Fix mixed precision fine-tuning for text-to-image-lora-sdxl example.

[sajadn]

What does this PR do?

Fixes #6442
Part of #6552

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline?
• Did you read our philosophy doc (important for complex PRs)?
• Was this discussed/approved via a GitHub issue or the forum? Please add a link to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

To Test?

training:

CUDA_VISIBLE_DEVICES=0 accelerate launch train_text_to_image_lora_sdxl.py \
  --pretrained_model_name_or_path="stabilityai/stable-diffusion-xl-base-1.0" \
  --pretrained_vae_model_name_or_path=madebyollin/sdxl-vae-fp16-fix \
  --dataset_name="lambdalabs/pokemon-blip-captions" \
  --mixed_precision="fp16" \
  --output_dir="text_to_image_example" \
  --resolution=1024 \
  --train_batch_size=1 \
  --gradient_accumulation_steps=4 --gradient_checkpointing \
  --learning_rate=1e-4 \
  --lr_scheduler="constant" \
  --lr_warmup_steps=0 \
  --max_train_steps=8 --checkpointing_steps=2 --checkpoints_total_limit=2 \
  --use_8bit_adam \
  --seed="0" 

resuming:

CUDA_VISIBLE_DEVICES=0 accelerate launch train_text_to_image_lora_sdxl.py \
  --pretrained_model_name_or_path="stabilityai/stable-diffusion-xl-base-1.0" \
  --pretrained_vae_model_name_or_path=madebyollin/sdxl-vae-fp16-fix \
  --dataset_name="lambdalabs/pokemon-blip-captions" \
  --mixed_precision="fp16" \
  --output_dir="text_to_image_example" \
  --resolution=1024 \
  --train_batch_size=1 \
  --gradient_accumulation_steps=4 --gradient_checkpointing \
  --learning_rate=1e-4 \
  --lr_scheduler="constant" \
  --lr_warmup_steps=0 \
  --max_train_steps=8 --checkpointing_steps=2 --checkpoints_total_limit=2 \
  --use_8bit_adam \
  --seed="0" \
  --resume_from_checkpoint="latest"

Who can review?

@sayakpaul

[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.

[sayakpaul]

Wow. I just ran it and it went smoothly. Thanks much!

[sayakpaul]

Shouldn't this be text_encoder_two_?

[sayakpaul]

Just a single comment.

[sajadn]

You're right! That was a terrible mistake! Review again please.

[sayakpaul]

Super!

[loboere]

I'm still getting the error ValueError: Attempting to unscale FP16 gradients.

Name: diffusers
Version: 0.26.2