add --network_train_unet_only option

[knishika62]

I don't know why, but adding --network_train_unet_only makes the faces more similar.

[marhensa]

I don't know why, but adding --network_train_unet_only makes the faces more similar.

where do you add it?

[knishika62]

where do you add it?

Advanced options (default OFF)

[marhensa]

where do you add it?

Advanced options (default OFF)

like add it on the run arguments?

python app.py --network_train_unet_only or where to put it on Advanced Options?

because I don't see where to put it on Advanced Options GUI

[knishika62]

because I don't see where to put it on Advanced Options GUI

Put it in the advanced options GUI. Is it the lowest? like a this.

network_train_unet_only [pulldown] off(default) | on

If it is on, add this option to cli.

accelerate launch --mixed_precision bf16 --num_cpu_threads_per_process 1 sd-scripts/flux_train_network.py --pretrained_model_name_or_path "D:\fluxgym\models\unet\flux1-dev.sft" --clip_l "D:\fluxgym\models\clip\clip_l.safetensors" --t5xxl "D:\fluxgym\models\clip\t5xxl_fp16.safetensors" --ae "D:\fluxgym\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 adamw8bit --learning_rate 8e-4 --cache_text_encoder_outputs --cache_text_encoder_outputs_to_disk --fp8_base --highvram --max_train_epochs 15 --save_every_n_epochs 5 --dataset_config "D:\fluxgym\dataset.toml" --output_dir "D:\fluxgym\outputs" --output_name test-8e4-dm4 --timestep_sampling shift --discrete_flow_shift 3.1582 --model_prediction_type raw --guidance_scale 1 --loss_type l2 --network_train_unet_only