fixed bug in defining embed dim for UNet1D

[SammyAgrawal]

What does this PR do?

Fixes #12110

Hi! First time contributor to diffusers, apologies for anything I'm doing wrong in the PR. I noticed what I believe is a very simple and small bug in creating 1D UNets. A value is hardcoded that should not be. I noticed this while trying to debug why my UNet kept giving me shape errors. I just transitioned from using a 2d model to 1d for my personal use case, and this was causing me a headache! I am basically solving my own "Good First Issue" (linked above) noticed during my usage. I believe the Issue itself explains the problem and solution pretty straightforwardly.

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.

Tagging @sayakpaul @yiyixuxu @DN6

[SammyAgrawal]

I should note that although I tried to follow the contributing doc, there were parts that were hard for me to replicate / understand.
For example, should I not be trying to merge into main? I only saw that my fork branch should not be main.

Will happily go back and fix this before review if asked. For example, I ran python tests/models/unets/test_models_unet_1d.py but it failed with:

Traceback (most recent call last):
  File "/home/jovyan/Samarth/diffusers/tests/models/unets/test_models_unet_1d.py", line 29, in <module>
    from ..test_modeling_common import ModelTesterMixin, UNetTesterMixin
ImportError: attempted relative import with no known parent package

Which is unrelated to my modification. I think this means I set it up wrong?

[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.

[DN6]

I think it might be better to introduce an optional time_embedding_dim argument to the class init and configure this way

diffusers/src/diffusers/models/unets/unet_2d_condition.py

Line 549 in f8ba5cd

time_embed_dim = time_embedding_dim or block_out_channels[0] * 2

[SammyAgrawal]

@DN6 I updated to follow the Unet2d interface more closely, as you recommended.

[SammyAgrawal]

@DN6 Requesting re-review

[SammyAgrawal]

Bit confused what you mean - 0 or 5 would result in 5, just like None or 5. So if the user does not provide a truth-y value, it defaults to block out channels.

Regardless, accepted the commit as None behaves (I believe the same way) the right way

[DN6]

Thank you @SammyAgrawal 👍🏽