* Update pipeline_stable_diffusion_instruct_pix2pix.py
Add `cross_attention_kwargs` to `__call__` method of `StableDiffusionInstructPix2PixPipeline`, which are passed to UNet.
* Update documentation for pipeline_stable_diffusion_instruct_pix2pix.py
* Update docstring
* Update docstring
* Fix typing import
* make _callback_tensor_inputs consistent between sdxl pipelines
* forgot this one
* fix failing test
* fix test_components_function
* fix controlnet inpaint tests
* Merged isinstance calls to make the code simpler.
* Corrected formatting errors using ruff.
---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: YiYi Xu <yixu310@gmail.com>
Fix `added_cond_kwargs` when using IP-Adapter
Fix error when using IP-Adapter in pipeline and passing `ip_adapter_image_embeds` instead of `ip_adapter_image`
Co-authored-by: YiYi Xu <yixu310@gmail.com>
* Expand `diffusers-cli env`
* SafeTensors -> Safetensors
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
* Move `safetensors_version = "not installed"` to `else`
* Update `safetensors_version` checking
* Add GPU detection for Linux, Mac OS, and Windows
* Add accelerator detection to environment command
* Add is_peft_version to import_utils
* Update env.py
* Add `huggingface_hub` reference
* Add `transformers` reference
* Add reference for `huggingface_hub`
* Fix print statement in env.py for unusual OS
* Up
* Fix platform information in env.py
* up
* Fix import order in env.py
* ruff
* make style
* Fix platform system check in env.py
* Fix run method return type in env.py
* 🤗
* No need f-string
* Remove location info
* Remove accelerate config
* Refactor env.py to remove accelerate config
* feat: Add support for `bitsandbytes` library in environment command
---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: Dhruv Nair <dhruv.nair@gmail.com>
* find & replace all FloatTensors to Tensor
* apply formatting
* Update torch.FloatTensor to torch.Tensor in the remaining files
* formatting
* Fix the rest of the places where FloatTensor is used as well as in documentation
* formatting
* Update new file from FloatTensor to Tensor
* Remove dead code
* PylancereportGeneralTypeIssues: Strings nested within an f-string cannot use the same quote character as the f-string prior to Python 3.12.
* Remove dead code
SDXL LoRA weights for text encoders should be decoupled on save
The method checks if at least one of unet, text_encoder and
text_encoder_2 lora weights are passed, which was not reflected in the
implentation.
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: YiYi Xu <yixu310@gmail.com>
`model_output.shape` may only have rank 1.
There are warnings related to use of random keys.
```
tests/schedulers/test_scheduler_flax.py: 13 warnings
/Users/phillypham/diffusers/src/diffusers/schedulers/scheduling_ddpm_flax.py:268: FutureWarning: normal accepts a single key, but was given a key array of shape (1, 2) != (). Use jax.vmap for batching. In a future JAX version, this will be an error.
noise = jax.random.normal(split_key, shape=model_output.shape, dtype=self.dtype)
tests/schedulers/test_scheduler_flax.py::FlaxDDPMSchedulerTest::test_betas
/Users/phillypham/virtualenv/diffusers/lib/python3.9/site-packages/jax/_src/random.py:731: FutureWarning: uniform accepts a single key, but was given a key array of shape (1,) != (). Use jax.vmap for batching. In a future JAX version, this will be an error.
u = uniform(key, shape, dtype, lo, hi) # type: ignore[arg-type]
```
* Add Ascend NPU support for SDXL fine-tuning and fix the model saving bug when using DeepSpeed.
* fix check code quality
* Decouple the NPU flash attention and make it an independent module.
* add doc and unit tests for npu flash attention.
---------
Co-authored-by: mhh001 <mahonghao1@huawei.com>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
FlaxStableDiffusionSafetyChecker sets main_input_name to "clip_input".
This makes StableDiffusionSafetyChecker consistent.
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: YiYi Xu <yixu310@gmail.com>
* Added get_velocity function to EulerDiscreteScheduler.
* Fix white space on blank lines
* Added copied from statement
* back to the original.
---------
Co-authored-by: Ruining Li <ruining@robots.ox.ac.uk>
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
swap the order for do_classifier_free_guidance concat with repeat
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: Dhruv Nair <dhruv.nair@gmail.com>
* Check for latents, before calling prepare_latents - sdxlImg2Img
* Added latents check for all the img2img pipeline
* Fixed silly mistake while checking latents as None
A new function compute_dream_and_update_latents has been added to the
training utilities that allows you to do DREAM rectified training in line
with the paper https://arxiv.org/abs/2312.00210. The method can be used
with an extra argument in the train_text_to_image.py script.
Co-authored-by: Jimmy <39@🇺🇸.com>
* Convert channel order to BGR for the watermark encoder. Convert the watermarked BGR images back to RGB. Fixes#6292
* Revert channel order before stacking images to overcome limitations that negative strides are currently not supported
---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
