Skip to content

GitLab

Switch branch/tag
  1. 28 May, 2025 1 commit
  3. 19 May, 2025 2 commits
  5. 02 May, 2025 1 commit
  7. 01 May, 2025 1 commit
  9. 15 Apr, 2025 1 commit
  11. 09 Apr, 2025 1 commit
  13. 08 Apr, 2025 1 commit
    • Benjamin Bossan's avatar
      [LoRA] Implement hot-swapping of LoRA (#9453) · fb544996
      Benjamin Bossan authored 
      * [WIP][LoRA] Implement hot-swapping of LoRA

This PR adds the possibility to hot-swap LoRA adapters. It is WIP.

Description

As of now, users can already load multiple LoRA adapters. They can
offload existing adapters or they can unload them (i.e. delete them).
However, they cannot "hotswap" adapters yet, i.e. substitute the weights
from one LoRA adapter with the weights of another, without the need to
create a separate LoRA adapter.

Generally, hot-swapping may not appear not super useful but when the
model is compiled, it is necessary to prevent recompilation. See #9279
for more context.

Caveats

To hot-swap a LoRA adapter for another, these two adapters should target
exactly the same layers and the "hyper-parameters" of the two adapters
should be identical. For instance, the LoRA alpha has to be the same:
Given that we keep the alpha from the first adapter, the LoRA scaling
would be incorrect for the second adapter otherwise.

Theoretically, we could override the scaling dict with the alpha values
derived from the second adapter's config, but changing the dict will
trigger a guard for recompilation, defeating the main purpose of the
feature.

I also found that compilation flags can have an impact on whether this
works or not. E.g. when passing "reduce-overhead", there will be errors
of the type:

> input name: arg861_1. data pointer changed from 139647332027392 to
139647331054592

I don't know enough about compilation to determine whether this is
problematic or not.

Current state

This is obviously WIP right now to collect feedback and discuss which
direction to take this. If this PR turns out to be useful, the
hot-swapping functions will be added to PEFT itself and can be imported
here (or there is a separate copy in diffusers to avoid the need for a
min PEFT version to use this feature).

Moreover, more tests need to be added to better cover this feature,
although we don't necessarily need tests for the hot-swapping
functionality itself, since those tests will be added to PEFT.

Furthermore, as of now, this is only implemented for the unet. Other
pipeline components have yet to implement this feature.

Finally, it should be properly documented.

I would like to collect feedback on the current state of the PR before
putting more time into finalizing it.

* Reviewer feedback

* Reviewer feedback, adjust test

* Fix, doc

* Make fix

* Fix for possible g++ error

* Add test for recompilation w/o hotswapping

* Make hotswap work

Requires https://github.com/huggingface/peft/pull/2366

More changes to make hotswapping work. Together with the mentioned PEFT
PR, the tests pass for me locally.

List of changes:

- docstring for hotswap
- remove code copied from PEFT, import from PEFT now
- adjustments to PeftAdapterMixin.load_lora_adapter (unfortunately, some
  state dict renaming was necessary, LMK if there is a better solution)
- adjustments to UNet2DConditionLoadersMixin._process_lora: LMK if this
  is even necessary or not, I'm unsure what the overall relationship is
  between this and PeftAdapterMixin.load_lora_adapter
- also in UNet2DConditionLoadersMixin._process_lora, I saw that there is
  no LoRA unloading when loading the adapter fails, so I added it
  there (in line with what happens in PeftAdapterMixin.load_lora_adapter)
- rewritten tests to avoid shelling out, make the test more precise by
  making sure that the outputs align, parametrize it
- also checked the pipeline code mentioned in this comment:
  https://github.com/huggingface/diffusers/pull/9453#issuecomment-2418508871;


  when running this inside the with
  torch._dynamo.config.patch(error_on_recompile=True) context, there is
  no error, so I think hotswapping is now working with pipelines.

* Address reviewer feedback:

- Revert deprecated method
- Fix PEFT doc link to main
- Don't use private function
- Clarify magic numbers
- Add pipeline test

Moreover:
- Extend docstrings
- Extend existing test for outputs != 0
- Extend existing test for wrong adapter name

* Change order of test decorators

parameterized.expand seems to ignore skip decorators if added in last
place (i.e. innermost decorator).

* Split model and pipeline tests

Also increase test coverage by also targeting conv2d layers (support of
which was added recently on the PEFT PR).

* Reviewer feedback: Move decorator to test classes

... instead of having them on each test method.

* Apply suggestions from code review
Co-authored-by: default avatarhlky <hlky@hlky.ac>

* Reviewer feedback: version check, TODO comment

* Add enable_lora_hotswap method

* Reviewer feedback: check _lora_loadable_modules

* Revert changes in unet.py

* Add possibility to ignore enabled at wrong time

* Fix docstrings

* Log possible PEFT error, test

* Raise helpful error if hotswap not supported

I.e. for the text encoder

* Formatting

* More linter

* More ruff

* Doc-builder complaint

* Update docstring:

- mention no text encoder support yet
- make it clear that LoRA is meant
- mention that same adapter name should be passed

* Fix error in docstring

* Update more methods with hotswap argument

- SDXL
- SD3
- Flux

No changes were made to load_lora_into_transformer.

* Add hotswap argument to load_lora_into_transformer

For SD3 and Flux. Use shorter docstring for brevity.

* Extend docstrings

* Add version guards to tests

* Formatting

* Fix LoRA loading call to add prefix=None

See:
https://github.com/huggingface/diffusers/pull/10187#issuecomment-2717571064



* Run make fix-copies

* Add hot swap documentation to the docs

* Apply suggestions from code review
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>

---------
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>
Co-authored-by: default avatarhlky <hlky@hlky.ac>
Co-authored-by: default avatarYiYi Xu <yixu310@gmail.com>
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>
      fb544996
  15. 02 Apr, 2025 1 commit
  17. 03 Mar, 2025 1 commit
  19. 26 Feb, 2025 1 commit
  21. 24 Feb, 2025 1 commit
  23. 12 Feb, 2025 1 commit
  25. 11 Feb, 2025 1 commit
  27. 03 Feb, 2025 2 commits
  29. 19 Jan, 2025 1 commit
  31. 14 Jan, 2025 1 commit
    • Marc Sun's avatar
      [FEAT] DDUF format (#10037) · fbff43ac
      Marc Sun authored 
      

* load and save dduf archive

* style

* switch to zip uncompressed

* updates

* Update src/diffusers/pipelines/pipeline_utils.py
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>

* Update src/diffusers/pipelines/pipeline_utils.py
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>

* first draft

* remove print

* switch to dduf_file for consistency

* switch to huggingface hub api

* fix log

* add a basic test

* Update src/diffusers/configuration_utils.py
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>

* Update src/diffusers/pipelines/pipeline_utils.py
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>

* Update src/diffusers/pipelines/pipeline_utils.py
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>

* fix

* fix variant

* change saving logic

* DDUF - Load transformers components manually (#10171)

* update hfh version

* Load transformers components manually

* load encoder from_pretrained with state_dict

* working version with transformers and tokenizer !

* add generation_config case

* fix tests

* remove saving for now

* typing

* need next version from transformers

* Update src/diffusers/configuration_utils.py
Co-authored-by: default avatarLucain <lucain@huggingface.co>

* check path corectly

* Apply suggestions from code review
Co-authored-by: default avatarLucain <lucain@huggingface.co>

* udapte

* typing

* remove check for subfolder

* quality

* revert setup changes

* oups

* more readable condition

* add loading from the hub test

* add basic docs.

* Apply suggestions from code review
Co-authored-by: default avatarLucain <lucain@huggingface.co>

* add example

* add

* make functions private

* Apply suggestions from code review
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>

* minor.

* fixes

* fix

* change the precdence of parameterized.

* error out when custom pipeline is passed with dduf_file.

* updates

* fix

* updates

* fixes

* updates

* fix xfail condition.

* fix xfail

* fixes

* sharded checkpoint compat

* add test for sharded checkpoint

* add suggestions

* Update src/diffusers/models/model_loading_utils.py
Co-authored-by: default avatarYiYi Xu <yixu310@gmail.com>

* from suggestions

* add class attributes to flag dduf tests

* last one

* fix logic

* remove comment

* revert changes

---------
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>
Co-authored-by: default avatarLucain <lucain@huggingface.co>
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>
Co-authored-by: default avatarYiYi Xu <yixu310@gmail.com>
      fbff43ac
  33. 13 Jan, 2025 1 commit
  35. 31 Dec, 2024 1 commit
  37. 05 Dec, 2024 1 commit
  39. 18 Nov, 2024 1 commit
  41. 16 Oct, 2024 1 commit
  43. 25 Sep, 2024 1 commit
  45. 16 Sep, 2024 1 commit
  47. 10 Aug, 2024 1 commit
  49. 05 Aug, 2024 2 commits
  51. 03 Aug, 2024 2 commits
  53. 02 Aug, 2024 1 commit
    • Sayak Paul's avatar
      [Core] Add PAG support for PixArtSigma (#8921) · 7b98c4cc
      Sayak Paul authored 
      * feat: add pixart sigma pag.

* inits.

* fixes

* fix

* remove print.

* copy paste methods to the pixart pag mixin

* fix-copies

* add documentation.

* add tests.

* remove correction file.

* remove pag_applied_layers

* empty
      7b98c4cc
  55. 26 Jul, 2024 1 commit
    • Sayak Paul's avatar
      [Chore] add `LoraLoaderMixin` to the inits (#8981) · d87fe95f
      Sayak Paul authored 
      

* introduce  to promote reusability.

* up

* add more tests

* up

* remove comments.

* fix fuse_nan test

* clarify the scope of fuse_lora and unfuse_lora

* remove space

* rewrite fuse_lora a bit.

* feedback

* copy over load_lora_into_text_encoder.

* address dhruv's feedback.

* fix-copies

* fix issubclass.

* num_fused_loras

* fix

* fix

* remove mapping

* up

* fix

* style

* fix-copies

* change to SD3TransformerLoRALoadersMixin

* Apply suggestions from code review
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>

* up

* handle wuerstchen

* up

* move lora to lora_pipeline.py

* up

* fix-copies

* fix documentation.

* comment set_adapters().

* fix-copies

* fix set_adapters() at the model level.

* fix?

* fix

* loraloadermixin.

---------
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>
      d87fe95f
  57. 25 Jul, 2024 2 commits
    • YiYi Xu's avatar
      Revert "[LoRA] introduce LoraBaseMixin to promote reusability." (#8976) · 62863bb1
      YiYi Xu authored 
      Revert "[LoRA] introduce LoraBaseMixin to promote reusability. (#8774)"

This reverts commit 527430d0.
      62863bb1
    • Sayak Paul's avatar
      [LoRA] introduce LoraBaseMixin to promote reusability. (#8774) · 527430d0
      Sayak Paul authored 
      

* introduce  to promote reusability.

* up

* add more tests

* up

* remove comments.

* fix fuse_nan test

* clarify the scope of fuse_lora and unfuse_lora

* remove space

* rewrite fuse_lora a bit.

* feedback

* copy over load_lora_into_text_encoder.

* address dhruv's feedback.

* fix-copies

* fix issubclass.

* num_fused_loras

* fix

* fix

* remove mapping

* up

* fix

* style

* fix-copies

* change to SD3TransformerLoRALoadersMixin

* Apply suggestions from code review
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>

* up

* handle wuerstchen

* up

* move lora to lora_pipeline.py

* up

* fix-copies

* fix documentation.

* comment set_adapters().

* fix-copies

* fix set_adapters() at the model level.

* fix?

* fix

---------
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>
      527430d0
  59. 22 Jul, 2024 2 commits
  61. 02 Jul, 2024 1 commit
  63. 01 Jul, 2024 1 commit
  65. 26 Jun, 2024 1 commit
  67. 25 Jun, 2024 1 commit