Beautiful Doc string added into the UNetMidBlock2D class. #5389

hi-sushanta · 2023-10-13T16:03:59Z

Once more time I added a new doc string into the model file among the UNetMidBlock2D class.

Chaged Code Looks Like:

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.

…standing other developers what are doing and where it's using.

This changes suggest by maintener. Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

Add suggested text Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

I changed the Parameter to Args text.

proper indentation set in this file.

a little bit of change in the act_fun argument line.

Doc string

similar doc-string add to have in the original diffusion repository.

Added Beutifull doc-string into the UNetMidBlock2D class.

sayakpaul

Thank you!

Good for me to merge once the CI is green.

sayakpaul · 2023-10-17T06:14:26Z

src/diffusers/models/unet_2d_blocks.py

+        attn_groups (`Optional[int]`, *optional*, defaults to None): The number of groups for the attention blocks.
+        resnet_pre_norm (`bool`, *optional*, defaults to `True`): Whether to use pre-normalization for the resnet blocks.
+        add_attention (`bool`, *optional*, defaults to `True`): Whether to add attention blocks.
+        attention_head_dim (`int`, *optional*, defaults to 1): The number of attention heads.


Checking with @patrickvonplaten:

Is attention_head_dim related to the number of attention heads?

Yes, the attention_head_dim parameter is related to the number of attention heads

attention_head_dim is the dimension of a single attention head and is used to calculate the number of attention heads in the block.

diffusers/src/diffusers/models/unet_2d_blocks.py

Line 580 in 680e596

heads=in_channels // attention_head_dim,

So while it does contribute to determining the number of attention heads, it does not actually set the number of heads. I would change this to

attention_head_dim: Dimension of a single attention head. The number of attention heads is determined based on this value and the number of input channels.

Okay, now I also update the doc string for the attention_head_dim parameter.

DN6 · 2023-10-17T07:06:50Z

src/diffusers/models/unet_2d_blocks.py

+        resnet_eps (`float`, *optional*, 1e-6 ): The epsilon value for the resnet blocks.
+        resnet_time_scale_shift (`str`, *optional*, defaults to `default`): The time scale shift for the resnet blocks.
+        resnet_act_fn (`str`, *optional*, defaults to `swish`): The activation function for the resnet blocks.
+        resnet_groups (`int`, *optional*, defaults to 32): The number of groups for the resnet blocks.


Could we maybe add some context on how the parameters are applied in the blocks?

e.g.
resnet_groups: The number of groups to use in the group normalization layers of the resnet.
resnet_timescale_shift: The type of normalization to apply to the time embeddings.

Okay, I understand, Now I added a new one that looks like that if it's OK.

resnet_groups: The number of groups to use in the group normalization layers of the resnet blocks. This can help to improve performance and reduce memory usage.

resnet_time_scale_shift: The type of normalization to apply to the time embeddings. This can help to improve the performance of the model on tasks with long-range temporal dependencies.

Hi @hi-sushanta. I think the phrase "improve performance of the model" is dependent on a number of factors, and reducing the number of groups doesn't necessarily lead to memory savings. So I would remove those additional sentences.

Okay, I am done to remove additional sentences.

I replaced the definition in this parameter resnet_time_scale_shift and resnet_groups.

I remove additional sentences into the resnet_groups argument.

I replaced my definition with the maintainer definition in the attention_head_dim parameter.

sayakpaul

Thank you for working on it!

DN6 · 2023-10-18T13:02:37Z

Hi @hi-sushanta Can you please run make style && make quality on your PR so that the code quality checks pass.

hi-sushanta · 2023-10-18T15:33:55Z

I already successfully reformatted the file using the black package. However, when I ran the make style && make quality commands, multiple other files were reformatted, even though I did not touch those files.

…e#5389) * I added a new doc string to the class. This is more flexible to understanding other developers what are doing and where it's using. * Update src/diffusers/models/unet_2d_blocks.py This changes suggest by maintener. Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * Update src/diffusers/models/unet_2d_blocks.py Add suggested text Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * Update unet_2d_blocks.py I changed the Parameter to Args text. * Update unet_2d_blocks.py proper indentation set in this file. * Update unet_2d_blocks.py a little bit of change in the act_fun argument line. * I run the black command to reformat style in the code * Update unet_2d_blocks.py similar doc-string add to have in the original diffusion repository. * Update unet_2d_blocks.py Added Beutifull doc-string into the UNetMidBlock2D class. * Update unet_2d_blocks.py I replaced the definition in this parameter resnet_time_scale_shift and resnet_groups. * Update unet_2d_blocks.py I remove additional sentences into the resnet_groups argument. * Update unet_2d_blocks.py I replaced my definition with the maintainer definition in the attention_head_dim parameter. * I am using black package for reformated my file --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> Co-authored-by: YiYi Xu <yixu310@gmail.com>

hi-sushanta and others added 17 commits October 5, 2023 06:34

I added a new doc string to the class. This is more flexible to under…

8425cd4

…standing other developers what are doing and where it's using.

Merge branch 'main' into doc_string

8a0e77d

Merge branch 'main' into doc_string

cf3a816

Merge branch 'main' into doc_string

a0cd96f

Update src/diffusers/models/unet_2d_blocks.py

746a8e8

This changes suggest by maintener. Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

Update src/diffusers/models/unet_2d_blocks.py

6e56886

Add suggested text Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

Update unet_2d_blocks.py

627fd9f

I changed the Parameter to Args text.

Update unet_2d_blocks.py

ae4f7f2

proper indentation set in this file.

Update unet_2d_blocks.py

f0bea43

a little bit of change in the act_fun argument line.

Merge branch 'main' into doc_string

872a4a5

I run the black command to reformat style in the code

3546f6d

Merge branch 'main' into doc_string

12534f4

Merge branch 'main' into doc_string

48afb4b

Merge pull request #1 from hi-sushanta/doc_string

0bb06b6

Doc string

Update unet_2d_blocks.py

01a9fc9

similar doc-string add to have in the original diffusion repository.

Merge branch 'huggingface:main' into main

04e6efb

Update unet_2d_blocks.py

eb9611d

Added Beutifull doc-string into the UNetMidBlock2D class.

hi-sushanta changed the title ~~Doc string~~ Beautiful Doc string added into the UNetMidBlock2D class. Oct 15, 2023

Merge branch 'main' into doc-string

298881a

patrickvonplaten requested review from DN6, yiyixuxu and sayakpaul October 16, 2023 13:30

Merge branch 'main' into doc-string

73dcca6

sayakpaul approved these changes Oct 17, 2023

View reviewed changes

DN6 reviewed Oct 17, 2023

View reviewed changes

hi-sushanta added 4 commits October 17, 2023 17:47

Update unet_2d_blocks.py

680e596

I replaced the definition in this parameter resnet_time_scale_shift and resnet_groups.

Update unet_2d_blocks.py

5db0ce9

I remove additional sentences into the resnet_groups argument.

Update unet_2d_blocks.py

ea792c9

I replaced my definition with the maintainer definition in the attention_head_dim parameter.

Merge branch 'main' into doc-string

af87b15

sayakpaul approved these changes Oct 18, 2023

View reviewed changes

sayakpaul requested a review from DN6 October 18, 2023 03:42

I am using black package for reformated my file

87a74ce

DN6 approved these changes Oct 18, 2023

View reviewed changes

Merge branch 'main' into doc-string

f99aac0

patrickvonplaten merged commit 36a0bac into huggingface:main Oct 18, 2023
10 of 11 checks passed

hi-sushanta deleted the doc-string branch October 19, 2023 01:06

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Beautiful Doc string added into the UNetMidBlock2D class. #5389

Beautiful Doc string added into the UNetMidBlock2D class. #5389

hi-sushanta commented Oct 13, 2023 •

edited

Loading

sayakpaul left a comment

sayakpaul Oct 17, 2023

hi-sushanta Oct 17, 2023

DN6 Oct 17, 2023 •

edited

Loading

hi-sushanta Oct 17, 2023

DN6 Oct 17, 2023 •

edited

Loading

hi-sushanta Oct 17, 2023

DN6 Oct 17, 2023

hi-sushanta Oct 17, 2023

sayakpaul left a comment

DN6 commented Oct 18, 2023

hi-sushanta commented Oct 18, 2023

Beautiful Doc string added into the UNetMidBlock2D class. #5389

Beautiful Doc string added into the UNetMidBlock2D class. #5389

Conversation

hi-sushanta commented Oct 13, 2023 • edited Loading

Before submitting

Who can review?

sayakpaul left a comment

Choose a reason for hiding this comment

sayakpaul Oct 17, 2023

Choose a reason for hiding this comment

hi-sushanta Oct 17, 2023

Choose a reason for hiding this comment

DN6 Oct 17, 2023 • edited Loading

Choose a reason for hiding this comment

hi-sushanta Oct 17, 2023

Choose a reason for hiding this comment

DN6 Oct 17, 2023 • edited Loading

Choose a reason for hiding this comment

hi-sushanta Oct 17, 2023

Choose a reason for hiding this comment

DN6 Oct 17, 2023

Choose a reason for hiding this comment

hi-sushanta Oct 17, 2023

Choose a reason for hiding this comment

sayakpaul left a comment

Choose a reason for hiding this comment

DN6 commented Oct 18, 2023

hi-sushanta commented Oct 18, 2023

hi-sushanta commented Oct 13, 2023 •

edited

Loading

DN6 Oct 17, 2023 •

edited

Loading

DN6 Oct 17, 2023 •

edited

Loading