* feat: support dora loras from community
* safe-guard dora operations under peft version.
* pop use_dora when False
* make dora lora from kohya work.
* fix: kohya conversion utils.
* add a fast test for DoRA compatibility..
* add a nightly test.
* Added mark_step for sdxl to run with pytorch xla. Also updated README with instructions for xla
* adding soft dependency on torch_xla
* fix some styling
---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: Dhruv Nair <dhruv.nair@gmail.com>
* Add new text encoder
* add transformers depth
* More
* Correct conversion script
* Fix more
* Fix more
* Correct more
* correct text encoder
* Finish all
* proof that in works in run local xl
* clean up
* Get refiner to work
* Add red castle
* Fix batch size
* Improve pipelines more
* Finish text2image tests
* Add img2img test
* Fix more
* fix import
* Fix embeddings for classic models (#3888)
Fix embeddings for classic SD models.
* Allow multiple prompts to be passed to the refiner (#3895)
* finish more
* Apply suggestions from code review
* add watermarker
* Model offload (#3889)
* Model offload.
* Model offload for refiner / img2img
* Hardcode encoder offload on img2img vae encode
Saves some GPU RAM in img2img / refiner tasks so it remains below 8 GB.
---------
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* correct
* fix
* clean print
* Update install warning for `invisible-watermark`
* add: missing docstrings.
* fix and simplify the usage example in img2img.
* fix setup for watermarking.
* Revert "fix setup for watermarking."
This reverts commit 491bc9f5a6.
* fix: watermarking setup.
* fix: op.
* run make fix-copies.
* make sure tests pass
* improve convert
* make tests pass
* make tests pass
* better error message
* fiinsh
* finish
* Fix final test
---------
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
* add
* clean
* up
* clean up more
* fix more tests
* Improve docs further
* improve
* more fixes docs
* Improve docs more
* Update src/diffusers/models/unet_2d_condition.py
* fix
* up
* update doc links
* make fix-copies
* add safety checker and watermarker to stage 3 doc page code snippets
* speed optimizations docs
* memory optimization docs
* make style
* add watermarking snippets to doc string examples
* make style
* use pt_to_pil helper functions in doc strings
* skip mps tests
* Improve safety
* make style
* new logic
* fix
* fix bad onnx design
* make new stable diffusion upscale pipeline model arguments optional
* define has_nsfw_concept when non-pil output type
* lowercase linked to notebook name
---------
Co-authored-by: William Berman <WLBberman@gmail.com>
* [MS Text To Video} Add first text to video
* upload
* make first model example
* match unet3d params
* make sure weights are correcctly converted
* improve
* forward pass works, but diff result
* make forward work
* fix more
* finish
* refactor video output class.
* feat: add support for a video export utility.
* fix: opencv availability check.
* run make fix-copies.
* add: docs for the model components.
* add: standalone pipeline doc.
* edit docstring of the pipeline.
* add: right path to TransformerTempModel
* add: first set of tests.
* complete fast tests for text to video.
* fix bug
* up
* three fast tests failing.
* add: note on slow tests
* make work with all schedulers
* apply styling.
* add slow tests
* change file name
* update
* more correction
* more fixes
* finish
* up
* Apply suggestions from code review
* up
* finish
* make copies
* fix pipeline tests
* fix more tests
* Apply suggestions from code review
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
* apply suggestions
* up
* revert
---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
* add: support for BLIP generation.
* add: support for editing synthetic images.
* remove unnecessary comments.
* add inits and run make fix-copies.
* version change of diffusers.
* fix: condition for loading the captioner.
* default conditions_input_image to False.
* guidance_amount -> cross_attention_guidance_amount
* fix inputs to check_inputs()
* fix: attribute.
* fix: prepare_attention_mask() call.
* debugging.
* better placement of references.
* remove torch.no_grad() decorations.
* put torch.no_grad() context before the first denoising loop.
* detach() latents before decoding them.
* put deocding in a torch.no_grad() context.
* add reconstructed image for debugging.
* no_grad(0
* apply formatting.
* address one-off suggestions from the draft PR.
* back to torch.no_grad() and add more elaborate comments.
* refactor prepare_unet() per Patrick's suggestions.
* more elaborate description for .
* formatting.
* add docstrings to the methods specific to pix2pix zero.
* suspecting a redundant noise prediction.
* needed for gradient computation chain.
* less hacks.
* fix: attention mask handling within the processor.
* remove attention reference map computation.
* fix: cross attn args.
* fix: prcoessor.
* store attention maps.
* fix: attention processor.
* update docs and better treatment to xa args.
* update the final noise computation call.
* change xa args call.
* remove xa args option from the pipeline.
* add: docs.
* first test.
* fix: url call.
* fix: argument call.
* remove image conditioning for now.
* 🚨 add: fast tests.
* explicit placement of the xa attn weights.
* add: slow tests 🐢
* fix: tests.
* edited direction embedding should be on the same device as prompt_embeds.
* debugging message.
* debugging.
* add pix2pix zero pipeline for a non-deterministic test.
* debugging/
* remove debugging message.
* make caption generation _
* address comments (part I).
* address PR comments (part II)
* fix: DDPM test assertion.
* refactor doc.
* address PR comments (part III).
* fix: type annotation for the scheduler.
* apply styling.
* skip_mps and add note on embeddings in the docs.
* convert __main__ to a function call and call it
* add missing type hint
* make style check pass
* move loading to src/diffusers
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* [Lora] first upload
* add first lora version
* upload
* more
* first training
* up
* correct
* improve
* finish loaders and inference
* up
* up
* fix more
* up
* finish more
* finish more
* up
* up
* change year
* revert year change
* Change lines
* Add cloneofsimo as co-author.
Co-authored-by: Simo Ryu <cloneofsimo@gmail.com>
* finish
* fix docs
* Apply suggestions from code review
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Co-authored-by: Suraj Patil <surajp815@gmail.com>
* upload
* finish
Co-authored-by: Simo Ryu <cloneofsimo@gmail.com>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Co-authored-by: Suraj Patil <surajp815@gmail.com>
* add AudioDiffusionPipeline and LatentAudioDiffusionPipeline
* add docs to toc
* fix tests
* fix tests
* fix tests
* fix tests
* fix tests
* Update pr_tests.yml
Fix tests
* parent 499ff34b3e
author teticio <teticio@gmail.com> 1668765652 +0000
committer teticio <teticio@gmail.com> 1669041721 +0000
parent 499ff34b3e
author teticio <teticio@gmail.com> 1668765652 +0000
committer teticio <teticio@gmail.com> 1669041704 +0000
add colab notebook
[Flax] Fix loading scheduler from subfolder (#1319)
[FLAX] Fix loading scheduler from subfolder
Fix/Enable all schedulers for in-painting (#1331)
* inpaint fix k lms
* onnox as well
* up
Correct path to schedlure (#1322)
* [Examples] Correct path
* uP
Avoid nested fix-copies (#1332)
* Avoid nested `# Copied from` statements during `make fix-copies`
* style
Fix img2img speed with LMS-Discrete Scheduler (#896)
Casting `self.sigmas` into a different dtype (the one of original_samples) is not advisable. In my img2img pipeline this leads to a long running time in the `integrate.quad` call later on- by long I mean more than 10x slower.
Co-authored-by: Anton Lozhkov <anton@huggingface.co>
Fix the order of casts for onnx inpainting (#1338)
Legacy Inpainting Pipeline for Onnx Models (#1237)
* Add legacy inpainting pipeline compatibility for onnx
* remove commented out line
* Add onnx legacy inpainting test
* Fix slow decorators
* pep8 styling
* isort styling
* dummy object
* ordering consistency
* style
* docstring styles
* Refactor common prompt encoding pattern
* Update tests to permanent repository home
* support all available schedulers until ONNX IO binding is available
Co-authored-by: Anton Lozhkov <aglozhkov@gmail.com>
* updated styling from PR suggested feedback
Co-authored-by: Anton Lozhkov <aglozhkov@gmail.com>
Jax infer support negative prompt (#1337)
* support negative prompts in sd jax pipeline
* pass batched neg_prompt
* only encode when negative prompt is None
Co-authored-by: Juan Acevedo <jfacevedo@google.com>
Update README.md: Minor change to Imagic code snippet, missing dir error (#1347)
Minor change to Imagic Readme
Missing dir causes an error when running the example code.
make style
change the sample model (#1352)
* Update alt_diffusion.mdx
* Update alt_diffusion.mdx
Add bit diffusion [WIP] (#971)
* Create bit_diffusion.py
Bit diffusion based on the paper, arXiv:2208.04202, Chen2022AnalogBG
* adding bit diffusion to new branch
ran tests
* tests
* tests
* tests
* tests
* removed test folders + added to README
* Update README.md
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* move Mel to module in pipeline construction, make librosa optional
* fix imports
* fix copy & paste error in comment
* fix style
* add missing register_to_config
* fix class docstrings
* fix class docstrings
* tweak docstrings
* tweak docstrings
* update slow test
* put trailing commas back
* respect alphabetical order
* remove LatentAudioDiffusion, make vqvae optional
* move Mel from models back to pipelines :-)
* allow loading of pretrained audiodiffusion models
* fix tests
* fix dummies
* remove reference to latent_audio_diffusion in docs
* unused import
* inherit from SchedulerMixin to make loadable
* Apply suggestions from code review
* Apply suggestions from code review
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* [Proposal] Support loading from safetensors if file is present.
* Style.
* Fix.
* Adding some test to check loading logic.
+ modify download logic to not download pytorch file if not necessary.
* Fixing the logic.
* Adressing comments.
* factor out into a function.
* Remove dead function.
* Typo.
* Extra fetch only if safetensors is there.
* Apply suggestions from code review
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* 2x speedup using memory efficient attention
* remove einops dependency
* Swap K, M in op instantiation
* Simplify code, remove unnecessary maybe_init call and function, remove unused self.scale parameter
* make xformers a soft dependency
* remove one-liner functions
* change one letter variable to appropriate names
* Remove Env variable dependency, remove MemoryEfficientCrossAttention class and use enable_xformers_memory_efficient_attention method
* Add memory efficient attention toggle to img2img and inpaint pipelines
* Clearer management of xformers' availability
* update optimizations markdown to add info about memory efficient attention
* add benchmarks for TITAN RTX
* More detailed explanation of how the mem eff benchmark were ran
* Removing autocast from optimization markdown
* import_utils: import torch only if is available
Co-authored-by: Nouamane Tazi <nouamane98@gmail.com>
* add accelerate to load models with smaller memory footprint
* remove low_cpu_mem_usage as it is reduntant
* move accelerate init weights context to modelling utils
* add test to ensure results are the same when loading with accelerate
* add tests to ensure ram usage gets lower when using accelerate
* move accelerate logic to single snippet under modelling utils and remove it from configuration utils
* format code using to pass quality check
* fix imports with isor
* add accelerate to test extra deps
* only import accelerate if device_map is set to auto
* move accelerate availability check to diffusers import utils
* format code
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* Fix is_onnx_available
Fix: If user install onnxruntime-gpu, is_onnx_available() will return False.
* add more onnxruntime candidates
* Run `make style`
Co-authored-by: anton-l <anton@huggingface.co>
