这玩意不太好改，得重新训练模型了

對西方人來說，龍的確是偏負面的，不過算了實用就行

of course.

```accelerate launch --num_cpu_threads_per_process=2 "train_network.py" --pretrained_model_name_or_path="D:/stable-diffusion-webui/models/Stable-diffusion/v1-5-pruned-emaonly.safetensors" --train_data_dir="D:\train\longpao\v1" --resolution=512,768 --output_dir="D:\train\longpao\models" --logging_dir="D:\train\longpao\log" --network_alpha="128" --save_model_as=safetensors --network_module=networks.lora --text_encoder_lr=5e-5 --unet_lr=0.0001 --network_dim=128 --output_name="dragon_gown_v1" --lr_scheduler_num_cycles="1" --learning_rate="0.0001" --lr_scheduler="constant" --train_batch_size="2" --max_train_steps="2000" --save_every_n_epochs="2" --mixed_precision="bf16" --save_precision="bf16" --seed="1234" --caption_extension=".txt" --cache_latents --optimizer_type="AdamW8bit" --max_data_loader_n_workers="1" --clip_skip=2 --bucket_reso_steps=64 --xformers --bucket_no_upscale```

Thank you so 🥰, I'm gonna try it, Sorry for many questions, can u also share number of repeatings and epoch also, Thanks

@Astro9 16 repeatings，19 images and 14 epochs