GitLab

* disable peft input autocast

* use new peft method name; only disable peft input autocast if submodule layerwise casting active

* add test; reference PeftInputAutocastDisableHook in peft docs

* add load_lora_weights test

* casted -> cast

* Update tests/lora/utils.py

* update

* update

* make style

* remove dynamo disable

* add coauthor
Co-Authored-By: Dhruv Nair <dhruv.nair@gmail.com>

* update

* update

* update

* update mixin

* add some basic tests

* update

* update

* non_blocking

* improvements

* update

* norm.* -> norm

* apply suggestions from review

* add example

* update hook implementation to the latest changes from pyramid attention broadcast

* deinitialize should raise an error

* update doc page

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

* update docs

* update

* refactor

* fix _always_upcast_modules for asym ae and vq_model

* fix lumina embedding forward to not depend on weight dtype

* refactor tests

* add simple lora inference tests

* _always_upcast_modules -> _precision_sensitive_module_patterns

* remove todo comments about review; revert changes to self.dtype in unets because .dtype on ModelMixin should be able to handle fp8 weight case

* check layer dtypes in lora test

* fix UNet1DModelTests::test_layerwise_upcasting_inference

* _precision_sensitive_module_patterns -> _skip_layerwise_casting_patterns based on feedback

* skip test in NCSNppModelTests

* skip tests for AutoencoderTinyTests

* skip tests for AutoencoderOobleckTests

* skip tests for UNet1DModelTests - unsupported pytorch operations

* layerwise_upcasting -> layerwise_casting

* skip tests for UNetRLModelTests; needs next pytorch release for currently unimplemented operation support

* add layerwise fp8 pipeline test

* use xfail

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

* add assertion with fp32 comparison; add tolerance to fp8-fp32 vs fp32-fp32 comparison (required for a few models' test to pass)

* add note about memory consumption on tesla CI runner for failing test

---------
Co-authored-by: Dhruv Nair <dhruv.nair@gmail.com>
Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* print

* remove print.

* print

* update slice.

* empty

* allow big lora tests to run on the CI.

* print

* print.

* print

* print

* print

* print

* more

* print

* remove print.

* remove print

* directly place on cuda.

* remove pipeline.

* remove

* fix

* fix

* spaces

* quality

* updates

* directly place flux controlnet pipeline on cuda.

* torch_device instead of cuda.

* style

* device placement.

* fixes

* add big gpu marker for mochi; rename test correctly

* address feedback

* fix

---------
Co-authored-by: Aryan <aryan@huggingface.co>

* update

* fix make copies

* update

* add relevant markers to the integration test suite.

* add copied.

* fox-copies

* temporarily add print.

* directly place on CUDA as CPU isn't that big on the CIO.

* fixes to fuse_lora, aryan was right.

* fixes

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

* fix: lora unloading when using expanded Flux LoRAs.

* fix argument name.
Co-authored-by: a-r-r-o-w <contact.aryanvs@gmail.com>

* docs.

---------
Co-authored-by: a-r-r-o-w <contact.aryanvs@gmail.com>

add slow and nightly markers to sd3 lora integation.

* check for base_layer key in transformer state dict

* test_lora_expansion_works_for_absent_keys

* check

* Update tests/lora/test_lora_layers_flux.py
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* check

* test_lora_expansion_works_for_absent_keys/test_lora_expansion_works_for_extra_keys

* absent->extra

---------
Co-authored-by: hlky <hlky@hlky.ac>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* feat: support unload_lora_weights() for Flux Control.

* tighten test

* minor

* updates

* meta device fixes.

* update

* make style

* update

* update

* update

* make style

* single file related changes

* update

* fix

* update single file urls and docs

* update

* fix

* fixes to tests

* fixture

* fixes

updates

* misc lora test improvements.

* updates

* fixes to tests

* sana lora training tests and misc.

* remove push to hub

* Update examples/dreambooth/train_dreambooth_lora_sana.py
Co-authored-by: Aryan <aryan@huggingface.co>

---------
Co-authored-by: Aryan <aryan@huggingface.co>

add integration tests for lora expansion stuff in Flux.

* lora expansion with dummy zeros.

* updates

* fix working 🥳

* working.

* use torch.device meta for state dict expansion.

* tests
Co-authored-by: a-r-r-o-w <contact.aryanvs@gmail.com>

* fixes

* fixes

* switch to debug

* fix

* Apply suggestions from code review
Co-authored-by: Aryan <aryan@huggingface.co>

* fix stuff

* docs

---------
Co-authored-by: a-r-r-o-w <contact.aryanvs@gmail.com>
Co-authored-by: Aryan <aryan@huggingface.co>

fix

* 1217

* 1217

* 1217

* update

* reverse

* add test

* update test

* make style

* update

* make style

---------
Co-authored-by: Aryan <aryan@huggingface.co>

remove nullop imports

* feat: lora support for SANA.

* make fix-copies

* rename test class.

* attention_kwargs -> cross_attention_kwargs.

* Revert "attention_kwargs -> cross_attention_kwargs."

This reverts commit 23433bf9bccc12e0f2f55df26bae58a894e8b43b.

* exhaust 119 max line limit

* sana lora fine-tuning script.

* readme

* add a note about the supported models.

* Apply suggestions from code review
Co-authored-by: Aryan <aryan@huggingface.co>

* style

* docs for attention_kwargs.

* remove lora_scale from pag pipeline.

* copy fix

---------
Co-authored-by: Aryan <aryan@huggingface.co>

* add lora support for ltx

* add tests

* fix copied from comments

* update

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

* add test for expanding lora and normal lora error

* Update tests/lora/test_lora_layers_flux.py

* fix things.

* Update src/diffusers/loaders/peft.py

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

* add a test to ensure set_adapters() and attn kwargs outs match

* remove print

* fix

* Apply suggestions from code review
Co-authored-by: Benjamin Bossan <BenjaminBossan@users.noreply.github.com>

* assertFalse.

---------
Co-authored-by: Benjamin Bossan <BenjaminBossan@users.noreply.github.com>

* update


---------
Co-authored-by: yiyixuxu <yixu310@gmail.com>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* fix condition argument in xfail.

* revert init changes.

* skip nan lora tests on PyTorch 2.5.1 CPU.

* cog

* use xfail

* correct xfail

* add condition

* tests

* Update handle single blocks on _convert_xlabs_flux_lora_to_diffusers to fix bug on updating keys and old_state_dict


---------
Co-authored-by: raul_ar <raul.moreno.salinas@autoretouch.com>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* feat: add lora support to Mochi-1.

* feat: save_lora_adapter.

* fix test

* fix test asser

* fix format

* Update test_lora_layers_sd3.py

* add first draft.

* fix

* updates.

* updates.

* updates

* updates

* updates.

* fix-copies

* lora constants.

* add tests

* Apply suggestions from code review
Co-authored-by: Benjamin Bossan <BenjaminBossan@users.noreply.github.com>

* docstrings.

---------
Co-authored-by: Benjamin Bossan <BenjaminBossan@users.noreply.github.com>

* move lora integration tests to nightly./

* remove slow marker in the workflow where not needed.

* log a warning when there are missing keys in the LoRA loading.

* handle missing keys and unexpected keys better.

* add tests

* fix-copies.

* updates

* tests

* concat warning.

* Add Differential Diffusion to Kolors (#9423)

* Added diff diff support for kolors img2img

* Fized relative imports

* Fized relative imports

* Added diff diff support for Kolors

* Fized import issues

* Added map

* Fized import issues

* Fixed naming issues

* Added diffdiff support for Kolors img2img pipeline

* Removed example docstrings

* Added map input

* Updated latents
Co-authored-by: Álvaro Somoza <asomoza@users.noreply.github.com>

* Updated `original_with_noise`
Co-authored-by: Álvaro Somoza <asomoza@users.noreply.github.com>

* Improved code quality

---------
Co-authored-by: Álvaro Somoza <asomoza@users.noreply.github.com>

* FluxMultiControlNetModel (#9647)

* tests

* Update src/diffusers/loaders/lora_pipeline.py
Co-authored-by: YiYi Xu <yixu310@gmail.com>

* fix

---------
Co-authored-by: M Saqlain <118016760+saqlain2204@users.noreply.github.com>
Co-authored-by: Álvaro Somoza <asomoza@users.noreply.github.com>
Co-authored-by: hlky <hlky@hlky.ac>
Co-authored-by: YiYi Xu <yixu310@gmail.com>

* add lora

increase transformers version in test_low_cpu_mem_usage_with_loading

fix dora test.

* allow loras to be loaded with low_cpu_mem_usage.

* add flux support but note https://github.com/huggingface/diffusers/pull/9510\#issuecomment-2378316687



* low_cpu_mem_usage.

* fix-copies

* fix-copies again

* tests

* _LOW_CPU_MEM_USAGE_DEFAULT_LORA

* _peft_version default.

* version checks.

* version check.

* version check.

* version check.

* require peft 0.13.1.

* explicitly specify low_cpu_mem_usage=False.

* docs.

* transformers version 4.45.2.

* update

* fix

* empty

* better name initialize_dummy_state_dict.

* doc todos.

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

* style

* fix-copies

---------
Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* handle dora.

* print test

* debug

* fix

* fix-copies

* update logits

* add warning in the test.

* make is_dora check consistent.

* fix-copies

add a note on the versions.

* support kohya flux loras that have tes.