we add "<segment:face,0.7>photo of" in the very end of our prompt?
we add "<segment:face,0.7>photo of" in the very end of our prompt?
Sto get a better quality face?
DYes, the face gets detected, segmented, and recreated, 'cuz you may not get high quality faces in distant (half/ full body) generations
DI find the results for this better than the yolo one.
Fye i plan that
Ftrue rtx 4090 cant run fp16 - not enough vram
Fit is. but gguf is i think more for lower vram much lower
Fi am waiting kohya for this
Fyes it inpaints face
Flike adetailer
Fyou have install error
Fyou probably didnt follow this video
FYouTubeSECourses
Welcome to the Pillars of AI Application Libraries: How to Install Tutorial! 
Are you ready to embark on a journey of installing AI libraries and applications with ease? 
In this video, we'll guide you through the process of installing Python, Git, Visual Studio C++ Compile tools, and FFmpeg on your Windows 10 machine. 
We'll also show you ho...
Are you ready to embark on a journey of installing AI libraries and applications with ease? In this video, we'll guide you through the process of installing Python, Git, Visual Studio C++ Compile tools, and FFmpeg on your Windows 10 machine. We'll also show you ho...F10 epoch
Fbut that lora training brought realims is not there yet
PDinos still need to get better .. 
.. awesome image
.. awesome imageF
Fwithout prompting it generates smiling
F
F20 epoch
Fbut i see the impact of poor lightning / focus of dataset
F
F
Fdmn model learnt even my broken teeth haah
Fi will now retrain batch size 1 with up to 30 epochs
Fafter 30 epochs it became cooked
Fmaybe maximum 40 epochs
Fit learnt my body shape too perfectly
F
FT5 attention mask + T5 XXL speed is 4.52 second / it on L40S - training everything
Dfor generation is Euler Simple still the best bet?
FiPNDM is best i think
Dsimple?
Dscheduler I mean
Fnormal - default - didnt change it actually
MFLUX LoRA Training Simplified: From Zero to Hero with Kohya SS GUI (8GB GPU, Windows) Tutorial Guide
do we have a video for this but using runpod ?
do we have a video for this but using runpod ?
Mi found already, nevermind
FYouTubeSECourses
Unlock the power of FLUX LoRA training, even if you're short on GPUs or looking to boost speed and scale! This comprehensive guide takes you from novice to expert, showing you how to use Kohya GUI for creating top-notch FLUX LoRAs in the cloud. We'll cover everything: maximizing quality, optimizing speed, and finding the best deals. With our exc...
Z@Dr. Furkan Gözükara Flux doesn't support negative prompts does it? i'm making batman but his nipples keep showing for some weird reason
Fit doesnt support by default true. i saw some things at some places but never tried they work or not and dont remembering it
Ftry to change prompt add more detailed clothing
Farmor
Zwhich clip and t5 models are best for text?
Zright now i'm using clip-l vit-l-14-text and t5xxl_enconly.safetensors and it's messing up text
Fi use fp16 t5 and default clip l for flux
FDI found it learnt a mfking mosquito in one photo lol
MHello. Which setting sets the clip l learning rate in the config? Text encoder learning rate?
Fyes text encoder learning rate
F
