fix required instance_column and image_column cast

[davidberenstein1957]

What does this PR do?

• fix to force casting and image column to decode to ensure this is done for feature sets where this is not the default
• fix to allow for not passing an instance_prompt

Fixes # NA

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.

[davidberenstein1957]

cc @linoytsaban @sayakpaul, minor fixes. I can create separate issue to link them to if preferred.

[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]

Thanks @davidberenstein1957! re: instance_prompt we fall back to it in case captions are not provided / there's an issue with provided captions, so I think it's safer to keep it. Is there some motivation I'm missing for removing the requirement?

[sayakpaul]

Why is this needed?

[davidberenstein1957]

@sayakpaul fix to force casting and image column to decode to ensure this is done for feature sets where decoding automatically is not the default

[sayakpaul]

Okay let's have that as a comment too. Thanks for explaining!

[sayakpaul]

re: instance_prompt we fall back to it in case captions are not provided / there's an issue with provided captions, so I think it's safer to keep it. Is there some motivation I'm missing for removing the requirement?

I think okay if it's made a non-required argument, though.

[davidberenstein1957]

@sayakpaul @linoytsaban I just realised these fixes might also need to be applied to other dreambooth scripts?

[linoytsaban]

I just realised these fixes might also need to be applied to other dreambooth scripts?

@davidberenstein1957 yes :) this can apply to -

• dreambooth/train_dreambooth_lora_sdxl.py
• dreambooth/train_dreambooth_lora_sd3.py
• dreambooth/train_dreambooth_flux.py
• dreambooth/train_dreambooth_sdxl.py
• dreambooth/train_dreambooth_sd3.py
• advanced_diffusion_training/train_dreambooth_lora_sdxl_advanced.py
• advanced_diffusion_training/train_dreambooth_lora_flux_advanced.py
• advanced_diffusion_training/train_dreambooth_lora_sd15_advanced.py

[linoytsaban]

I think okay if it's made a non-required argument, though.

@sayakpaul @davidberenstein1957 I'm not opposed to it either, but it might be good to add a informative warning/throw an error in places we relied on instance_prompt being provided

[davidberenstein1957]

@linoytsaban will make sure to add some context if neither one of the required params is provided.

[sayakpaul]

Error sounds good. We could tackle this just for a single script and open the rest to the community. I don't think we need to be exhaustive at all.

[linoytsaban]

Error sounds good. We could tackle this just for a single script and open the rest to the community. I don't think we need to be exhaustive at all.

Yeah I agree

[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.

[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.