[Examples] fix: prior preservation setting in DreamBooth LoRA SDXL script.

[sayakpaul]

What does this PR do?

Fixes: #7239.

Example training command:

accelerate launch train_dreambooth_lora_sdxl.py \
  --pretrained_model_name_or_path="stabilityai/stable-diffusion-xl-base-1.0"  \
  --instance_data_dir="dog" \
  --pretrained_vae_model_name_or_path="madebyollin/sdxl-vae-fp16-fix" \
  --output_dir="dog-sdxl" \
  --mixed_precision="fp16" \
  --instance_prompt="a photo of sks dog" \
  --resolution=1024 \
  --train_batch_size=1 \
  --gradient_accumulation_steps=4 \
  --learning_rate=1e-4 \
  --report_to="wandb" \
  --lr_scheduler="constant" \
  --lr_warmup_steps=0 \
  --max_train_steps=500 \
  --validation_prompt="A photo of sks dog in a bucket" \
  --validation_epochs=25 \
  --seed="0" \
  --with_prior_preservation --class_data_dir="dog_class" --class_prompt="photo of a dog"

FYI: @shawnrushefsky.

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[linoytsaban]

Great! sorry for the bug 🐞

[Evan-Sukhoi]

The live doc in unavailible currently. Do you remember how it was fixed?

[sayakpaul]

https://github.com/huggingface/diffusers/pull/7242/files explains the changes.