GitLab

* 7529 do not disable autocast for cuda devices

* Remove typecasting error check for non-mps platforms, as a correct autocast implementation makes it a non-issue

* add autocast fix to other training examples

* disable native_amp for dreambooth (sdxl)

* disable native_amp for pix2pix (sdxl)

* remove tests from remaining files

* disable native_amp on huggingface accelerator for every training example that uses it

* convert more usages of autocast to nullcontext, make style fixes

* make style fixes

* style.

* Empty-Commit

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

you cannot specify `type="bool"` and `action="store_true"` at the same time.
remove excessive and buggy `type=bool`.
Co-authored-by: Linoy Tsaban <57615435+linoytsaban@users.noreply.github.com>

* post-release

* quality

* add edm style training

* style

* finish adding edm training feature

* import fix

* fix latents mean

* minor adjustments

* add edm to readme

* style

* fix autocast and scheduler config issues when using edm

* style

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

switch to logger.warning

adding the type gives you

```
TypeError: _StoreTrueAction.__init__() got an unexpected keyword argument 'type'
```

* add tags for diffusers training

* add tags for diffusers training

* add tags for diffusers training

* add tags for diffusers training

* add tags for diffusers training

* add tags for diffusers training

* add dora tags for drambooth lora scripts

* style

* add is_dora arg

* style

* add dora training feature to sd 1.5 script

* added notes about DoRA training

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

* fix bug in micro-conditioning of class images

* fix bug in micro-conditioning of class images

* style

* post release

* style

* Empty-Commit

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

change to 2024

* add noise_offset param

* micro conditioning - wip

* image processing adjusted and moved to support micro conditioning

* change time ids to be computed inside train loop

* change time ids to be computed inside train loop

* change time ids to be computed inside train loop

* time ids shape fix

* move token replacement of validation prompt to the same section of instance prompt and class prompt

* add offset noise to sd15 advanced script

* fix token loading during validation

* fix token loading during validation in sdxl script

* a little clean

* style

* a little clean

* style

* sdxl script - a little clean + minor path fix

sd 1.5 script - change default resolution value

* ad 1.5 script - minor path fix

* fix missing comma in code example in model card

* clean up commented lines

* style

* remove time ids computed outside training loop - no longer used now that we utilize micro-conditioning, as all time ids are now computed inside the training loop

* style

* [WIP] - added draft readme, building off of examples/dreambooth/README.md

* readme

* readme

* readme

* readme

* readme

* readme

* readme

* readme

* removed --crops_coords_top_left from CLI args

* style

* fix missing shape bug due to missing RGB if statement

* add blog mention at the start of the reamde as well

* Update examples/advanced_diffusion_training/README.md
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* change note to render nicely as well

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

* sd1.5 support in separate script

A quick adaptation to support people interested in using this method on 1.5 models.

* sd15 prompt text encoding and unet conversions

as per @linoytsaban 's recommendations. Testing would be appreciated,

* Readability and quality improvements

Removed some mentions of SDXL, and some arguments that don't apply to sd 1.5, and cleaned up some comments.

* make style/quality commands

* tracker rename and run-it doc

* Update examples/advanced_diffusion_training/train_dreambooth_lora_sd15_advanced.py

* Update examples/advanced_diffusion_training/train_dreambooth_lora_sd15_advanced.py

---------
Co-authored-by: Linoy Tsaban <57615435+linoytsaban@users.noreply.github.com>

* fixes bugs:
1. redundant retraction
2. param clone
3. stopping optimization of text encoder params

* param upscaling

* style

resolve conflicts

* Fixes #6418 Advanced Dreambooth LoRa Training

* change order of import to fix nit

* fix nit, use cast_training_params

* remove torch.compile fix, will move to a new PR

* remove unnecessary import

* post release

* style

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

* unwrap text encoder when saving hook only for full text encoder tuning

* unwrap text encoder when saving hook only for full text encoder tuning

* save embeddings in each checkpoint as well

* save embeddings in each checkpoint as well

* save embeddings in each checkpoint as well

* Update examples/advanced_diffusion_training/train_dreambooth_lora_sdxl_advanced.py
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

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

[bug fix] using snr gamma and prior preservation loss in the dreambooth lora sdxl training scripts (#6356)

* change timesteps used to calculate snr when --with_prior_preservation is enabled

* change timesteps used to calculate snr when --with_prior_preservation is enabled (canonical script)

* style

* revert canonical script to before snr gamma change

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

* Add WebUI format support to Advanced Training Script

* style

---------
Co-authored-by: multimodalart <joaopaulo.passos+multimodal@gmail.com>

fix keys for lora format on advanced training scripts

* Fix ProdigyOPT in SDXL Dreambooth script

* style

* style

* Add PEFT to Advanced Training Script

* style

* style

* ✨ style ✨



* change order for logic operation

* add lora alpha

* style

* Align PEFT to new format

* Update train_dreambooth_lora_sdxl_advanced.py

Apply #6355 fix

---------
Co-authored-by: multimodalart <joaopaulo.passos+multimodal@gmail.com>

* add: script to train lcm lora for sdxl with 🤗

 datasets

* suit up the args.

* remove comments.

* fix num_update_steps

* fix batch unmarshalling

* fix num_update_steps_per_epoch

* fix; dataloading.

* fix microconditions.

* unconditional predictions debug

* fix batch size.

* no need to use use_auth_token

* Apply suggestions from code review
Co-authored-by: Suraj Patil <surajp815@gmail.com>

* make vae encoding batch size an arg

* final serialization in kohya

* style

* state dict rejigging

* feat: no separate teacher unet.

* debug

* fix state dict serialization

* debug

* debug

* debug

* remove prints.

* remove kohya utility and make style

* fix serialization

* fix

* add test

* add peft dependency.

* add: peft

* remove peft

* autocast device determination from accelerator

* autocast

* reduce lora rank.

* remove unneeded space

* Apply suggestions from code review
Co-authored-by: Suraj Patil <surajp815@gmail.com>

* style

* remove prompt dropout.

* also save in native diffusers ckpt format.

* debug

* debug

* debug

* better formation of the null embeddings.

* remove space.

* autocast fixes.

* autocast fix.

* hacky

* remove lora_sayak

* Apply suggestions from code review
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>

* style

* make log validation leaner.

* move back enabled in.

* fix: log_validation call.

* add: checkpointing tests

* taking my chances to see if disabling autocasting has any effect?

* start debugging

* name

* name

* name

* more debug

* more debug

* index

* remove index.

* print length

* print length

* print length

* move unet.train() after add_adapter()

* disable some prints.

* enable_adapters() manually.

* remove prints.

* some changes.

* fix params_to_optimize

* more fixes

* debug

* debug

* remove print

* disable grad for certain contexts.

* Add support for IPAdapterFull (#5911)

* Add support for IPAdapterFull
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

---------
Co-authored-by: YiYi Xu <yixu310@gmail.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Fix a bug in `add_noise` function  (#6085)

* fix

* copies

---------
Co-authored-by: yiyixuxu <yixu310@gmail,com>

* [Advanced Diffusion Script] Add Widget default text (#6100)

add widget

* [Advanced Training Script] Fix pipe example (#6106)

* IP-Adapter for StableDiffusionControlNetImg2ImgPipeline (#5901)

* adapter for StableDiffusionControlNetImg2ImgPipeline

* fix-copies

* fix-copies

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

* IP adapter support for most pipelines (#5900)

* support ip-adapter in src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py

* support ip-adapter in src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_attend_and_excite.py

* support ip-adapter in src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py

* update tests

* support ip-adapter in src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_panorama.py

* support ip-adapter in src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_sag.py

* support ip-adapter in src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py

* support ip-adapter in src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_text2img.py

* support ip-adapter in src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_img2img.py

* support ip-adapter in src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_ldm3d.py

* revert changes to sd_attend_and_excite and sd_upscale

* make style

* fix broken tests

* update ip-adapter implementation to latest

* apply suggestions from review

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

* fix: lora_alpha

* make vae casting conditional/

* param upcasting

* propagate comments from https://github.com/huggingface/diffusers/pull/6145

Co-authored-by: dg845 <dgu8957@gmail.com>

* [Peft] fix saving / loading when unet is not "unet" (#6046)

* [Peft] fix saving / loading when unet is not "unet"

* Update src/diffusers/loaders/lora.py
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* undo stablediffusion-xl changes

* use unet_name to get unet for lora helpers

* use unet_name

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

* [Wuerstchen] fix fp16 training and correct lora args (#6245)

fix fp16 training
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

* [docs] fix: animatediff docs (#6339)

fix: animatediff docs

* add: note about the new script in readme_sdxl.

* Revert "[Peft] fix saving / loading when unet is not "unet" (#6046)"

This reverts commit 4c7e983bb5929320bab08d70333eeb93f047de40.

* Revert "[Wuerstchen] fix fp16 training and correct lora args (#6245)"

This reverts commit 0bb9cf0216e501632677895de6574532092282b5.

* Revert "[docs] fix: animatediff docs (#6339)"

This reverts commit 11659a6f74b5187f601eeeeeb6f824dda73d0627.

* remove tokenize_prompt().

* assistive comments around enable_adapters() and diable_adapters().

---------
Co-authored-by: Suraj Patil <surajp815@gmail.com>
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: Fabio Rigano <57982783+fabiorigano@users.noreply.github.com>
Co-authored-by: YiYi Xu <yixu310@gmail.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: yiyixuxu <yixu310@gmail,com>
Co-authored-by: apolinário <joaopaulo.passos@gmail.com>
Co-authored-by: Charchit Sharma <charchitsharma11@gmail.com>
Co-authored-by: Aryan V S <contact.aryanvs@gmail.com>
Co-authored-by: dg845 <dgu8957@gmail.com>
Co-authored-by: Kashif Rasul <kashif.rasul@gmail.com>

[advanced dreambooth lora sdxl training script] load pipeline for inference only if validation prompt is used (#6171)

* load pipeline for inference only if validation prompt is used

* move things outside

* load pipeline for inference only if validation prompt is used

* fix readme when validation prompt is used

---------
Co-authored-by: linoytsaban <linoy@huggingface.co>
Co-authored-by: apolinário <joaopaulo.passos@gmail.com>

add widget

* add cache latents

* style

* Update train_dreambooth_lora_sdxl_advanced.py

* remove global function args from dreamboothdataset class

* style

* style

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

* improve help tags

* style fix

* changes token_abstraction type to string.
support multiple concepts for pivotal using a comma separated string.

* style fixup

* changed logger to warning (not yet available)

* moved the token_abstraction parsing to be in the same block as where we create the mapping of identifier to token

---------
Co-authored-by: Linoy <linoy@huggingface.co>

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

* improve help tags

* style fix

---------
Co-authored-by: Linoy <linoy@huggingface.co>

* Post Release: v0.24.0

* post pone deprecation

* post pone deprecation

* Add model_index.json

readme

[examples/advanced_diffusion_training] bug fixes and improvements for LoRA Dreambooth SDXL advanced training script (#5935)

* imports and readme bug fixes

* bug fix - ensures text_encoder params are dtype==float32 (when using pivotal tuning) even if the rest of the model is loaded in fp16

* added pivotal tuning to readme

* mapping token identifier to new inserted token in validation prompt (if used)

* correct default value of --train_text_encoder_frac

* change default value of  --adam_weight_decay_text_encoder

* validation prompt generations when using pivotal tuning bug fix

* style fix

* textual inversion embeddings name change

* style fix

* bug fix - stopping text encoder optimization halfway

* readme - will include token abstraction and new inserted tokens when using pivotal tuning
- added type to --num_new_tokens_per_abstraction

* style fix

---------
Co-authored-by: Linoy Tsaban <linoy@huggingface.co>

Adds an advanced version of the SD-XL DreamBooth LoRA training script supporting pivotal tuning  (#5883)

* sdxl dreambooth lora training script with pivotal tuning

* bug fix - args missing from parse_args

* code quality fixes

* comment unnecessary code from TokenEmbedding handler class

* fixup

---------
Co-authored-by: Linoy Tsaban <linoy@huggingface.co>