Search Setup for Free

Software Engineering Courses (SECourses)•3y ago

Is 6700 steps enough for sdxl?

Is 6700 steps enough for sdxl?

Software Engineering Courses (SECourses)Join

Hello everyone. I am Dr. Furkan Gözükara. PhD Computer Engineer. SECourses is a dedicated YouTube channel for the following topics : Tech, AI, News, Science, Robotics, Singularity, ComfyUI, SwarmUI, ML, Artificial Intelligence, Humanoid Robots, Wan 2.2, FLUX, Krea, Qwen Image, VLMs, Stable Diffusion

11,654Members

View on Discord

Resources

Similar Threads

Was this page helpful?

Product

Pricing Docs Communities

Resources

About Blog Changelog Contributors

Legal

Terms Privacy Cookies EULA

Twitter GitHub Discord

© 2026 Hedgehog Software, LLC. All rights reserved.

© 2026 Hedgehog Software, LLC

Twitter GitHub Discord

More

Communities Docs About Terms Privacy

mikemenders

M

mikemenders•10/9/23, 5:14 AM

and here is a few portrait

40450-2492411505-bbl_woman_blue_dress_short_hair_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many.png

40449-1065977894-bbl_woman_blue_dress_short_hair_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many.png

40447-1578977074-bbl_woman_green_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sit.png

40446-1230343363-bbl_woman_green_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sit.png

40445-1387945743-bbl_woman_green_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sit.png

mikemenders

M

mikemenders•10/9/23, 5:19 AM

generates stable images even under aux10 model

40464-2163930282-bbl_woman_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sitting_i.png

40463-1992744207-bbl_woman_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sitting_i.png

40462-1257673000-bbl_woman_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sitting_i.png

40461-1801482298-bbl_woman_dress_short_hair_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_peop.png

40460-1897398850-bbl_woman_dress_short_hair_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_peop.png

40459-1473175472-bbl_woman_dress_short_hair_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_peop.png

40458-3196843830-bbl_woman_dress_short_hair_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_peop.png

mikemenders

M

mikemenders•10/9/23, 5:26 AM

With artUniverse model:

40489-3077305307-bbl_woman_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sitting_i.png

40488-3117019385-bbl_woman_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sitting_i.png

40487-2920139718-bbl_woman_dress_sitting_in_a_chair_restaurant_some_man_in_background_portrait_look_at_viewer_and_many_people_sitting_i.png

Furkan Gözükara SECourses

FFurkan Gözükara SECourses i think memory efficient attention doing nothing 😄

Metaphysix

M

Metaphysix•10/9/23, 5:39 AM

Agree I made some test for speed/memory. No gradient checkpointing:
xformers + Mem Efficient Attention = xformers without Mem. E. attention
Same speed, same memory usage.

mikemenders

M

mikemenders•10/9/23, 6:37 AM

I noticed a very interesting thing. I'm training for style and I haven't changed anything else, just the number of epochs from 10 to 30, and I get completely different sample images and the training values are absolutely different. I'm using Prodigy now.

Praveen Miriyala

PPraveen Miriyala I'm training LoRa SDXL on my 3060 with 12GB VRAM, 768*768 resolution and it's sh...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 9:27 AM

yep looking decent

mikemenders

Mmikemenders It's interesting, because I also have an RTX 3060 12 GB card and I can train a p...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 9:27 AM

512x512 much lower resolution

JJM Is 6700 steps enough for sdxl?

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 9:28 AM

it is decent steps yes

Metaphysix

MMetaphysix Agree I made some test for speed/memory. No gradient checkpointing: xformers + M...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 9:28 AM

yep i noticed same

Furkan Gözükara SECourses

FFurkan Gözükara SECourses yep looking decent

Praveen Miriyala

P

Praveen Miriyala•10/9/23, 9:28 AM

but result was not good

mikemenders

Mmikemenders I noticed a very interesting thing. I'm training for style and I haven't changed...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 9:28 AM

i never got good results with Prodigy

Praveen Miriyala

PPraveen Miriyala but result was not good ☹️

Furkan Gözükara SECourses

F

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

try bf16 too

Furkan Gözükara SECourses

F

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

but reason could be 768x768

Furkan Gözükara SECourses

F

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

also add this parameter to optimizer : --max_grad_norm=0.0

Furkan Gözükara SECourses

F

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

i will do another research for LoRA

Praveen Miriyala

P

Praveen Miriyala•10/9/23, 9:30 AM

ok i will follow these tips for next training

J

JMOP•10/9/23, 9:38 AM

: A tensor with all NaNs was produced in VAE. This could
be because there's not enough precision to represent the picture. Try adding --no-half-vae commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check. How to fix?

SuperTurboHero

S

SuperTurboHero•10/9/23, 10:05 AM

I wonder what is quality difference if I don't use classification images? training would be much faster, but would I lose a lot of quality?

mikemenders

M

mikemenders•10/9/23, 10:40 AM

I used this parameters for Prodigy, and trained SD 1.5 base model: "use_bias_correction=True" "weight_decay=0.5" "decouple=True" "betas=(0.9, 0.99)" "d_coef=2" "safeguard_warmup=False"

JJM : A tensor with all NaNs was produced in VAE. This could be because there's not ...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 10:41 AM

this happens when model totally overtrained

Furkan Gözükara SECourses

FFurkan Gözükara SECourses this happens when model totally overtrained

J

JMOP•10/9/23, 10:57 AM

I don't use with model lora I use model checkpoint Every model doesn't work.Basicly sd 1.5 I found this problem.

Sandstormleader

S

Sandstormleader•10/9/23, 10:57 AM

so is SDXL training essentially the same as 1.5 training just using an SDXL model? @Dr. Furkan Gözükara

Sandstormleader

SSandstormleader so is SDXL training essentially the same as 1.5 training just using an SDXL mode...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 11:15 AM

the parameters changes too

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 11:15 AM

learning rate

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 11:15 AM

Sandstormleader

S

Sandstormleader•10/9/23, 11:18 AM

so not 0.000001

Sandstormleader

SSandstormleader so not 0.000001

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 11:19 AM

yes

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 11:19 AM

I am searching best parameters for SDXL DreamBooth for weeks now

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 11:38 AM

Huge Stable Diffusion XL (SDXL) Text Encoder DreamBooth training comparison

U-NET is always trained

PNG info (prompts) are in alts

All images are 1024x1024 so download full sizes

https://twitter.com/GozukaraFurkan/status/1711345323900100643

Benjamin

B

Benjamin•10/9/23, 11:40 AM

v23.10.0 adetailer got released. Just noticed because now in the preview of the image a red box appears around the detected face

Furkan Gözükara SECourses

FFurkan Gözükara SECourses Huge Stable Diffusion XL (SDXL) Text Encoder DreamBooth training comparison U-N...

Benjamin

B

Benjamin•10/9/23, 11:41 AM

good job, from a first glance it does not look like it is changing that much? what is your impression

Benjamin

BBenjamin v23.10.0 adetailer got released. Just noticed because now in the preview of the ...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 12:44 PM

2023-10-07

- v23.10.0
- 허깅페이스 모델을 다운로드 실패했을 때, 계속 다운로드를 시도하지 않음
- img2img에서 img2img단계를 건너뛰는 기능 추가
- live preview에서 감지 단계를 보여줌 (PR #352)

Benjamin

BBenjamin good job, from a first glance it does not look like it is changing that much? wh...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 12:45 PM

i think it still adds some more extra details

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 12:45 PM

so 2e-6 looks like nice. 3e-6 also nice but a little bit starting to cook

Furkan Gözükara SECourses

FFurkan Gözükara SECourses i think it still adds some more extra details

Benjamin

B

Benjamin•10/9/23, 12:47 PM

how did it influence the training time?

Benjamin

B

Benjamin•10/9/23, 12:48 PM

ah you already wrote in patreon it is slower

Benjamin

BBenjamin ah you already wrote in patreon it is slower

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 12:54 PM

ye a little bit more slower

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 12:54 PM

like 1.5 second / it to 1.75 second / it

Stefanoid9343

S

Stefanoid9343•10/9/23, 1:11 PM

Thank you so much to @Dr. Furkan Gözükara and @Sandstormleader after several hours of training and solving some errors the model came as I expected this is definitely a game changer process

Furkan Gözükara SECourses

FFurkan Gözükara SECourses i never got good results with Prodigy

mikemenders

M

mikemenders•10/9/23, 1:14 PM

I have now made a perfect Lora with Prodigy, but you have to use Clip Skip 2 when generating and then it will be flexible.

mikemenders

Mmikemenders I have now made a perfect Lora with Prodigy, but you have to use Clip Skip 2 whe...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 1:20 PM

interesting never tested clip skip 2. this is sd 1.5?

Stefanoid9343

SStefanoid9343 Thank you so much to @Dr. Furkan Gözükara and @Sandstormleader after several hou...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 1:20 PM

awesome

Furkan Gözükara SECourses

FFurkan Gözükara SECourses interesting never tested clip skip 2. this is sd 1.5?

mikemenders

M

mikemenders•10/9/23, 1:23 PM

Yes, I trained with RealisticVision 5.1, and I generated the images in the same model.

mikemenders

M

mikemenders•10/9/23, 1:24 PM

If I generated on clip skip 1, the model was inflexible, but on clip skip 2 everything was fine.
And I thought right: I trained it with clip skip 2 and then clip skip 1 can be generated well.

Metaphysix

M

Metaphysix•10/9/23, 1:31 PM

So now the SDXL training works properly with TE, but what's up with 1.5 dreambooth?

rafraf

R

rafraf•10/9/23, 1:51 PM

what is the best/fastest gpu deal on Runpod to do inference on Auto1111? A100 or A6000? I find A6000 very slow, slower than 4900...

Metaphysix

MMetaphysix So now the SDXL training works properly with TE, but what's up with 1.5 dreamboo...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 1:51 PM

yes but using kohya

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 1:52 PM

i think sd 1.5 dreambooth also should work but need good hyper parameter testing. also kohya missing EMA

rafraf

Rrafraf what is the best/fastest gpu deal on Runpod to do inference on Auto1111? A100 or...

Furkan Gözükara SECourses

F

Furkan Gözükara SECourses•10/9/23, 1:52 PM

4090 could be fastest with price ratio

Furkan Gözükara SECourses

FFurkan Gözükara SECourses 4090 could be fastest with price ratio

rafraf

R

rafraf•10/9/23, 1:56 PM

ok, yea, that's what i thought. a6000 takes twice as long to infer