Modularize Dreambooth LoRA SD inferencing during and after training

[rootonchair]

What does this PR do?

Part of #6545

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?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[yiyixuxu]

cc @sayakpaul
should we merge this in?

[sayakpaul]

Why do we need to initialize it here?

[rootonchair]

We don't. I will delete this line

[sayakpaul]

Looks very good. Thanks a lot for working on this.

Could you also provide some sample training commands with which we could test this?

[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.

[rootonchair]

Hi @sayakpaul, you could test it with the following line

accelerate launch examples/dreambooth/train_dreambooth_lora.py --pretrained_model_name_or_path="runwayml/stable-diffusion-v1-5" --instance_data_dir="dog" --output_dir=dog_lora --instance_prompt="a photo of sks dog" --train_batch_size=1 --resolution=512 --checkpointing_steps=100 --learning_rate=1e-4 --max_train_steps=500 --validation_prompt="a photo of sks dog in a bucket" --validation_epochs=50 --seed="0"

[sayakpaul]

Just one change requested. Thanks a lot for working on this!

Ran with your command:

accelerate launch train_dreambooth_lora.py --pretrained_model_name_or_path="runwayml/stable-diffusion-v1-5" --instance_data_dir="dog" --output_dir=dog_lora --instance_prompt="a photo of sks dog" --train_batch_size=1 --resolution=512 --checkpointing_steps=100 --learning_rate=1e-4 --max_train_steps=500 --validation_prompt="a photo of sks dog in a bucket" --validation_epochs=50 --seed="0" --report_to="wandb"

Working perfectly: https://wandb.ai/sayakpaul/dreambooth-lora/runs/ligw3fg4

[sayakpaul]

Thanks a lot for working on this!