GitLab

change to 2024

add ipo and hinge loss to dpo trainer

* Update train_diffusion_dpo.py

Address #6702

* Update train_diffusion_dpo_sdxl.py

* Empty-Commit

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

* fix minsnr implementation for v-prediction case

* format code

* always compute snr when snr_gamma is specified

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

* initial commit for unconditional/class-conditional consistency training script

* make style

* Add entry for consistency training script in community README.

* Move consistency training script from community to research_projects/consistency_training

* Add requirements.txt and README to research_projects/consistency_training directory.

* Manually revert community README changes for consistency training.

* Fix path to script after moving script to research projects.

* Add option to load U-Net weights from pretrained model.

---------
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Update pipeline_controlnet_sd_xl.py

* Update pipeline_controlnet_xs_sd_xl.py

* move unets to  module 🦋

* parameterize unet-level import.

* fix flax unet2dcondition model import

* models __init__

* mildly depcrecating models.unet_2d_blocks in favor of models.unets.unet_2d_blocks.

* noqa

* correct depcrecation behaviour

* inherit from the actual classes.

* Empty-Commit

* backwards compatibility for unet_2d.py

* backward compatibility for unet_2d_condition

* bc for unet_1d

* bc for unet_1d_blocks

apply fixes from #6547

Bumps [jinja2](https://github.com/pallets/jinja) from 3.1.2 to 3.1.3.
- [Release notes](https://github.com/pallets/jinja/releases)
- [Changelog](https://github.com/pallets/jinja/blob/main/CHANGES.rst)
- [Commits](https://github.com/pallets/jinja/compare/3.1.2...3.1.3

)

---
updated-dependencies:
- dependency-name: jinja2
  dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

pass tracker name as argumentw

* base template file - train_instruct_pix2pix.py

* additional import and parser argument requried for lora

* finetune only instructpix2pix model -- no need to include these layers

* inject lora layers

* freeze unet model -- only lora layers are trained

* training modifications to train only lora parameters

* store only lora parameters

* move train script to research project

* run quality and style code checks

* move train script to a new folder

* add README

* update README

* update references in README

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

* add: experimental script for diffusion dpo training.

* random_crop cli.

* fix: caption tokenization.

* fix: pixel_values index.

* fix: grad?

* debug

* fix: reduction.

* fixes in the loss calculation.

* style

* fix: unwrap call.

* fix: validation inference.

* add: initial sdxl script

* debug

* make sure images in the tuple are of same res

* fix model_max_length

* report print

* boom

* fix: numerical issues.

* fix: resolution

* comment about resize.

* change the order of the training transformation.

* save call.

* debug

* remove print

* manually detaching necessary?

* use the same vae for validation.

* add: readme.

* handle rest of the stuff related to deprecated lora stuff.

* fix: copies

* don't modify the uNet in-place.

* fix: temporal autoencoder.

* manually remove lora layers.

* don't copy unet.

* alright

* remove lora attn processors from unet3d

* fix: unet3d.

* styl

* Empty-Commit

* files added

* fixing code quality

* fixing code quality

* fixing code quality

* fixing code quality

* sorted import block

* seperated import wandb

* ruff on script

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

* update

* update

* update

* update

* update

* make style

* remove docs

* update

* move to research folder.

* fix-copies

* remove _toctree entry.

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

Bump transformers in /examples/research_projects/realfill

Bumps [transformers](https://github.com/huggingface/transformers) from 4.34.0 to 4.36.0.
- [Release notes](https://github.com/huggingface/transformers/releases)
- [Commits](https://github.com/huggingface/transformers/compare/v4.34.0...v4.36.0

)

---
updated-dependencies:
- dependency-name: transformers
  dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Fix typos and trim trailing whitespaces

* Harmonize HF environment variables + deprecate use_auth_token

* fix import

* fix

* ruff format

* not need to use doc-builder's black styling as the doc is styled in ruff

* make fix-copies

* comment

* use run_ruff

* Set usedforsecurity=False in hashlib methods (FIPS compliance)

* update version dependency

* bump hfh version

* bump hfh version

Correct code for distributed training

* Add realfill

* Move realfill folder

* Fix some format issues

* update training examples to use HFAPI.

* update training example.

* reflect the changes in the korean version too.

* Empty-Commit

* chore: fix typos

* Update src/diffusers/pipelines/shap_e/renderer.py
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>

---------
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>

Min-SNR Gamma: correct the fix for SNR weighted loss in v-prediction by adding 1 to SNR rather than the resulting loss weights
Co-authored-by: bghira <bghira@users.github.com>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* [SDXL Flax] Add research folder

* Add co-author
Co-authored-by: Juan Acevedo <jfacevedo@google.com>

---------
Co-authored-by: Juan Acevedo <jfacevedo@google.com>

* fix all

* make fix copies

* make fix copies

add compute_snr() to training utils.

* merge with main

* fix flax example

* fix onnx example

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

Co-authored-by: bghira <bghira@users.github.com>

* Resetting rdm pr

* Fixed styles

* Fixed style

* Moved to rdm folder+fixed slight errors

* Removed config diff

* Started adding tests

* Adding retrieved images

* Fixed faiss import

* Fixed import errors

* Fixing tests

* Added require_faiss

* Updated dependency table

* Attempt solving consistency test

* Fixed truncation and vocab size issue

* Passed common tests

* Finished up cpu testing on pipeline

* Passed all tests locally

* Removed some slow tests

* Removed diffs from test_pipeline_common

* Remove logs

* Removed diffs from test_pipelines_common

* Fixed style

* Fully fixed styles on diffs

* Fixed name

* Proper rename

* Fixed dummies

* Fixed issue with dummyonnx

* Fixed black style

* Fixed dummies

* Changed ordering

* Fixed logging

* Fixing

* Fixing

* quality

* Debugging regex

* Fix dummies with guess

* Fixed typo

* Attempt fix dummies

* black

* ruff

* fixed ordering

* Logging

* Attempt fix

* Attempt fix dummy

* Attempt fixing styles

* Fixed faiss dependency

* Removed unnecessary deprecations

* Finished up main changes

* Added doc

* Passed tests

* Fixed tests

* Remove invisible watermark

* Fixed ruff errors

* Added prompt embed to tests

* Added tests and made retriever an optional component

* Fixed styles

* Made faiss a dependency of pipeline

* Logging

* Fixed dummies

* Make pipeline test work

* Fixed style

* Moved to research projects

* Remove diff

* Fixed style error

---------
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* push to hub docs

* fix typo

* feedback

* make style

add controlent script,

* add: utility for image grid.

* add: return type.

* change necessary places.

* add to utility page.

* Multiply lr scheduler steps by `num_processes`.

* Stop multiplying steps by gradient accumulation.