I didn't go through it comprehensively, but I did look at it
I didn't go through it comprehensively, but I did look at it
Fso dataset matters most
Fand consistency
Fif not consistent captions help
SI used the default of 10 in flux gym, and 1 in Kohya
VI think that is a major difference, not enough time to learn the style. I have noticed this myself. The major drawback about fine tuning for styles, I think, is that it needs more image repeats, but I haven’t done enough testing yet. Was the dataset 1:1?
Syes, the data set was the same. I'll give it another try with higher repeats in kohya and see if it's the difference.
SThe default epochs in gym is 16 -- is that how many epochs I would put into Koyha to try and replicate results? It seems low
Fdefault number of epochs really meaningless
Fit depends on each dataset
Fhow many images you have total?
S21, but the style is very consistent. I could add more but I'm just trying to get a feel for the settings
VYou really need to look at the training log in flux gym, all the settings are different. I believe it uses Adam8bit, and a slightly faster learning rate. I also think training is in fp8.
VSare you referring to the train.bat file?
VI think that’s it. If you right click and open in a text program you can see the settings.
VDid you assign a trigger word during training/captioning and use it in both images when testing?
Fyou should add more
Fwe cant train fp8
Fall trainings are done at bf16
Fkohya doesnt know how to do fp8 training :d
Wnew Kohya with Torch 2.5 speed increase about 20% - 25%
W
WI used 3090
Famazing
Fi dont have yet. our kohya configs are transferable to onetrainer
Mwhere can i find that khoya configs
F
Full Instructions, Configs, Installers, Information and Links Shared Post (the one used in the tutorial) 
https://www.patreon.com/posts/112099700Patreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
FYouTubeSECourses
If you want to train FLUX with maximum possible quality, this is the tutorial looking for. In this comprehensive tutorial, you will learn how to install Kohya GUI and use it to fully Fine-Tune / DreamBooth FLUX model. After that how to use SwarmUI to compare generated checkpoints / models and find the very best one to generate most amazing image...
Fit has both lora and fine tuning configs
Mthank you Dr.
V
V@Saxon The training parameters are all here. Just match them in kohya. The lora network dim was trained at rank 4 btw.
accelerate launch ^
--mixed_precision bf16 ^
--num_cpu_threads_per_process 1 ^
sd-scripts/flux_train_network.py ^
--pretrained_model_name_or_path "C:\pinokio\api\fluxgym.git\models\unet\flux1-dev.sft" ^
--clip_l "C:\pinokio\api\fluxgym.git\models\clip\clip_l.safetensors" ^
--t5xxl "C:\pinokio\api\fluxgym.git\models\clip\t5xxl_fp16.safetensors" ^
--ae "C:\pinokio\api\fluxgym.git\models\vae\ae.sft" ^
--cache_latents_to_disk ^
--save_model_as safetensors ^
--sdpa --persistent_data_loader_workers ^
--max_data_loader_n_workers 2 ^
--seed 42 ^
--gradient_checkpointing ^
--mixed_precision bf16 ^
--save_precision bf16 ^
--network_module networks.lora_flux ^
--network_dim 4 ^
--optimizer_type adafactor ^
--optimizer_args "relative_step=False" "scale_parameter=False" "warmup_init=False" ^
--split_mode ^
--network_args "train_blocks=single" ^
--lr_scheduler constant_with_warmup ^
--max_grad_norm 0.0 ^--sample_prompts="C:\pinokio\api\fluxgym.git\outputs\k44\sample_prompts.txt" --sample_every_n_steps="500" ^
--learning_rate 8e-4 ^
--cache_text_encoder_outputs ^
--cache_text_encoder_outputs_to_disk ^
--fp8_base ^
--highvram ^
--max_train_epochs 16 ^
--save_every_n_epochs 4 ^
--dataset_config "C:\pinokio\api\fluxgym.git\outputs\k44\dataset.toml" ^
--output_dir "C:\pinokio\api\fluxgym.git\outputs\k44" ^
--output_name k44 ^
--timestep_sampling shift ^
--discrete_flow_shift 3.1582 ^
--model_prediction_type raw ^
--guidance_scale 1 ^
--loss_type l2 ^
accelerate launch ^
--mixed_precision bf16 ^
--num_cpu_threads_per_process 1 ^
sd-scripts/flux_train_network.py ^
--pretrained_model_name_or_path "C:\pinokio\api\fluxgym.git\models\unet\flux1-dev.sft" ^
--clip_l "C:\pinokio\api\fluxgym.git\models\clip\clip_l.safetensors" ^
--t5xxl "C:\pinokio\api\fluxgym.git\models\clip\t5xxl_fp16.safetensors" ^
--ae "C:\pinokio\api\fluxgym.git\models\vae\ae.sft" ^
--cache_latents_to_disk ^
--save_model_as safetensors ^
--sdpa --persistent_data_loader_workers ^
--max_data_loader_n_workers 2 ^
--seed 42 ^
--gradient_checkpointing ^
--mixed_precision bf16 ^
--save_precision bf16 ^
--network_module networks.lora_flux ^
--network_dim 4 ^
--optimizer_type adafactor ^
--optimizer_args "relative_step=False" "scale_parameter=False" "warmup_init=False" ^
--split_mode ^
--network_args "train_blocks=single" ^
--lr_scheduler constant_with_warmup ^
--max_grad_norm 0.0 ^--sample_prompts="C:\pinokio\api\fluxgym.git\outputs\k44\sample_prompts.txt" --sample_every_n_steps="500" ^
--learning_rate 8e-4 ^
--cache_text_encoder_outputs ^
--cache_text_encoder_outputs_to_disk ^
--fp8_base ^
--highvram ^
--max_train_epochs 16 ^
--save_every_n_epochs 4 ^
--dataset_config "C:\pinokio\api\fluxgym.git\outputs\k44\dataset.toml" ^
--output_dir "C:\pinokio\api\fluxgym.git\outputs\k44" ^
--output_name k44 ^
--timestep_sampling shift ^
--discrete_flow_shift 3.1582 ^
--model_prediction_type raw ^
--guidance_scale 1 ^
--loss_type l2 ^
Fhe is actively working on but i havent tested yet
SI went through and changed all the settings I could find, but I wasn't sure how to change the network_dim or network_args. I'm also not clear on the epochs, so I just set it to 16 because I'm not sure what the distinction is between epochs and max_train_epochs.
VI believe network dim id lora rank and alpha, which should be changed from 128 to what you see there. Epoch is like a turn, max epoch is max turns you want. The kohya_ss GitHub wiki tells a bit more about the settings.
VFloras yes
Fi dont know base models
Si see what it was -- I was matching the settings to the dreambooth tab -- network dim is in the lora tab. Is there a reason that I'm too naive to know that these settings would not work well in dreambooth?
Mhey everyone, when i am finetuning stable diffusionXL 1.0 base on a woman's picture. why does the base model always have that indian clothing style. so the problem is that when i am tuninng my model, the model can only perform good while i generate a picture of the character i have trained in indian clothes. Need help here
Jwhat token are you using for the woman? could be the token alone is related somehow to indian clothing styles
Fyou should use some good fine tuned model of sdxl
Fas a base
MRealVisXL_v4 , this is the model i am using as base model
Mi will try to change that
Fthere is v5