Skip to content

How To Properly Setup OneTrainer Concepts And Some Other Options Quick Tutorial Before Big One

Jump to bottom
FurkanGozukara edited this page Oct 21, 2025 · 1 revision

How To Properly Setup OneTrainer Concepts And Some Other Options - Quick Tutorial Before Big One

Full tutorial link > https://www.youtube.com/watch?v=yPOadldf6bI

How To Properly Setup OneTrainer Concepts And Some Other Options - Quick Tutorial Before Big One

image Hits Patreon BuyMeACoffee Furkan Gözükara Medium Codio Furkan Gözükara Medium

YouTube Channel Furkan Gözükara LinkedIn Udemy Twitter Follow Furkan Gözükara

OneTrainer Stable Diffusion XL (SDXL) Fine Tuning Best Presets : https://www.patreon.com/posts/96028218

The Very Best OneTrainer Workflow & Config For SD 1.5 Based Models DreamBooth / Full Fine Tuning :

https://www.patreon.com/posts/97381002

Full Workflow For Newbie Stable Diffusion Trainers For SD 1.5 Models & SDXL Models Training With DreamBooth & LoRA :

https://www.patreon.com/posts/full-workflow-sd-98620163

5200 Ground Truth Regularization Images for Man & Woman (each one has 5200 images) :

https://www.patreon.com/posts/massive-4k-woman-87700469

OneTrainer repo : https://github.com/Nerogar/OneTrainer

Video Transcription

  • 00:00:03 in this video i going to Show you How to use our very best one trainer

  • 00:00:05 use our very best one trainer configurations actually i am working on

  • 00:00:08 configurations actually i am working on a full public tutorial Which in which i

  • 00:00:11 a full public tutorial Which in which i will Share All of The Hyper parameters

  • 00:00:14 will Share All of The Hyper parameters and settings in details and i will

  • 00:00:16 and settings in details and i will explain How to use One trainer but it is getting

  • 00:00:18 explain How to use One trainer but it is getting delayed therefore Today i going to Show

  • 00:00:21 delayed therefore Today i going to Show you How to use these

  • 00:00:23 you How to use these configurations We have configurations

  • 00:00:25 configurations We have configurations for 1.5 based mod

  • 00:00:28 for 1.5 based mod and

  • 00:00:30 and mod so it doesn't matter both of working

  • 00:00:33 mod so it doesn't matter both of working exactly Same downlo the that vram supports

  • 00:00:38 exactly Same downlo the that vram supports for example if you have 12 gab

  • 00:00:41 for example if you have 12 gab use the 10.6 gab if you have 16 gab use

  • 00:00:45 use the 10.6 gab if you have 16 gab use the Tire 2 if you have 24 gab use this

  • 00:00:48 the Tire 2 if you have 24 gab use this One it is Same for sxl As Well so Let's

  • 00:00:52 One it is Same for sxl As Well so Let's for example use this 14.5 GB sxl

  • 00:01:04 After downloading it go to One trainer installation and paste it

  • 00:01:07 to One trainer installation and paste it into Training presets folder so This is

  • 00:01:11 into Training presets folder so This is you need to put inside Training presets

  • 00:01:15 you need to put inside Training presets Start One trainer

  • 00:01:36 him from the settings configuration this was The

  • 00:01:39 settings configuration this was The downloaded One so what you need to

  • 00:01:42 downloaded One so what you need to change Here is change your workspace

  • 00:01:44 change Here is change your workspace directory according to Your hard Drive

  • 00:01:46 directory according to Your hard Drive for example Let's it As My F Drive and

  • 00:01:51 for example Let's it As My F Drive and Let's Say One

  • 00:01:58 trainer so All of theor

  • 00:02:12 and Select the folder and Get selected in Here you don't Need To anything el in

  • 00:02:15 in Here you don't Need To anything el in the mod Select mods downloading from

  • 00:02:19 the mod Select mods downloading from hugging Face repositories you canly give

  • 00:02:21 hugging Face repositories you canly give the mod from here

  • 00:02:32 cck supported model output Destination Now This is Super important you need to

  • 00:02:34 Now This is Super important you need to Give full pad Don't Forget that Give

  • 00:02:37 Give full pad Don't Forget that Give full pad with safe T Source

  • 00:02:40 full pad with safe T Source extension these are all the best

  • 00:02:42 extension these are all the best settings so You don't need to change

  • 00:02:44 settings so You don't need to change anything

  • 00:02:46 anything there in the

  • 00:02:48 there in the data enable latent

  • 00:02:51 data enable latent caching it increases the Speed ​​ and also

  • 00:02:55 caching it increases the Speed ​​and also reduces vram and the most important part is

  • 00:02:59 reduces vram and the most important part is concepts in the concepts you need to ADD your Concept

  • 00:03:02 concepts in the concepts you need to ADD your Concept One by One for example

  • 00:03:05 One by One for example Let's ADD Training images Concept so

  • 00:03:08 Let's ADD Training images Concept so After clicking there type As Training images

  • 00:03:11 After clicking there type As Training images Give any name you can enable or

  • 00:03:14 Give any name you can enable or disable concepts All of The enabled concepts will be Used

  • 00:03:16 disable concepts All of The enabled concepts will be Used One trainer doesn't have Dream

  • 00:03:18 One trainer doesn't have Dream

  • 00:03:24 however We are able to uti regularization images that We have from un

  • 00:03:26 regularization images that We have from un that i have collected Select the p of

  • 00:03:30 that i have collected Select the p of the images for example My images are

  • 00:03:32 the images for example My images are Here will Select them

  • 00:03:35 Here will Select them ok Here This is ST 1.5 based config so

  • 00:03:40 ok Here This is ST 1.5 based config so Let's Select the

  • 00:03:43 Let's Select the folder and in

  • 00:03:46 folder and in Here you need to set from text file Per

  • 00:03:50 Here you need to set from text file Per sample or from single text file so from text

  • 00:03:54 sample or from single text file so from text file Per sample means That it will

  • 00:03:58 file Per sample means That it will look for the captions with the same name

  • 00:04:01 look for the captions with the same name As the file name And it will read them

  • 00:04:04 As the file name And it will read them however i don't use captioning for

  • 00:04:05 however i don't use captioning for Training persons so make a caption Here

  • 00:04:09 Training persons so make a caption Here like Let's Say caption

  • 00:04:12 like Let's Say caption and Let's edit it OK Let's Open

  • 00:04:17 and Let's edit it OK Let's Open with it and type ohw Man or woman or

  • 00:04:20 with it and type ohw Man or woman or Whatever i will explain All of this in

  • 00:04:23 Whatever i will explain All of this in the Other video in details Select so it

  • 00:04:27 the Other video in details Select so it will use this caption for All of The

  • 00:04:43 humans and i don't know is doing en ring of Sampa i don't

  • 00:04:46 of Sampa i don't suggest All of these options are Dis update

  • 00:04:50 suggest All of these options are Dis update will see the updated im

  • 00:05:07 1. Train ep it will repeat One Times these

  • 00:05:10 it will repeat One Times these Training

  • 00:05:12 Training images so it will be 15 steps with B

  • 00:05:16 images so it will be 15 steps with B size One for my Training images and Now

  • 00:05:18 size One for my Training images and Now i going to ADD the regularization images

  • 00:05:21 i going to ADD the regularization images so ADD Another Concept and

  • 00:05:24 so ADD Another Concept and İnel

  • 00:05:49 All L pat

  • 00:06:12 and edit it Forget and tye Man because these are all Man images that We are

  • 00:06:14 these are all Man images that We are using CL and Select see like selected

  • 00:06:21 using CL and Select see like selected and

  • 00:06:30 repeat Need div

  • 00:06:32 div 500 and put value Here What does this me

  • 00:06:38 500 and put value Here What does this me this means that in Every ep it will use

  • 00:06:41 this means that in Every ep it will use only 15 images of regularization images

  • 00:06:45 only 15 images of regularization images randomly i have tested the effect of

  • 00:06:48 randomly i have tested the effect of this it is shared on the patreon As Well

  • 00:06:50 this it is shared on the patreon As Well This is improving the Training As

  • 00:06:53 This is improving the Training As expected and image augmentation also

  • 00:06:56 expected and image augmentation also Close update preview and is working

  • 00:07:06 Training Concept

  • 00:07:09 Concept Training number of you to train for exle

  • 00:07:20 fores 150 working nice and

  • 00:07:32 Can extremely important Now they have ADD Save file prefix Which i love it i

  • 00:07:35 ADD Save file prefix Which i love it i told The developer To ADD so you can Save

  • 00:07:38 told The developer To ADD so you can Save After Every ep and Get checkpoint

  • 00:07:41 After Every ep and Get checkpoint and you can Say like My Training One so

  • 00:07:46 and you can Say like My Training One so this will be the saved file names

  • 00:07:48 this will be the saved file names prefixes they will be saved inside your

  • 00:07:51 prefixes they will be saved inside your workspace folder Wherever you

  • 00:07:53 workspace folder Wherever you have ok hopefully see you in the full

  • 00:07:56 have ok hopefully see you in the full tutorial later

Clone this wiki locally