere that allows it

dave5555•3/10/23, 2:22 AM

current work around now is to turn off img2img influence and turn off init image in open pose and run a batch this way , and then re run it again as normal img2img map using depth control added on , but it takes double the time it would take to just be able to batch the control nets in the first place

dave5555•3/10/23, 2:24 AM

im on a 1080 and am running out of vram with some control net things im trying to do , i have never used command line args for low vram , or med vram or anything , would that help me get more control net without sacrificing anything?

FFurkan Gözükara SECourses I made a new video about how to generate nice text with controlnet. which one o...

Haashim•3/10/23, 3:01 AM

Almost too many to choose. I like the ones with nice shadow

Haashim•3/10/23, 3:02 AM

Fires ones are cool too. A little tougher to read some though.

Ddave5555 i was trying to port over the kohya trainer from colab to jupyter and get hung u...

Tesli23•3/10/23, 3:30 AM

It's possible but requires a few modifications to cells. I ported dreambooth colab to Jupyter notebook

TTesli23 It's possible but requires a few modifications to cells. I ported dreambooth col...

dave5555•3/10/23, 5:04 AM

what are main things? is it something to do with what depencies colab already has? or something in the workspace file system? sorry im noob

dave5555•3/10/23, 5:05 AM

i use the pytorch notebook , so doesnt that have pytorch pre installed?

dave5555•3/10/23, 5:07 AM

runpod also has a guide for bypassing google gpu and ssh in to runpod instance , i wonder if this is better?

Steel•3/10/23, 5:41 AM

wow they finally fixed dreambooth lora

vjavad1443OP•3/10/23, 8:09 AM

@Furkan Gözükara SECourses I feel something isn't right. I have followed your method for training the face but it takes up to 2 hours ( 30-45 min generating classification & the rest for training ) and the final result isn't bad. But I see other people are able to do this job in 10-15 min with a decent quality.

Have you tested fast stable diffusion ?

mathieu•3/10/23, 9:08 AM

hello, i'm doing DreamBooth Update tutorial but i get this error when i clic 'TRAIN' :

Exception training model: 'name 'str2optimizer8bit_blockwise' is not defined'.

Mmathieu hello, i'm doing DreamBooth Update tutorial but i get this error when i clic 'T...

Furkan Gözükara SECourses•3/10/23, 9:26 AM

hi. what graphic card do you have

Vvjavad1443 @Furkan Gözükara SECourses I feel something isn't right. I have followed your me...

Furkan Gözükara SECourses•3/10/23, 9:26 AM

it totally depends on your gpu and number of training images you have

HHaashim Almost too many to choose. I like the ones with nice shadow

Furkan Gözükara SECourses•3/10/23, 9:26 AM

Ddave5555 cool, i like the ones with shadows best , the ones that interplay with the backg...

Furkan Gözükara SECourses•3/10/23, 9:29 AM

yes this is controlnet i made a new video

FFurkan Gözükara SECourses hi. what graphic card do you have

mathieu•3/10/23, 9:31 AM

RTX 4090
12 vCPU 62 GB RAM

5 GB Disk
150 GB Volume
VRAM was up to 24G

mathieu•3/10/23, 9:34 AM

is it because of the FP16 / BF16 option ?

Mmathieu RTX 4090 12 vCPU 62 GB RAM 5 GB Disk 150 GB Volume VRAM was up to 24G

Furkan Gözükara SECourses•3/10/23, 9:34 AM

i see so gpu is not the case

Furkan Gözükara SECourses•3/10/23, 9:34 AM

it must be either installation or settings mistake

Furkan Gözükara SECourses•3/10/23, 9:35 AM

you followed this install tutorial?

Furkan Gözükara SECourses•3/10/23, 9:35 AM

https://youtu.be/oI0GEnrySS0

YouTubeSECourses - Software Engineering Courses

How To Install New DreamBooth Extension On RunPod - Automatic1111 W...

RunPod: https://bit.ly/RunPodIO. Discord : https://bit.ly/SECoursesDiscord. How to install and start using #DreamBooth extension on #RunPod SD Web UI template. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on

https://www.patreon.com/SECourses

DreamBooth extension: ht...

Furkan Gözükara SECourses•3/10/23, 9:35 AM

and how to config explained in here i mean dreambooth settings https://youtu.be/sRdtVanSRl4

YouTubeSECourses - Software Engineering Courses

Epic Web UI DreamBooth Update - New Best Settings - 10 Stable Diffu...

RunPod: https://bit.ly/RunPodIO. Discord : https://bit.ly/SECoursesDiscord. Web UI DreamBooth got epic update and we tested all new features to find the best settings (10). If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on

https://www.patreon.com/SECourses

Playlist of...

mathieu•3/10/23, 9:40 AM

i followed the second
the first video got uploaded 1 day ago ???
you are crazy fast !!!

Neuromasters•3/10/23, 10:30 AM

what manipulations can improve the detail of the background?

01628-851013001-snowapocalypse_film_Harry_Potter_1.3_clear_face__eyes_reflections_detailed_eyes_perfect_eyes_rendered_eyes.jpg

Mmathieu i followed the second the first video got uploaded 1 day ago ??? you are crazy ...

Furkan Gözükara SECourses•3/10/23, 11:00 AM

yes

NNeuromasters what manipulations can improve the detail of the background?

Furkan Gözükara SECourses•3/10/23, 11:01 AM

it looks blurry

Furkan Gözükara SECourses•3/10/23, 11:01 AM

perhaps inpainting

Furkan Gözükara SECourses•3/10/23, 11:01 AM

or hd upscaler

vjavad1443OP•3/10/23, 2:22 PM

Ok, after I trained my model one time now I wanted to train one of them that looks better and continue the training

So I created a new model and in source checkpoint I selected one of them but when I clicked on Create model I get this error:

FileNotFoundError: [Errno 2] No such file or directory: 'models/dreambooth/der_2640/der_2640.yaml' -> 'models/dreambooth/der_2640/der_2640.yaml'
IndexError: list index out of range

Which is correct, because the path for .safetensors are here:
/model/stable_diffusion/der/der_2640.yaml

But as you see it's trying to read from another path that file isn't existed
models/dreambooth/der_2640/der_2640.yaml

What should I do in this case?

Furkan Gözükara SECourses•3/10/23, 2:25 PM

make sure files are in that folder

vjavad1443OP•3/10/23, 2:29 PM

you mean do I need to move those files to that path manually ?

Furkan Gözükara SECourses•3/10/23, 2:30 PM

yes

Furkan Gözükara SECourses•3/10/23, 2:30 PM

you are running this on

Furkan Gözükara SECourses•3/10/23, 2:30 PM

colab?

vjavad1443OP•3/10/23, 2:30 PM

no, on runpod