-
Notifications
You must be signed in to change notification settings - Fork 4.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add UniDiffuser model and pipeline #2963
Commits on May 5, 2023
-
Fix a bug of pano when not doing CFG (huggingface#3030)
* Fix a bug of pano when not doing CFG * enhance code quality * apply formatting. --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 115e382 - Browse repository at this point
Copy the full SHA 115e382View commit details -
Text2video zero refinements (huggingface#3070)
* fix progress bar issue in pipeline_text_to_video_zero.py. Copy scheduler after first backward * fix tensor loading in test_text_to_video_zero.py * make style && make quality
Configuration menu - View commit details
-
Copy full SHA for 10c54cb - Browse repository at this point
Copy the full SHA 10c54cbView commit details -
Configuration menu - View commit details
-
Copy full SHA for 945f300 - Browse repository at this point
Copy the full SHA 945f300View commit details -
[Tests] Speed up panorama tests (huggingface#3067)
* fix: norm group test for UNet3D. * chore: speed up the panorama tests (fast). * set default value of _test_inference_batch_single_identical. * fix: batch_sizes default value.
Configuration menu - View commit details
-
Copy full SHA for 322b5cb - Browse repository at this point
Copy the full SHA 322b5cbView commit details -
Configuration menu - View commit details
-
Copy full SHA for af0c3a7 - Browse repository at this point
Copy the full SHA af0c3a7View commit details -
Adds profiling flags, computes train metrics average. (huggingface#3053)
* WIP controlnet training - bugfix --streaming - bugfix running report_to!='wandb' - adds memory profile before validation * Adds final logging statement. * Sets train epochs to 11. Looking at a longer ~16ep run, we see only good validation images after ~11ep: https://wandb.ai/andsteing/controlnet_fill50k/runs/3j2hx6n8 * Removes --logging_dir (it's not used). * Adds --profile flags. * Updates --output_dir=runs/fill-circle-{timestamp}. * Compute mean of `train_metrics`. Previously `train_metrics[-1]` was logged, resulting in very bumpy train metrics. * Improves logging a bit. - adds l2_grads gradient norm logging - adds steps_per_sec - sets walltime as x coordinate of train/step - logs controlnet_params config * Adds --ccache (doesn't really help though). * minor fix in controlnet flax example (huggingface#2986) * fix the error when push_to_hub but not log validation * contronet_from_pt & controlnet_revision * add intermediate checkpointing to the guide * Bugfix --profile_steps * Sets `RACKER_PROJECT_NAME='controlnet_fill50k'`. * Logs fractional epoch. * Adds relative `walltime` metric. * Adds `StepTraceAnnotation` and uses `global_step` insetad of `step`. * Applied `black`. * Streamlines commands in README a bit. * Removes `--ccache`. This makes only a very small difference (~1 min) with this model size, so removing the option introduced in cdb3cc. * Re-ran `black`. * Update examples/controlnet/README.md Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * Converts spaces to tab. * Removes repeated args. * Skips first step (compilation) in profiling * Updates README with profiling instructions. * Unifies tabs/spaces in README. * Re-ran style & quality. --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 7a1d100 - Browse repository at this point
Copy the full SHA 7a1d100View commit details -
Configuration menu - View commit details
-
Copy full SHA for bbabf3f - Browse repository at this point
Copy the full SHA bbabf3fView commit details -
Configuration menu - View commit details
-
Copy full SHA for 068d6b4 - Browse repository at this point
Copy the full SHA 068d6b4View commit details -
[Tests] parallelize (huggingface#3078)
* [Tests] parallelize * finish folder structuring * Parallelize tests more * Correct saving of pipelines * make sure logging level is correct * try again * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Configuration menu - View commit details
-
Copy full SHA for 9dd6058 - Browse repository at this point
Copy the full SHA 9dd6058View commit details -
Throw deprecation warning for return_cached_folder (huggingface#3092)
Throw deprecation warning
Configuration menu - View commit details
-
Copy full SHA for 74907ee - Browse repository at this point
Copy the full SHA 74907eeView commit details -
Allow SD attend and excite pipeline to work with any size output imag…
…es (huggingface#2835) Allow stable diffusion attend and excite pipeline to work with any size output image. Re: huggingface#2476, huggingface#2603
Configuration menu - View commit details
-
Copy full SHA for 0be9f8b - Browse repository at this point
Copy the full SHA 0be9f8bView commit details -
[docs] Update community pipeline docs (huggingface#2989)
* update community pipeline docs * fix formatting * explain sharing workflows
Configuration menu - View commit details
-
Copy full SHA for d6ae0ae - Browse repository at this point
Copy the full SHA d6ae0aeView commit details -
Add to support Guess Mode for StableDiffusionControlnetPipleline (hug…
…gingface#2998) * add guess mode (WIP) * fix uncond/cond order * support guidance_scale=1.0 and batch != 1 * remove magic coeff * add docstring * add intergration test * add document to controlnet.mdx * made the comments a bit more explanatory * fix table
Configuration menu - View commit details
-
Copy full SHA for 7f3cb6d - Browse repository at this point
Copy the full SHA 7f3cb6dView commit details -
Configuration menu - View commit details
-
Copy full SHA for 1184b36 - Browse repository at this point
Copy the full SHA 1184b36View commit details -
remvoe one line as requested by gc team (huggingface#3077)
remvoe one line
Configuration menu - View commit details
-
Copy full SHA for fa6a6b4 - Browse repository at this point
Copy the full SHA fa6a6b4View commit details -
ddpm custom timesteps (huggingface#3007)
add custom timesteps test add custom timesteps descending order check docs timesteps -> custom_timesteps can only pass one of num_inference_steps and timesteps
Configuration menu - View commit details
-
Copy full SHA for a256f84 - Browse repository at this point
Copy the full SHA a256f84View commit details -
Fix breaking change in
pipeline_stable_diffusion_controlnet.py
(hug……gingface#3118) fix breaking change
Configuration menu - View commit details
-
Copy full SHA for e9cb03e - Browse repository at this point
Copy the full SHA e9cb03eView commit details -
Configuration menu - View commit details
-
Copy full SHA for c98e41d - Browse repository at this point
Copy the full SHA c98e41dView commit details -
[Bug fix] Fix img2img processor with safety checker (huggingface#3127)
Fix img2img processor with safety checker
Configuration menu - View commit details
-
Copy full SHA for 653b3c1 - Browse repository at this point
Copy the full SHA 653b3c1View commit details -
[Bug fix] Make sure correct timesteps are chosen for img2img (hugging…
…face#3128) Make sure correct timesteps are chosen for img2img
Configuration menu - View commit details
-
Copy full SHA for 7fa3b6c - Browse repository at this point
Copy the full SHA 7fa3b6cView commit details -
Configuration menu - View commit details
-
Copy full SHA for cc14690 - Browse repository at this point
Copy the full SHA cc14690View commit details -
Fix config deprecation (huggingface#3129)
* Better deprecation message * Better deprecation message * Better doc string * Fixes * fix more * fix more * Improve __getattr__ * correct more * fix more * fix * Improve more * more improvements * fix more * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * make style * Fix all rest & add tests & remove old deprecation fns --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Configuration menu - View commit details
-
Copy full SHA for 6f12a36 - Browse repository at this point
Copy the full SHA 6f12a36View commit details -
feat: verfication of multi-gpu support for select examples. (huggingf…
…ace#3126) * feat: verfication of multi-gpu support for select examples. * add: multi-gpu training sections to the relvant doc pages.
Configuration menu - View commit details
-
Copy full SHA for 16ddd8b - Browse repository at this point
Copy the full SHA 16ddd8bView commit details -
Configuration menu - View commit details
-
Copy full SHA for 07731e9 - Browse repository at this point
Copy the full SHA 07731e9View commit details -
Optimize log_validation in train_controlnet_flax (huggingface#3110)
extract pipeline from log_validation
Configuration menu - View commit details
-
Copy full SHA for 7a39b0f - Browse repository at this point
Copy the full SHA 7a39b0fView commit details -
Configuration menu - View commit details
-
Copy full SHA for 7ae597f - Browse repository at this point
Copy the full SHA 7ae597fView commit details -
Correct textual inversion readme (huggingface#3145)
* Update README.md * Apply suggestions from code review
Configuration menu - View commit details
-
Copy full SHA for 00a5e55 - Browse repository at this point
Copy the full SHA 00a5e55View commit details -
Add unet act fn to other model components (huggingface#3136)
Adding act fn config to the unet timestep class embedding and conv activation. The custom activation defaults to silu which is the default activation function for both the conv act and the timestep class embeddings so default behavior is not changed. The only unet which use the custom activation is the stable diffusion latent upscaler https://huggingface.co/stabilityai/sd-x2-latent-upscaler/blob/main/unet/config.json (I ran a script against the hub to confirm). The latent upscaler does not use the conv activation nor the timestep class embeddings so we don't change its behavior.
Configuration menu - View commit details
-
Copy full SHA for ff5b99b - Browse repository at this point
Copy the full SHA ff5b99bView commit details -
Configuration menu - View commit details
-
Copy full SHA for c8eaea5 - Browse repository at this point
Copy the full SHA c8eaea5View commit details -
[ckpt loader] Allow loading the Inpaint and Img2Img pipelines, while …
…loading a ckpt model (huggingface#2705) * [ckpt loader] Allow loading the Inpaint and Img2Img pipelines, while loading a ckpt model * Address review comment from PR * PyLint formatting * Some more pylint fixes, unrelated to our change * Another pylint fix * Styling fix
Configuration menu - View commit details
-
Copy full SHA for 1fac211 - Browse repository at this point
Copy the full SHA 1fac211View commit details -
add from_ckpt method as Mixin (huggingface#2318)
* add mixin class for pipeline from original sd ckpt * Improve * make style * merge main into * Improve more * fix more * up * Apply suggestions from code review * finish docs * rename * make style --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for f3300a8 - Browse repository at this point
Copy the full SHA f3300a8View commit details -
Add TensorRT SD/txt2img Community Pipeline to diffusers along with Te…
…nsorRT utils (huggingface#2974) * Add SD/txt2img Community Pipeline to diffusers along with TensorRT utils Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * update installation command Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * update tensorrt installation Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * changes 1. Update setting of cache directory 2. Address comments: merge utils and pipeline code. 3. Address comments: Add section in README Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * apply make style Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> --------- Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for fc3760d - Browse repository at this point
Copy the full SHA fc3760dView commit details -
Correct
Transformer2DModel.forward
docstring (huggingface#3074)⚙️chore(transformer_2d) update function signature for encoder_hidden_states
Configuration menu - View commit details
-
Copy full SHA for 6058378 - Browse repository at this point
Copy the full SHA 6058378View commit details -
Update pipeline_stable_diffusion_inpaint_legacy.py (huggingface#2903)
* Update pipeline_stable_diffusion_inpaint_legacy.py * fix preprocessing of Pil images with adequate batch size * revert map * add tests * reformat * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * next try to fix the style * wth is this * Update testing_utils.py * Update testing_utils.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py * Update test_stable_diffusion_inpaint_legacy.py --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for e5335f3 - Browse repository at this point
Copy the full SHA e5335f3View commit details -
Modified altdiffusion pipline to support altdiffusion-m18 (huggingfac…
…e#2993) * Modified altdiffusion pipline to support altdiffusion-m18 * Modified altdiffusion pipline to support altdiffusion-m18 * Modified altdiffusion pipline to support altdiffusion-m18 * Modified altdiffusion pipline to support altdiffusion-m18 * Modified altdiffusion pipline to support altdiffusion-m18 * Modified altdiffusion pipline to support altdiffusion-m18 * Modified altdiffusion pipline to support altdiffusion-m18 --------- Co-authored-by: root <fulong_ye@163.com>
Configuration menu - View commit details
-
Copy full SHA for 4afb911 - Browse repository at this point
Copy the full SHA 4afb911View commit details -
controlnet training resize inputs to multiple of 8 (huggingface#3135)
controlnet training center crop input images to multiple of 8 The pipeline code resizes inputs to multiples of 8. Not doing this resizing in the training script is causing the encoded image to have different height/width dimensions than the encoded conditioning image (which uses a separate encoder that's part of the controlnet model). We resize and center crop the inputs to make sure they're the same size (as well as all other images in the batch). We also check that the initial resolution is a multiple of 8.
Configuration menu - View commit details
-
Copy full SHA for b320c6b - Browse repository at this point
Copy the full SHA b320c6bView commit details -
adding custom diffusion training to diffusers examples (huggingface#3031
) * diffusers==0.14.0 update * custom diffusion update * custom diffusion update * custom diffusion update * custom diffusion update * custom diffusion update * custom diffusion update * custom diffusion * custom diffusion * custom diffusion * custom diffusion * custom diffusion * apply formatting and get rid of bare except. * refactor readme and other minor changes. * misc refactor. * fix: repo_id issue and loaders logging bug. * fix: save_model_card. * fix: save_model_card. * fix: save_model_card. * add: doc entry. * refactor doc,. * custom diffusion * custom diffusion * custom diffusion * apply style. * remove tralining whitespace. * fix: toctree entry. * remove unnecessary print. * custom diffusion * custom diffusion * custom diffusion test * custom diffusion xformer update * custom diffusion xformer update * custom diffusion xformer update --------- Co-authored-by: Nupur Kumari <nupurkumari@Nupurs-MacBook-Pro.local> Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> Co-authored-by: Nupur Kumari <nupurkumari@nupurs-mbp.wifi.local.cmu.edu>
Configuration menu - View commit details
-
Copy full SHA for 3417b1f - Browse repository at this point
Copy the full SHA 3417b1fView commit details -
Configuration menu - View commit details
-
Copy full SHA for 4a6aee9 - Browse repository at this point
Copy the full SHA 4a6aee9View commit details -
Update custom_diffusion.mdx (huggingface#3165)
Add missing newlines for rendering the links correctly
Configuration menu - View commit details
-
Copy full SHA for 3306b04 - Browse repository at this point
Copy the full SHA 3306b04View commit details -
Added distillation for quantization example on textual inversion. (hu…
…ggingface#2760) * Added distillation for quantization example on textual inversion. Signed-off-by: Ye, Xinyu <xinyu.ye@intel.com> * refined readme and code style. Signed-off-by: Ye, Xinyu <xinyu.ye@intel.com> * Update text2images.py * refined code of model load and added compatibility check. Signed-off-by: Ye, Xinyu <xinyu.ye@intel.com> * fixed code style. Signed-off-by: Ye, Xinyu <xinyu.ye@intel.com> * fix C403 [*] Unnecessary `list` comprehension (rewrite as a `set` comprehension) Signed-off-by: Ye, Xinyu <xinyu.ye@intel.com> --------- Signed-off-by: Ye, Xinyu <xinyu.ye@intel.com>
Configuration menu - View commit details
-
Copy full SHA for abd21da - Browse repository at this point
Copy the full SHA abd21daView commit details -
Update Noise Autocorrelation Loss Function for Pix2PixZero Pipeline (h…
…uggingface#2942) * Update Pix2PixZero Auto-correlation Loss * Add fast inversion tests * Clarify purpose and mark as deprecated Fix inversion prompt broadcasting * Register modules set to `None` in config for `test_save_load_optional_components` * Update new tests to coordinate with huggingface#2953
Configuration menu - View commit details
-
Copy full SHA for f99a9ff - Browse repository at this point
Copy the full SHA f99a9ffView commit details -
[DreamBooth] add text encoder LoRA support in the DreamBooth training…
… script (huggingface#3130) * add: LoRA text encoder support for DreamBooth example. * fix initialization. * fix: modification call. * add: entry in the readme. * use dog dataset from hub. * fix: params to clip. * add entry to the LoRA doc. * add: tests for lora. * remove unnecessary list comprehension./
Configuration menu - View commit details
-
Copy full SHA for 206b9b6 - Browse repository at this point
Copy the full SHA 206b9b6View commit details -
Update Habana Gaudi documentation (huggingface#3169)
* Update Habana Gaudi doc * Fix tables
Configuration menu - View commit details
-
Copy full SHA for 6ad4392 - Browse repository at this point
Copy the full SHA 6ad4392View commit details -
Add model offload to x4 upscaler (huggingface#3187)
* Add model offload to x4 upscaler * fix
Configuration menu - View commit details
-
Copy full SHA for 4366b0c - Browse repository at this point
Copy the full SHA 4366b0cView commit details -
[docs] Deterministic algorithms (huggingface#3172)
deterministic algos
Configuration menu - View commit details
-
Copy full SHA for e21784e - Browse repository at this point
Copy the full SHA e21784eView commit details -
Update custom_diffusion.mdx to credit the author (huggingface#3163)
* Update custom_diffusion.mdx * fix: unnecessary list comprehension.
Configuration menu - View commit details
-
Copy full SHA for 006ae03 - Browse repository at this point
Copy the full SHA 006ae03View commit details -
Fix TensorRT community pipeline device set function (huggingface#3157)
pass silence_dtype_warnings as kwarg Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for dac4d4a - Browse repository at this point
Copy the full SHA dac4d4aView commit details -
make
from_flax
work for controlnet (huggingface#3161)fix from_flax Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for c98a055 - Browse repository at this point
Copy the full SHA c98a055View commit details -
[docs] Clarify training args (huggingface#3146)
* clarify training arg * apply feedback
Configuration menu - View commit details
-
Copy full SHA for 15a90e2 - Browse repository at this point
Copy the full SHA 15a90e2View commit details -
Multi Vector Textual Inversion (huggingface#3144)
* Multi Vector * Improve * fix multi token * improve test * make style * Update examples/test_examples.py * Apply suggestions from code review Co-authored-by: Suraj Patil <surajp815@gmail.com> * update * Finish * Apply suggestions from code review --------- Co-authored-by: Suraj Patil <surajp815@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for cf35763 - Browse repository at this point
Copy the full SHA cf35763View commit details -
Add
Karras sigmas
to HeunDiscreteScheduler (huggingface#3160)* Add karras pattern to discrete heun scheduler * Add integration test * Fix failing CI on pytorch test on M1 (mps) --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for c729403 - Browse repository at this point
Copy the full SHA c729403View commit details -
Configuration menu - View commit details
-
Copy full SHA for 43c90b0 - Browse repository at this point
Copy the full SHA 43c90b0View commit details -
Fix bug in train_dreambooth_lora (huggingface#3183)
* Update train_dreambooth_lora.py fix bug * Update train_dreambooth_lora.py
Configuration menu - View commit details
-
Copy full SHA for 49c9b4c - Browse repository at this point
Copy the full SHA 49c9b4cView commit details -
[Community Pipelines] Update lpw_stable_diffusion pipeline (huggingfa…
…ce#3197) * Update lpw_stable_diffusion.py * fix cpu offload
Configuration menu - View commit details
-
Copy full SHA for a69502f - Browse repository at this point
Copy the full SHA a69502fView commit details -
Make sure VAE attention works with Torch 2_0 (huggingface#3200)
* Make sure attention works with Torch 2_0 * make style * Fix more
Configuration menu - View commit details
-
Copy full SHA for 3267649 - Browse repository at this point
Copy the full SHA 3267649View commit details -
Revert "[Community Pipelines] Update lpw_stable_diffusion pipeline" (h…
…uggingface#3201) Revert "[Community Pipelines] Update lpw_stable_diffusion pipeline (huggingface#3197)" This reverts commit 9965cb5.
Configuration menu - View commit details
-
Copy full SHA for de05ea0 - Browse repository at this point
Copy the full SHA de05ea0View commit details -
Configuration menu - View commit details
-
Copy full SHA for 8953209 - Browse repository at this point
Copy the full SHA 8953209View commit details -
fix mixed precision training on train_dreambooth_inpaint_lora (huggin…
…gface#3138) cast to weight dtype
Configuration menu - View commit details
-
Copy full SHA for 4e03663 - Browse repository at this point
Copy the full SHA 4e03663View commit details -
adding enable_vae_tiling and disable_vae_tiling functions (huggingfac…
…e#3225) adding enable_vae_tiling and disable_val_tiling functions
Configuration menu - View commit details
-
Copy full SHA for 167cb7a - Browse repository at this point
Copy the full SHA 167cb7aView commit details -
Add ControlNet v1.1 docs (huggingface#3226)
Add v1.1 docs
Configuration menu - View commit details
-
Copy full SHA for 0431637 - Browse repository at this point
Copy the full SHA 0431637View commit details -
Fix issue in maybe_convert_prompt (huggingface#3188)
When the token used for textual inversion does not have any special symbols (e.g. it is not surrounded by <>), the tokenizer does not properly split the replacement tokens. Adding a space for the padding tokens fixes this.
Configuration menu - View commit details
-
Copy full SHA for 9e2f445 - Browse repository at this point
Copy the full SHA 9e2f445View commit details -
Sync cache version check from transformers (huggingface#3179)
sync cache version check from transformers
Configuration menu - View commit details
-
Copy full SHA for 81950af - Browse repository at this point
Copy the full SHA 81950afView commit details -
Fix docs text inversion (huggingface#3166)
* Fix docs text inversion * Apply suggestions from code review
Configuration menu - View commit details
-
Copy full SHA for 711119a - Browse repository at this point
Copy the full SHA 711119aView commit details -
* 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>
Configuration menu - View commit details
-
Copy full SHA for 416f31a - Browse repository at this point
Copy the full SHA 416f31aView commit details -
Allow return pt x4 (huggingface#3236)
* Add all files * update
Configuration menu - View commit details
-
Copy full SHA for 7ad77dd - Browse repository at this point
Copy the full SHA 7ad77ddView commit details -
Allow fp16 attn for x4 upscaler (huggingface#3239)
* Add all files * update * Make sure vae is memory efficient for PT 1 * make style
Configuration menu - View commit details
-
Copy full SHA for 3acc879 - Browse repository at this point
Copy the full SHA 3acc879View commit details -
Configuration menu - View commit details
-
Copy full SHA for 4c73947 - Browse repository at this point
Copy the full SHA 4c73947View commit details -
Adds a document on token merging (huggingface#3208)
* add document on token merging. * fix headline. * fix: headline. * add some samples for comparison.
Configuration menu - View commit details
-
Copy full SHA for 81d7eba - Browse repository at this point
Copy the full SHA 81d7ebaView commit details -
[AudioLDM] Update docs to use updated ckpt (huggingface#3240)
* [AudioLDM] Update docs to use updated ckpt * make style
Configuration menu - View commit details
-
Copy full SHA for f83fbbd - Browse repository at this point
Copy the full SHA f83fbbdView commit details -
Configuration menu - View commit details
-
Copy full SHA for 4cc60b5 - Browse repository at this point
Copy the full SHA 4cc60b5View commit details -
Post release for 0.16.0 (huggingface#3244)
* Post release * fix more
Configuration menu - View commit details
-
Copy full SHA for 163c33b - Browse repository at this point
Copy the full SHA 163c33bView commit details -
[docs] only mention one stage (huggingface#3246)
* [docs] only mention one stage * add blurb on auto accepting --------- Co-authored-by: William Berman <WLBberman@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 59986b6 - Browse repository at this point
Copy the full SHA 59986b6View commit details -
Write model card in controlnet training script (huggingface#3229)
Write model card in controlnet training script.
Configuration menu - View commit details
-
Copy full SHA for a640f1b - Browse repository at this point
Copy the full SHA a640f1bView commit details -
[2064]: Add stochastic sampler (sample_dpmpp_sde) (huggingface#3020)
* [2064]: Add stochastic sampler * [2064]: Add stochastic sampler * [2064]: Add stochastic sampler * [2064]: Add stochastic sampler * [2064]: Add stochastic sampler * [2064]: Add stochastic sampler * [2064]: Add stochastic sampler * Review comments * [Review comment]: Add is_torchsde_available() * [Review comment]: Test and docs * [Review comment] * [Review comment] * [Review comment] * [Review comment] * [Review comment] --------- Co-authored-by: njindal <njindal@adobe.com>
Configuration menu - View commit details
-
Copy full SHA for 7880ed7 - Browse repository at this point
Copy the full SHA 7880ed7View commit details -
[Stochastic Sampler][Slow Test]: Cuda test fixes (huggingface#3257)
[Slow Test]: Cuda test fixes Co-authored-by: njindal <njindal@adobe.com>
Configuration menu - View commit details
-
Copy full SHA for 8def721 - Browse repository at this point
Copy the full SHA 8def721View commit details -
Remove required from tracker_project_name (huggingface#3260)
Remove required from tracker_project_name. As observed by https://github.com/off99555 in huggingface#2695 (comment), it already has a default value.
Configuration menu - View commit details
-
Copy full SHA for 97cf386 - Browse repository at this point
Copy the full SHA 97cf386View commit details -
adding required parameters while calling the get_up_block and get_dow…
…n_block (huggingface#3210) * removed unnecessary parameters from get_up_block and get_down_block functions * adding resnet_skip_time_act, resnet_out_scale_factor and cross_attention_norm to get_up_block and get_down_block functions --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for cf2bf70 - Browse repository at this point
Copy the full SHA cf2bf70View commit details -
[docs] Update interface in repaint.mdx (huggingface#3119)
Update repaint.mdx accomodate to huggingface#1701
Configuration menu - View commit details
-
Copy full SHA for 76e5941 - Browse repository at this point
Copy the full SHA 76e5941View commit details -
Update IF name to XL (huggingface#3262)
Co-authored-by: multimodalart <joaopaulo.passos+multimodal@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 1147c76 - Browse repository at this point
Copy the full SHA 1147c76View commit details -
Configuration menu - View commit details
-
Copy full SHA for cd13b10 - Browse repository at this point
Copy the full SHA cd13b10View commit details -
Fix typo in textual inversion JAX training script (huggingface#3123)
The pipeline is built as `pipe` but then used as `pipeline`.
Configuration menu - View commit details
-
Copy full SHA for 5f3b10a - Browse repository at this point
Copy the full SHA 5f3b10aView commit details -
AudioDiffusionPipeline - fix encode method after config changes (hugg…
…ingface#3114) * config fixes * deprecate get_input_dims
Configuration menu - View commit details
-
Copy full SHA for 9585b23 - Browse repository at this point
Copy the full SHA 9585b23View commit details -
Revert "Revert "[Community Pipelines] Update lpw_stable_diffusion pip…
…eline"" (huggingface#3265) Revert "Revert "[Community Pipelines] Update lpw_stable_diffusion pipeline" (huggingface#3201)" This reverts commit 91a2a80.
Configuration menu - View commit details
-
Copy full SHA for 79706a7 - Browse repository at this point
Copy the full SHA 79706a7View commit details -
Configuration menu - View commit details
-
Copy full SHA for 12868b1 - Browse repository at this point
Copy the full SHA 12868b1View commit details -
update notebook (huggingface#3259)
Co-authored-by: yiyixuxu <yixu@yis-macbook-pro.lan>
Configuration menu - View commit details
-
Copy full SHA for 2c87f65 - Browse repository at this point
Copy the full SHA 2c87f65View commit details -
[docs] add notes for stateful model changes (huggingface#3252)
* [docs] add notes for stateful model changes * Update docs/source/en/optimization/fp16.mdx Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * link to accelerate docs for discarding hooks --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Configuration menu - View commit details
-
Copy full SHA for a80f696 - Browse repository at this point
Copy the full SHA a80f696View commit details -
[LoRA] quality of life improvements in the loading semantics and docs (…
…huggingface#3180) * 👽 qol improvements for LoRA. * better function name? * fix: LoRA weight loading with the new format. * address Patrick's comments. * Apply suggestions from code review Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * change wording around encouraging the use of load_lora_weights(). * fix: function name. --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 72a8467 - Browse repository at this point
Copy the full SHA 72a8467View commit details -
[Community Pipelines] EDICT pipeline implementation (huggingface#3153)
* EDICT pipeline initial commit - Starting point taking from https://github.com/Joqsan/edict-diffusion * refactor __init__() method * minor refactoring * refactor scheduler code - remove scheduler and move its methods to the EDICTPipeline class * make CFG optional - refactor encode_prompt(). - include optional generator for sampling with vae. - minor variable renaming * add EDICT pipeline description to README.md * replace preprocess() with VaeImageProcessor * run make style and make quality commands --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 716c255 - Browse repository at this point
Copy the full SHA 716c255View commit details -
[Docs]zh translated docs update (huggingface#3245)
* zh translated docs update * update _toctree
Configuration menu - View commit details
-
Copy full SHA for c1dce20 - Browse repository at this point
Copy the full SHA c1dce20View commit details -
Update logging.mdx (huggingface#2863)
Fix typos
Configuration menu - View commit details
-
Copy full SHA for 5151f21 - Browse repository at this point
Copy the full SHA 5151f21View commit details -
Add multiple conditions to StableDiffusionControlNetInpaintPipeline (h…
…uggingface#3125) * try multi controlnet inpaint * multi controlnet inpaint * multi controlnet inpaint
Configuration menu - View commit details
-
Copy full SHA for ffe6e92 - Browse repository at this point
Copy the full SHA ffe6e92View commit details -
Let's make sure that dreambooth always uploads to the Hub (huggingfac…
…e#3272) * Update Dreambooth README * Adapt all docs as well * automatically write model card * fix * make style
Configuration menu - View commit details
-
Copy full SHA for 10d856a - Browse repository at this point
Copy the full SHA 10d856aView commit details -
Diffedit Zero-Shot Inpainting Pipeline (huggingface#2837)
* Update Pix2PixZero Auto-correlation Loss * Add Stable Diffusion DiffEdit pipeline * Add draft documentation and import code * Bugfixes and refactoring * Add option to not decode latents in the inversion process * Harmonize preprocessing * Revert "Update Pix2PixZero Auto-correlation Loss" This reverts commit b218062. * Update annotations * rename `compute_mask` to `generate_mask` * Update documentation * Update docs * Update Docs * Fix copy * Change shape of output latents to batch first * Update docs * Add first draft for tests * Bugfix and update tests * Add `cross_attention_kwargs` support for all pipeline methods * Fix Copies * Add support for PIL image latents Add support for mask broadcasting Update docs and tests Align `mask` argument to `mask_image` Remove height and width arguments * Enable MPS Tests * Move example docstrings * Fix test * Fix test * fix pipeline inheritance * Harmonize `prepare_image_latents` with StableDiffusionPix2PixZeroPipeline * Register modules set to `None` in config for `test_save_load_optional_components` * Move fixed logic to specific test class * Clean changes to other pipelines * Update new tests to coordinate with huggingface#2953 * Update slow tests for better results * Safety to avoid potential problems with torch.inference_mode * Add reference in SD Pipeline Overview * Fix tests again * Enforce determinism in noise for generate_mask * Fix copies * Widen test tolerance for fp16 based on `test_stable_diffusion_upscale_pipeline_fp16` * Add LoraLoaderMixin and update `prepare_image_latents` * clean up repeat and reg * bugfix * Remove invalid args from docs Suppress spurious warning by repeating image before latent to mask gen
Configuration menu - View commit details
-
Copy full SHA for 029a28f - Browse repository at this point
Copy the full SHA 029a28fView commit details -
add constant learning rate with custom rule (huggingface#3133)
* add constant lr with rules * add constant with rules in TYPE_TO_SCHEDULER_FUNCTION * add constant lr rate with rule * hotfix code quality * fix doc style * change name constant_with_rules to piecewise constant
Configuration menu - View commit details
-
Copy full SHA for fd47d7c - Browse repository at this point
Copy the full SHA fd47d7cView commit details -
Allow disabling torch 2_0 attention (huggingface#3273)
* Allow disabling torch 2_0 attention * make style * Update src/diffusers/models/attention.py
Configuration menu - View commit details
-
Copy full SHA for 220657b - Browse repository at this point
Copy the full SHA 220657bView commit details -
[doc] add link to training script (huggingface#3271)
add link to training script Co-authored-by: yiyixuxu <yixu@yis-macbook-pro.lan>
Configuration menu - View commit details
-
Copy full SHA for 1b95720 - Browse repository at this point
Copy the full SHA 1b95720View commit details -
temp disable spectogram diffusion tests (huggingface#3278)
The note-seq package throws an error on import because the default installed version of Ipython is not compatible with python 3.8 which we run in the CI. https://github.com/huggingface/diffusers/actions/runs/4830121056/jobs/8605954838#step:7:9
Configuration menu - View commit details
-
Copy full SHA for 08fbaaf - Browse repository at this point
Copy the full SHA 08fbaafView commit details -
Changed sample[0] to images[0] (huggingface#3304)
A pipeline object stores the results in `images` not in `sample`. Current code blocks don't work.
Configuration menu - View commit details
-
Copy full SHA for 799015c - Browse repository at this point
Copy the full SHA 799015cView commit details -
Configuration menu - View commit details
-
Copy full SHA for c38d004 - Browse repository at this point
Copy the full SHA c38d004View commit details -
Torch compile graph fix (huggingface#3286)
* fix more * Fix more * fix more * Apply suggestions from code review * fix * make style * make fix-copies * fix * make sure torch compile * Clean * fix test
Configuration menu - View commit details
-
Copy full SHA for 6a84a74 - Browse repository at this point
Copy the full SHA 6a84a74View commit details -
Postprocessing refactor img2img (huggingface#3268)
* refactor img2img VaeImageProcessor.postprocess * remove copy from for init, run_safety_checker, decode_latents Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> --------- Co-authored-by: yiyixuxu <yixu@yis-macbook-pro.lan> Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 863bb75 - Browse repository at this point
Copy the full SHA 863bb75View commit details -
[Torch 2.0 compile] Fix more torch compile breaks (huggingface#3313)
* Fix more torch compile breaks * add tests * Fix all * fix controlnet * fix more * Add Horace He as co-author. > > Co-authored-by: Horace He <horacehe2007@yahoo.com> * Add Horace He as co-author. Co-authored-by: Horace He <horacehe2007@yahoo.com> --------- Co-authored-by: Horace He <horacehe2007@yahoo.com>
Configuration menu - View commit details
-
Copy full SHA for c8cc4f0 - Browse repository at this point
Copy the full SHA c8cc4f0View commit details -
fix: scale_lr and sync example readme and docs. (huggingface#3299)
* fix: scale_lr and sync example readme and docs. * fix doc link.
Configuration menu - View commit details
-
Copy full SHA for 6e8d065 - Browse repository at this point
Copy the full SHA 6e8d065View commit details -
Update stable_diffusion.mdx (huggingface#3310)
fixed import statement
Configuration menu - View commit details
-
Copy full SHA for d38b4d9 - Browse repository at this point
Copy the full SHA d38b4d9View commit details -
Fix missing variable assign in DeepFloyd-IF-II (huggingface#3315)
Fix missing variable assign lol
Configuration menu - View commit details
-
Copy full SHA for 0d22064 - Browse repository at this point
Copy the full SHA 0d22064View commit details -
Correct doc build for patch releases (huggingface#3316)
Update build_documentation.yml
Configuration menu - View commit details
-
Copy full SHA for 5a75a8a - Browse repository at this point
Copy the full SHA 5a75a8aView commit details -
Add Stable Diffusion RePaint to community pipelines (huggingface#3320)
* Add Stable Diffsuion RePaint to community pipelines - Adds Stable Diffsuion RePaint to community pipelines - Add Readme enty for pipeline * Fix: Remove wrong import - Remove wrong import - Minor change in comments * Fix: Code formatting of stable_diffusion_repaint * Fix: ruff errors in stable_diffusion_repaint
Configuration menu - View commit details
-
Copy full SHA for 5ea3424 - Browse repository at this point
Copy the full SHA 5ea3424View commit details -
Fix multistep dpmsolver for cosine schedule (suitable for deepfloyd-i…
…f) (huggingface#3314) * fix multistep dpmsolver for cosine schedule (deepfloy-if) * fix a typo * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * update all dpmsolver (singlestep, multistep, dpm, dpm++) for cosine noise schedule * add test, fix style --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 7815c41 - Browse repository at this point
Copy the full SHA 7815c41View commit details -
[docs] Improve LoRA docs (huggingface#3311)
* update docs * add to toctree * apply feedback
Configuration menu - View commit details
-
Copy full SHA for 0e8f4f0 - Browse repository at this point
Copy the full SHA 0e8f4f0View commit details -
Added input pretubation (huggingface#3292)
* Added input pretubation * Fixed spelling
Configuration menu - View commit details
-
Copy full SHA for 7929587 - Browse repository at this point
Copy the full SHA 7929587View commit details -
Configuration menu - View commit details
-
Copy full SHA for 3e8d3d8 - Browse repository at this point
Copy the full SHA 3e8d3d8View commit details -
Configuration menu - View commit details
-
Copy full SHA for b4aa419 - Browse repository at this point
Copy the full SHA b4aa419View commit details -
Configuration menu - View commit details
-
Copy full SHA for 458847e - Browse repository at this point
Copy the full SHA 458847eView commit details -
Configuration menu - View commit details
-
Copy full SHA for 140ab74 - Browse repository at this point
Copy the full SHA 140ab74View commit details -
Add the SDE variant of DPM-Solver and DPM-Solver++ (huggingface#3344)
* add SDE variant of DPM-Solver and DPM-Solver++ * add test * fix typo * fix typo
Configuration menu - View commit details
-
Copy full SHA for 652dbaa - Browse repository at this point
Copy the full SHA 652dbaaView commit details -
Add upsample_size to AttnUpBlock2D, AttnDownBlock2D (huggingface#3275)
The argument `upsample_size` needs to be added to these modules to allow compatibility with other blocks that require this argument.
Configuration menu - View commit details
-
Copy full SHA for 1d213de - Browse repository at this point
Copy the full SHA 1d213deView commit details -
Add UniDiffuser classes to __init__ files, modify transformer block t…
…o support pre- and post-LN, add fast default tests, fix some bugs.
Configuration menu - View commit details
-
Copy full SHA for 434b255 - Browse repository at this point
Copy the full SHA 434b255View commit details -
Update fast tests to use test checkpoints stored on the hub and to be…
…tter match the reference UniDiffuser implementation.
Configuration menu - View commit details
-
Copy full SHA for 7097dd7 - Browse repository at this point
Copy the full SHA 7097dd7View commit details -
Configuration menu - View commit details
-
Copy full SHA for fc85263 - Browse repository at this point
Copy the full SHA fc85263View commit details -
Configuration menu - View commit details
-
Copy full SHA for 9d39bef - Browse repository at this point
Copy the full SHA 9d39befView commit details -
Configuration menu - View commit details
-
Copy full SHA for 1cb726a - Browse repository at this point
Copy the full SHA 1cb726aView commit details -
Add self.image_encoder, self.text_decoder to list of models to offloa…
…d to CPU in the enable_sequential_cpu_offload(...)/enable_model_cpu_offload(...) methods to make test_cpu_offload_forward_pass pass.
Configuration menu - View commit details
-
Copy full SHA for e62b32a - Browse repository at this point
Copy the full SHA e62b32aView commit details -
Configuration menu - View commit details
-
Copy full SHA for fc540b5 - Browse repository at this point
Copy the full SHA fc540b5View commit details
Commits on May 6, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 54c495f - Browse repository at this point
Copy the full SHA 54c495fView commit details
Commits on May 9, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 8dd7b0b - Browse repository at this point
Copy the full SHA 8dd7b0bView commit details -
Configuration menu - View commit details
-
Copy full SHA for 34a40ad - Browse repository at this point
Copy the full SHA 34a40adView commit details -
Configuration menu - View commit details
-
Copy full SHA for 0cddc3c - Browse repository at this point
Copy the full SHA 0cddc3cView commit details -
Fix bugs and make improvements to the UniDiffuser pipeline:
- Improve batch size inference and fix bugs when num_images_per_prompt or num_prompts_per_image > 1 - Add tests for num_images_per_prompt, num_prompts_per_image > 1 - Improve check_inputs, especially regarding checking supplied latents - Add reset_mode method so that mode inference can be re-enabled after mode is set manually - Fix some warnings related to accessing class members directly instead of through their config - Small amount of refactoring in pipeline_unidiffuser.py
Configuration menu - View commit details
-
Copy full SHA for 16fd515 - Browse repository at this point
Copy the full SHA 16fd515View commit details -
Configuration menu - View commit details
-
Copy full SHA for 5728328 - Browse repository at this point
Copy the full SHA 5728328View commit details
Commits on May 11, 2023
-
Add/edit docstrings for added classes and public pipeline methods. Al…
…so do some light refactoring.
Configuration menu - View commit details
-
Copy full SHA for abd6fca - Browse repository at this point
Copy the full SHA abd6fcaView commit details -
Configuration menu - View commit details
-
Copy full SHA for ae7d549 - Browse repository at this point
Copy the full SHA ae7d549View commit details -
Configuration menu - View commit details
-
Copy full SHA for 2b92111 - Browse repository at this point
Copy the full SHA 2b92111View commit details -
Configuration menu - View commit details
-
Copy full SHA for a46e1ec - Browse repository at this point
Copy the full SHA a46e1ecView commit details -
Move the UniDiffusers convert_from_ckpy.py script to diffusers/script…
…s/convert_unidiffuser_to_diffusers.py
Configuration menu - View commit details
-
Copy full SHA for a7f50f4 - Browse repository at this point
Copy the full SHA a7f50f4View commit details -
Configuration menu - View commit details
-
Copy full SHA for 8a57342 - Browse repository at this point
Copy the full SHA 8a57342View commit details -
Configuration menu - View commit details
-
Copy full SHA for 006ab49 - Browse repository at this point
Copy the full SHA 006ab49View commit details -
Configuration menu - View commit details
-
Copy full SHA for 8f2d325 - Browse repository at this point
Copy the full SHA 8f2d325View commit details -
Configuration menu - View commit details
-
Copy full SHA for a54d631 - Browse repository at this point
Copy the full SHA a54d631View commit details -
Remove outdated logic based on transformers version in UniDiffuser pi…
…peline __init__.py
Configuration menu - View commit details
-
Copy full SHA for fa9e387 - Browse repository at this point
Copy the full SHA fa9e387View commit details -
Remove dependency on einops by refactoring einops operations to pure …
…torch operations.
Configuration menu - View commit details
-
Copy full SHA for 19a20a5 - Browse repository at this point
Copy the full SHA 19a20a5View commit details -
Configuration menu - View commit details
-
Copy full SHA for 28dda62 - Browse repository at this point
Copy the full SHA 28dda62View commit details -
Add slow test on full checkpoint for joint mode and correct expected …
…image slices/text prefixes.
Configuration menu - View commit details
-
Copy full SHA for de8794c - Browse repository at this point
Copy the full SHA de8794cView commit details -
Configuration menu - View commit details
-
Copy full SHA for 7242f1b - Browse repository at this point
Copy the full SHA 7242f1bView commit details -
Fix mixed precision issue by wrapping the offending code with the tor…
…ch.autocast context manager.
Configuration menu - View commit details
-
Copy full SHA for 1a58958 - Browse repository at this point
Copy the full SHA 1a58958View commit details -
Revert "Fix mixed precision issue by wrapping the offending code with…
… the torch.autocast context manager." This reverts commit 1a58958.
Configuration menu - View commit details
-
Copy full SHA for f36df41 - Browse repository at this point
Copy the full SHA f36df41View commit details -
Configuration menu - View commit details
-
Copy full SHA for 1bc2b91 - Browse repository at this point
Copy the full SHA 1bc2b91View commit details -
Fix the mixed precision issue and add additional tests of the pipelin…
…e cuda/fp16 functionality.
Configuration menu - View commit details
-
Copy full SHA for 5341450 - Browse repository at this point
Copy the full SHA 5341450View commit details -
Configuration menu - View commit details
-
Copy full SHA for b1a6f22 - Browse repository at this point
Copy the full SHA b1a6f22View commit details -
Use a CLIPVisionModelWithProjection instead of CLIPVisionModel for im…
…age_encoder to better match the original UniDiffuser implementation.
Configuration menu - View commit details
-
Copy full SHA for 54cfa3b - Browse repository at this point
Copy the full SHA 54cfa3bView commit details -
Configuration menu - View commit details
-
Copy full SHA for 10e3774 - Browse repository at this point
Copy the full SHA 10e3774View commit details
Commits on May 12, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 4d656b5 - Browse repository at this point
Copy the full SHA 4d656b5View commit details
Commits on May 15, 2023
-
Configuration menu - View commit details
-
Copy full SHA for be4abff - Browse repository at this point
Copy the full SHA be4abffView commit details -
Add option to use fixed latents for UniDiffuserPipelineSlowTests and …
…fix issue in modeling_text_decoder.py.
Configuration menu - View commit details
-
Copy full SHA for 848b7e6 - Browse repository at this point
Copy the full SHA 848b7e6View commit details
Commits on May 16, 2023
-
Improve UniDiffuser docs, particularly the usage examples, and improv…
…e slow tests with new expected outputs.
Configuration menu - View commit details
-
Copy full SHA for e56fab2 - Browse repository at this point
Copy the full SHA e56fab2View commit details -
Configuration menu - View commit details
-
Copy full SHA for ecaf07f - Browse repository at this point
Copy the full SHA ecaf07fView commit details -
Configuration menu - View commit details
-
Copy full SHA for c161e29 - Browse repository at this point
Copy the full SHA c161e29View commit details
Commits on May 17, 2023
-
In image-to-text mode, sample from the autoencoder moment distributio…
…n instead of always getting its mode.
Configuration menu - View commit details
-
Copy full SHA for 926c7fb - Browse repository at this point
Copy the full SHA 926c7fbView commit details -
Configuration menu - View commit details
-
Copy full SHA for edbadcc - Browse repository at this point
Copy the full SHA edbadccView commit details
Commits on May 21, 2023
-
When encoding the image using the VAE, scale the image latents by the…
… VAE's scaling factor.
Configuration menu - View commit details
-
Copy full SHA for 6b35c03 - Browse repository at this point
Copy the full SHA 6b35c03View commit details -
Configuration menu - View commit details
-
Copy full SHA for f46593e - Browse repository at this point
Copy the full SHA f46593eView commit details -
Configuration menu - View commit details
-
Copy full SHA for ec7fb87 - Browse repository at this point
Copy the full SHA ec7fb87View commit details -
Configuration menu - View commit details
-
Copy full SHA for 029c96c - Browse repository at this point
Copy the full SHA 029c96cView commit details -
[docs] Fix docstring (huggingface#3334)
fix docstring Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 6644d11 - Browse repository at this point
Copy the full SHA 6644d11View commit details -
if dreambooth lora (huggingface#3360)
* update IF stage I pipelines add fixed variance schedulers and lora loading * added kv lora attn processor * allow loading into alternative lora attn processor * make vae optional * throw away predicted variance * allow loading into added kv lora layer * allow load T5 * allow pre compute text embeddings * set new variance type in schedulers * fix copies * refactor all prompt embedding code class prompts are now included in pre-encoding code max tokenizer length is now configurable embedding attention mask is now configurable * fix for when variance type is not defined on scheduler * do not pre compute validation prompt if not present * add example test for if lora dreambooth * add check for train text encoder and pre compute text embeddings
Configuration menu - View commit details
-
Copy full SHA for c221086 - Browse repository at this point
Copy the full SHA c221086View commit details -
Postprocessing refactor all others (huggingface#3337)
* add text2img * fix-copies * add * add all other pipelines * add * add * add * add * add * make style * style + fix copies --------- Co-authored-by: yiyixuxu <yixu310@gmail,com>
Configuration menu - View commit details
-
Copy full SHA for f670e08 - Browse repository at this point
Copy the full SHA f670e08View commit details -
[docs] Improve safetensors docstring (huggingface#3368)
* clarify safetensor docstring * fix typo * apply feedback
Configuration menu - View commit details
-
Copy full SHA for 7266fc1 - Browse repository at this point
Copy the full SHA 7266fc1View commit details -
add: a warning message when using xformers in a PT 2.0 env. (huggingf…
…ace#3365) * add: a warning message when using xformers in a PT 2.0 env. * 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>
Configuration menu - View commit details
-
Copy full SHA for 4b76097 - Browse repository at this point
Copy the full SHA 4b76097View commit details -
StableDiffusionInpaintingPipeline - resize image w.r.t height and wid…
…th (huggingface#3322) * StableDiffusionInpaintingPipeline now resizes input images and masks w.r.t to passed input height and width. Default is already set to 512. This addresses the common tensor mismatch error. Also moved type check into relevant funciton to keep main pipeline body tidy. * Fixed StableDiffusionInpaintingPrepareMaskAndMaskedImageTests Due to previous commit these tests were failing as height and width need to be passed into the prepare_mask_and_masked_image function, I have updated the code and added a height/width variable per unit test as it seemed more appropriate than the current hard coded solution * Added a resolution test to StableDiffusionInpaintPipelineSlowTests this unit test simply gets the input and resizes it into some that would fail (e.g. would throw a tensor mismatch error/not a mult of 8). Then passes it through the pipeline and verifies it produces output with correct dims w.r.t the passed height and width --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 6e297b4 - Browse repository at this point
Copy the full SHA 6e297b4View commit details -
Configuration menu - View commit details
-
Copy full SHA for fec7bd1 - Browse repository at this point
Copy the full SHA fec7bd1View commit details -
[docs] Adapt a model (huggingface#3326)
* first draft * apply feedback * conv_in.weight thrown away
Configuration menu - View commit details
-
Copy full SHA for e162d49 - Browse repository at this point
Copy the full SHA e162d49View commit details -
[docs] Load safetensors (huggingface#3333)
* safetensors * apply feedback * apply feedback * Apply suggestions from code review --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for caa080c - Browse repository at this point
Copy the full SHA caa080cView commit details -
Configuration menu - View commit details
-
Copy full SHA for 75c2f75 - Browse repository at this point
Copy the full SHA 75c2f75View commit details -
[Docs] Fix stable_diffusion.mdx typo (huggingface#3398)
Fix typo in last code block. Correct "prommpts" to "prompt"
Configuration menu - View commit details
-
Copy full SHA for f0c0f00 - Browse repository at this point
Copy the full SHA f0c0f00View commit details -
Support ControlNet v1.1 shuffle properly (huggingface#3340)
* add inferring_controlnet_cond_batch * Revert "add inferring_controlnet_cond_batch" This reverts commit abe8d63. * set guess_mode to True whenever global_pool_conditions is True Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * nit * add integration test --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 42eabb8 - Browse repository at this point
Copy the full SHA 42eabb8View commit details -
[Tests] better determinism (huggingface#3374)
* enable deterministic pytorch and cuda operations. * disable manual seeding. * make style && make quality for unet_2d tests. * enable determinism for the unet2dconditional model. * add CUBLAS_WORKSPACE_CONFIG for better reproducibility. * relax tolerance (very weird issue, though). * revert to torch manual_seed() where needed. * relax more tolerance. * better placement of the cuda variable and relax more tolerance. * enable determinism for 3d condition model. * relax tolerance. * add: determinism to alt_diffusion. * relax tolerance for alt diffusion. * dance diffusion. * dance diffusion is flaky. * test_dict_tuple_outputs_equivalent edit. * fix two more tests. * fix more ddim tests. * fix: argument. * change to diff in place of difference. * fix: test_save_load call. * test_save_load_float16 call. * fix: expected_max_diff * fix: paint by example. * relax tolerance. * add determinism to 1d unet model. * torch 2.0 regressions seem to be brutal * determinism to vae. * add reason to skipping. * up tolerance. * determinism to vq. * determinism to cuda. * determinism to the generic test pipeline file. * refactor general pipelines testing a bit. * determinism to alt diffusion i2i * up tolerance for alt diff i2i and audio diff * up tolerance. * determinism to audioldm * increase tolerance for audioldm lms. * increase tolerance for paint by paint. * increase tolerance for repaint. * determinism to cycle diffusion and sd 1. * relax tol for cycle diffusion 🚲 * relax tol for sd 1.0 * relax tol for controlnet. * determinism to img var. * relax tol for img variation. * tolerance to i2i sd * make style * determinism to inpaint. * relax tolerance for inpaiting. * determinism for inpainting legacy * relax tolerance. * determinism to instruct pix2pix * determinism to model editing. * model editing tolerance. * panorama determinism * determinism to pix2pix zero. * determinism to sag. * sd 2. determinism * sd. tolerance * disallow tf32 matmul. * relax tolerance is all you need. * make style and determinism to sd 2 depth * relax tolerance for depth. * tolerance to diffedit. * tolerance to sd 2 inpaint. * up tolerance. * determinism in upscaling. * tolerance in upscaler. * more tolerance relaxation. * determinism to v pred. * up tol for v_pred * unclip determinism * determinism to unclip img2img * determinism to text to video. * determinism to last set of tests * up tol. * vq cumsum doesn't have a deterministic kernel * relax tol * relax tol
Configuration menu - View commit details
-
Copy full SHA for 1965acf - Browse repository at this point
Copy the full SHA 1965acfView commit details -
[docs] Add transformers to install (huggingface#3388)
add transformers to install
Configuration menu - View commit details
-
Copy full SHA for 7b7b6bf - Browse repository at this point
Copy the full SHA 7b7b6bfView commit details -
[deepspeed] partial ZeRO-3 support (huggingface#3076)
* [deepspeed] partial ZeRO-3 support * cleanup * improve deepspeed fixes * Improve * make style --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for c998614 - Browse repository at this point
Copy the full SHA c998614View commit details -
Add omegaconf for tests (huggingface#3400)
Add omegaconfg
Configuration menu - View commit details
-
Copy full SHA for 1085f3e - Browse repository at this point
Copy the full SHA 1085f3eView commit details -
Fix various bugs with LoRA Dreambooth and Dreambooth script (huggingf…
…ace#3353) * Improve checkpointing lora * fix more * Improve doc string * Update src/diffusers/loaders.py * make stytle * Apply suggestions from code review * Update src/diffusers/loaders.py * Apply suggestions from code review * Apply suggestions from code review * better * Fix all * Fix multi-GPU dreambooth * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Fix all * make style * make style --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Configuration menu - View commit details
-
Copy full SHA for 188de89 - Browse repository at this point
Copy the full SHA 188de89View commit details -
Fix docker file (huggingface#3402)
* up * up
Configuration menu - View commit details
-
Copy full SHA for 89a8f73 - Browse repository at this point
Copy the full SHA 89a8f73View commit details -
Configuration menu - View commit details
-
Copy full SHA for cb4016d - Browse repository at this point
Copy the full SHA cb4016dView commit details -
[Docs] Add
sigmoid
beta_scheduler to docstrings of relevant Schedul……ers (huggingface#3399) * Add `sigmoid` beta scheduler to `DDPMScheduler` docstring * Add `sigmoid` beta scheduler to `RePaintScheduler` docstring --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 41763f6 - Browse repository at this point
Copy the full SHA 41763f6View commit details -
Configuration menu - View commit details
-
Copy full SHA for 622c3c6 - Browse repository at this point
Copy the full SHA 622c3c6View commit details -
Configuration menu - View commit details
-
Copy full SHA for f001e07 - Browse repository at this point
Copy the full SHA f001e07View commit details -
Improve fast tests (huggingface#3416)
Update pr_tests.yml
Configuration menu - View commit details
-
Copy full SHA for 80c2e55 - Browse repository at this point
Copy the full SHA 80c2e55View commit details -
attention refactor: the trilogy (huggingface#3387)
* Replace `AttentionBlock` with `Attention` * use _from_deprecated_attn_block check re: @patrickvonplaten
Configuration menu - View commit details
-
Copy full SHA for d749d57 - Browse repository at this point
Copy the full SHA d749d57View commit details -
[Docs] update the PT 2.0 optimization doc with latest findings (huggi…
…ngface#3370) * add: benchmarking stats for A100 and V100. * Apply suggestions from code review Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * address patrick's comments. * add: rtx 4090 stats * ⚔ benchmark reports done * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * 3313 pr link. * add: plots. Co-authored-by: Pedro <pedro@huggingface.co> * fix formattimg * update number percent. --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Configuration menu - View commit details
-
Copy full SHA for 6ce7f8f - Browse repository at this point
Copy the full SHA 6ce7f8fView commit details -
Fix style rendering (huggingface#3433)
* Fix style rendering. * Fix typo
Configuration menu - View commit details
-
Copy full SHA for 480b525 - Browse repository at this point
Copy the full SHA 480b525View commit details -
Configuration menu - View commit details
-
Copy full SHA for d3b3855 - Browse repository at this point
Copy the full SHA d3b3855View commit details -
Replace deprecated command with environment file (huggingface#3409)
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 9a31cce - Browse repository at this point
Copy the full SHA 9a31cceView commit details -
Configuration menu - View commit details
-
Copy full SHA for df625f4 - Browse repository at this point
Copy the full SHA df625f4View commit details -
add stable diffusion tensorrt img2img pipeline (huggingface#3419)
* add stable diffusion tensorrt img2img pipeline Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * update docstrings Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> --------- Signed-off-by: Asfiya Baig <asfiyab@nvidia.com>
Configuration menu - View commit details
-
Copy full SHA for 8065462 - Browse repository at this point
Copy the full SHA 8065462View commit details -
Refactor controlnet and add img2img and inpaint (huggingface#3386)
* refactor controlnet and add img2img and inpaint * First draft to get pipelines to work * make style * Fix more * Fix more * More tests * Fix more * Make inpainting work * make style and more tests * Apply suggestions from code review * up * make style * Fix imports * Fix more * Fix more * Improve examples * add test * Make sure import is correctly deprecated * Make sure everything works in compile mode * make sure authorship is correctly attributed
Configuration menu - View commit details
-
Copy full SHA for d5f65dc - Browse repository at this point
Copy the full SHA d5f65dcView commit details -
[Scheduler] DPM-Solver (++) Inverse Scheduler (huggingface#3335)
* Add DPM-Solver Multistep Inverse Scheduler * Add draft tests for DiffEdit * Add inverse sde-dpmsolver steps to tune image diversity from inverted latents * Fix tests --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 2b11926 - Browse repository at this point
Copy the full SHA 2b11926View commit details -
[Docs] Fix incomplete docstring for resnet.py (huggingface#3438)
Fix incomplete docstrings for resnet.py
Configuration menu - View commit details
-
Copy full SHA for 63abfce - Browse repository at this point
Copy the full SHA 63abfceView commit details -
fix tiled vae blend extent range (huggingface#3384)
fix tiled vae bleand extent range
Configuration menu - View commit details
-
Copy full SHA for 32162aa - Browse repository at this point
Copy the full SHA 32162aaView commit details -
Small update to "Next steps" section (huggingface#3443)
Small update to "Next steps" section: - PyTorch 2 is recommended. - Updated improvement figures.
Configuration menu - View commit details
-
Copy full SHA for 3f5a176 - Browse repository at this point
Copy the full SHA 3f5a176View commit details -
Allow arbitrary aspect ratio in IFSuperResolutionPipeline (huggingfac…
…e#3298) * Update pipeline_if_superresolution.py Allow arbitrary aspect ratio in IFSuperResolutionPipeline by using the input image shape * IFSuperResolutionPipeline: allow the user to override the height and width through the arguments * update IFSuperResolutionPipeline width/height doc string to match StableDiffusionInpaintPipeline conventions --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 3019e08 - Browse repository at this point
Copy the full SHA 3019e08View commit details -
Adding 'strength' parameter to StableDiffusionInpaintingPipeline (hug…
…gingface#3424) * Added explanation of 'strength' parameter * Added get_timesteps function which relies on new strength parameter * Added `strength` parameter which defaults to 1. * Swapped ordering so `noise_timestep` can be calculated before masking the image this is required when you aren't applying 100% noise to the masked region, e.g. strength < 1. * Added strength to check_inputs, throws error if out of range * Changed `prepare_latents` to initialise latents w.r.t strength inspired from the stable diffusion img2img pipeline, init latents are initialised by converting the init image into a VAE latent and adding noise (based upon the strength parameter passed in), e.g. random when strength = 1, or the init image at strength = 0. * WIP: Added a unit test for the new strength parameter in the StableDiffusionInpaintingPipeline still need to add correct regression values * Created a is_strength_max to initialise from pure random noise * Updated unit tests w.r.t new strength parameter + fixed new strength unit test * renamed parameter to avoid confusion with variable of same name * Updated regression values for new strength test - now passes * removed 'copied from' comment as this method is now different and divergent from the cpy * Update src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Ensure backwards compatibility for prepare_mask_and_masked_image created a return_image boolean and initialised to false * Ensure backwards compatibility for prepare_latents * Fixed copy check typo * Fixes w.r.t backward compibility changes * make style * keep function argument ordering same for backwards compatibility in callees with copied from statements * make fix-copies --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> Co-authored-by: William Berman <WLBberman@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for bb1172b - Browse repository at this point
Copy the full SHA bb1172bView commit details -
[WIP] Bugfix - Pipeline.from_pretrained is broken when the pipeline i…
…s partially downloaded (huggingface#3448) Added bugfix using f strings.
Configuration menu - View commit details
-
Copy full SHA for 2a16062 - Browse repository at this point
Copy the full SHA 2a16062View commit details -
Fix gradient checkpointing bugs in freezing part of models (requires_…
…grad=False) (huggingface#3404) * gradient checkpointing bug fix * bug fix; changes for reviews * reformat * reformat --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 68a97bd - Browse repository at this point
Copy the full SHA 68a97bdView commit details -
Make dreambooth lora more robust to orig unet (huggingface#3462)
* Make dreambooth lora more robust to orig unet * up
Configuration menu - View commit details
-
Copy full SHA for ce072e0 - Browse repository at this point
Copy the full SHA ce072e0View commit details -
Reduce peak VRAM by releasing large attention tensors (as soon as the…
…y're unnecessary) (huggingface#3463) Release large tensors in attention (as soon as they're no longer required). Reduces peak VRAM by nearly 2 GB for 1024x1024 (even after slicing), and the savings scale up with image size.
Configuration menu - View commit details
-
Copy full SHA for ee10c71 - Browse repository at this point
Copy the full SHA ee10c71View commit details -
Add min snr to text2img lora training script (huggingface#3459)
add min snr to text2img lora training script
Configuration menu - View commit details
-
Copy full SHA for 9388b3a - Browse repository at this point
Copy the full SHA 9388b3aView commit details -
Add inpaint lora scale support (huggingface#3460)
* add inpaint lora scale support * add inpaint lora scale test --------- Co-authored-by: yueyang.hyy <yueyang.hyy@alibaba-inc.com>
Configuration menu - View commit details
-
Copy full SHA for 2ef1b00 - Browse repository at this point
Copy the full SHA 2ef1b00View commit details -
[From ckpt] Fix from_ckpt (huggingface#3466)
* Correct from_ckpt * make style
Configuration menu - View commit details
-
Copy full SHA for 62d9c72 - Browse repository at this point
Copy the full SHA 62d9c72View commit details -
Configuration menu - View commit details
-
Copy full SHA for 368f9ad - Browse repository at this point
Copy the full SHA 368f9adView commit details -
Configuration menu - View commit details
-
Copy full SHA for 68441bf - Browse repository at this point
Copy the full SHA 68441bfView commit details -
Configuration menu - View commit details
-
Copy full SHA for eb7ae28 - Browse repository at this point
Copy the full SHA eb7ae28View commit details -
Configuration menu - View commit details
-
Copy full SHA for bb1e25a - Browse repository at this point
Copy the full SHA bb1e25aView commit details -
dreambooth docs torch.compile note (huggingface#3471)
* dreambooth docs torch.compile note * Update examples/dreambooth/README.md Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * Update examples/dreambooth/README.md Co-authored-by: Pedro Cuenca <pedro@huggingface.co> --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Configuration menu - View commit details
-
Copy full SHA for 9a195d7 - Browse repository at this point
Copy the full SHA 9a195d7View commit details -
Configuration menu - View commit details
-
Copy full SHA for 09ddb88 - Browse repository at this point
Copy the full SHA 09ddb88View commit details -
[docs] Textual inversion inference (huggingface#3473)
* add textual inversion inference to docs * add to toctree --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for e36596c - Browse repository at this point
Copy the full SHA e36596cView commit details -
[docs] Distributed inference (huggingface#3376)
* distributed inference * move to inference section * apply feedback * update with split_between_processes * apply feedback
Configuration menu - View commit details
-
Copy full SHA for 147da83 - Browse repository at this point
Copy the full SHA 147da83View commit details -
[{Up,Down}sample1d] explicit view kernel size as number elements in f…
…lattened indices (huggingface#3479) explicit view kernel size as number elements in flattened indices
Configuration menu - View commit details
-
Copy full SHA for 2d8e089 - Browse repository at this point
Copy the full SHA 2d8e089View commit details -
mps & onnx tests rework (huggingface#3449)
* Remove ONNX tests from PR. They are already a part of push_tests.yml. * Remove mps tests from PRs. They are already performed on push. * Fix workflow name for fast push tests. * Extract mps tests to a workflow. For better control/filtering. * Remove --extra-index-url from mps tests * Increase tolerance of mps test This test passes in my Mac (Ventura 13.3) but fails in the CI hardware (Ventura 13.2). I ran the local tests following the same steps that exist in the CI workflow. * Temporarily run mps tests on pr So we can test. * Revert "Temporarily run mps tests on pr" Tests passed, go back to running on push.
Configuration menu - View commit details
-
Copy full SHA for 53e37b8 - Browse repository at this point
Copy the full SHA 53e37b8View commit details -
[Attention processor] Better warning message when shifting to `AttnPr…
…ocessor2_0` (huggingface#3457) * add: debugging to enabling memory efficient processing * add: better warning message.
Configuration menu - View commit details
-
Copy full SHA for 8eae86d - Browse repository at this point
Copy the full SHA 8eae86dView commit details -
[Docs] add note on local directory path. (huggingface#3397)
add note on local directory path. Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 55ca69b - Browse repository at this point
Copy the full SHA 55ca69bView commit details -
Refactor full determinism (huggingface#3485)
* up * fix more * Apply suggestions from code review * fix more * fix more * Check it * Remove 16:8 * fix more * fix more * fix more * up * up * Test only stable diffusion * Test only two files * up * Try out spinning up processes that can be killed * up * Apply suggestions from code review * up * up
Configuration menu - View commit details
-
Copy full SHA for a8219e8 - Browse repository at this point
Copy the full SHA a8219e8View commit details -
Fix DPM single (huggingface#3413)
* Fix DPM single * add test * fix one more bug * Apply suggestions from code review Co-authored-by: StAlKeR7779 <stalkek7779@yandex.ru> --------- Co-authored-by: StAlKeR7779 <stalkek7779@yandex.ru>
Configuration menu - View commit details
-
Copy full SHA for a3e1153 - Browse repository at this point
Copy the full SHA a3e1153View commit details -
Add
use_Karras_sigmas
to DPMSolverSinglestepScheduler (huggingface#……3476) * add use_karras_sigmas * add karras test * add doc
Configuration menu - View commit details
-
Copy full SHA for 01b42e4 - Browse repository at this point
Copy the full SHA 01b42e4View commit details -
Configuration menu - View commit details
-
Copy full SHA for d22535a - Browse repository at this point
Copy the full SHA d22535aView commit details -
Configuration menu - View commit details
-
Copy full SHA for b78e854 - Browse repository at this point
Copy the full SHA b78e854View commit details -
[Docs] Korean translation (optimization, training) (huggingface#3488)
* feat) optimization kr translation * fix) typo, italic setting * feat) dreambooth, text2image kr * feat) lora kr * fix) LoRA * fix) fp16 fix * fix) doc-builder style * fix) fp16 일부 단어 수정 * fix) fp16 style fix * fix) opt, training docs update * feat) toctree update * feat) toctree update --------- Co-authored-by: Chanran Kim <seriousran@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for a9ac5a8 - Browse repository at this point
Copy the full SHA a9ac5a8View commit details -
DataLoader respecting EXIF data in Training Images (huggingface#3465)
* DataLoader will now bake in any transforms or image manipulations contained in the EXIF Images may have rotations stored in EXIF. Training using such images will cause those transforms to be ignored while training and thus produce unexpected results * Fixed the Dataloading EXIF issue in main DreamBooth training as well * Run make style (black & isort)
Configuration menu - View commit details
-
Copy full SHA for 634cf1f - Browse repository at this point
Copy the full SHA 634cf1fView commit details -
Configuration menu - View commit details
-
Copy full SHA for 5782887 - Browse repository at this point
Copy the full SHA 5782887View commit details -
feat: allow disk offload for diffuser models (huggingface#3285)
* allow disk offload for diffuser models * sort import * add max_memory argument * Changed sample[0] to images[0] (huggingface#3304) A pipeline object stores the results in `images` not in `sample`. Current code blocks don't work. * Typo in tutorial (huggingface#3295) * Torch compile graph fix (huggingface#3286) * fix more * Fix more * fix more * Apply suggestions from code review * fix * make style * make fix-copies * fix * make sure torch compile * Clean * fix test * Postprocessing refactor img2img (huggingface#3268) * refactor img2img VaeImageProcessor.postprocess * remove copy from for init, run_safety_checker, decode_latents Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> --------- Co-authored-by: yiyixuxu <yixu@yis-macbook-pro.lan> Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * [Torch 2.0 compile] Fix more torch compile breaks (huggingface#3313) * Fix more torch compile breaks * add tests * Fix all * fix controlnet * fix more * Add Horace He as co-author. > > Co-authored-by: Horace He <horacehe2007@yahoo.com> * Add Horace He as co-author. Co-authored-by: Horace He <horacehe2007@yahoo.com> --------- Co-authored-by: Horace He <horacehe2007@yahoo.com> * fix: scale_lr and sync example readme and docs. (huggingface#3299) * fix: scale_lr and sync example readme and docs. * fix doc link. * Update stable_diffusion.mdx (huggingface#3310) fixed import statement * Fix missing variable assign in DeepFloyd-IF-II (huggingface#3315) Fix missing variable assign lol * Correct doc build for patch releases (huggingface#3316) Update build_documentation.yml * Add Stable Diffusion RePaint to community pipelines (huggingface#3320) * Add Stable Diffsuion RePaint to community pipelines - Adds Stable Diffsuion RePaint to community pipelines - Add Readme enty for pipeline * Fix: Remove wrong import - Remove wrong import - Minor change in comments * Fix: Code formatting of stable_diffusion_repaint * Fix: ruff errors in stable_diffusion_repaint * Fix multistep dpmsolver for cosine schedule (suitable for deepfloyd-if) (huggingface#3314) * fix multistep dpmsolver for cosine schedule (deepfloy-if) * fix a typo * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update src/diffusers/schedulers/scheduling_dpmsolver_multistep.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * update all dpmsolver (singlestep, multistep, dpm, dpm++) for cosine noise schedule * add test, fix style --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * [docs] Improve LoRA docs (huggingface#3311) * update docs * add to toctree * apply feedback * Added input pretubation (huggingface#3292) * Added input pretubation * Fixed spelling * Update write_own_pipeline.mdx (huggingface#3323) * update controlling generation doc with latest goodies. (huggingface#3321) * [Quality] Make style (huggingface#3341) * Fix config dpm (huggingface#3343) * Add the SDE variant of DPM-Solver and DPM-Solver++ (huggingface#3344) * add SDE variant of DPM-Solver and DPM-Solver++ * add test * fix typo * fix typo * Add upsample_size to AttnUpBlock2D, AttnDownBlock2D (huggingface#3275) The argument `upsample_size` needs to be added to these modules to allow compatibility with other blocks that require this argument. * Rename --only_save_embeds to --save_as_full_pipeline (huggingface#3206) * Set --only_save_embeds to False by default Due to how the option is named, it makes more sense to behave like this. * Refactor only_save_embeds to save_as_full_pipeline * [AudioLDM] Generalise conversion script (huggingface#3328) Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Fix TypeError when using prompt_embeds and negative_prompt (huggingface#2982) * test: Added test case * fix: fixed type checking issue on _encode_prompt * fix: fixed copies consistency * fix: one copy was not sufficient * Fix pipeline class on README (huggingface#3345) Update README.md * Inpainting: typo in docs (huggingface#3331) Typo in docs Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Add `use_Karras_sigmas` to LMSDiscreteScheduler (huggingface#3351) * add karras sigma to lms discrete scheduler * add test for lms_scheduler karras * reformat test lms * Batched load of textual inversions (huggingface#3277) * Batched load of textual inversions - Only call resize_token_embeddings once per batch as it is the most expensive operation - Allow pretrained_model_name_or_path and token to be an optional list - Remove Dict from type annotation pretrained_model_name_or_path as it was not supported in this function - Add comment that single files (e.g. .pt/.safetensors) are supported - Add comment for token parameter - Convert token override log message from warning to info * Update src/diffusers/loaders.py Check for duplicate tokens Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update condition for None tokens --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * make fix-copies * [docs] Fix docstring (huggingface#3334) fix docstring Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * if dreambooth lora (huggingface#3360) * update IF stage I pipelines add fixed variance schedulers and lora loading * added kv lora attn processor * allow loading into alternative lora attn processor * make vae optional * throw away predicted variance * allow loading into added kv lora layer * allow load T5 * allow pre compute text embeddings * set new variance type in schedulers * fix copies * refactor all prompt embedding code class prompts are now included in pre-encoding code max tokenizer length is now configurable embedding attention mask is now configurable * fix for when variance type is not defined on scheduler * do not pre compute validation prompt if not present * add example test for if lora dreambooth * add check for train text encoder and pre compute text embeddings * Postprocessing refactor all others (huggingface#3337) * add text2img * fix-copies * add * add all other pipelines * add * add * add * add * add * make style * style + fix copies --------- Co-authored-by: yiyixuxu <yixu310@gmail,com> * [docs] Improve safetensors docstring (huggingface#3368) * clarify safetensor docstring * fix typo * apply feedback * add: a warning message when using xformers in a PT 2.0 env. (huggingface#3365) * add: a warning message when using xformers in a PT 2.0 env. * 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> * StableDiffusionInpaintingPipeline - resize image w.r.t height and width (huggingface#3322) * StableDiffusionInpaintingPipeline now resizes input images and masks w.r.t to passed input height and width. Default is already set to 512. This addresses the common tensor mismatch error. Also moved type check into relevant funciton to keep main pipeline body tidy. * Fixed StableDiffusionInpaintingPrepareMaskAndMaskedImageTests Due to previous commit these tests were failing as height and width need to be passed into the prepare_mask_and_masked_image function, I have updated the code and added a height/width variable per unit test as it seemed more appropriate than the current hard coded solution * Added a resolution test to StableDiffusionInpaintPipelineSlowTests this unit test simply gets the input and resizes it into some that would fail (e.g. would throw a tensor mismatch error/not a mult of 8). Then passes it through the pipeline and verifies it produces output with correct dims w.r.t the passed height and width --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * make style * [docs] Adapt a model (huggingface#3326) * first draft * apply feedback * conv_in.weight thrown away * [docs] Load safetensors (huggingface#3333) * safetensors * apply feedback * apply feedback * Apply suggestions from code review --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * make style * [Docs] Fix stable_diffusion.mdx typo (huggingface#3398) Fix typo in last code block. Correct "prommpts" to "prompt" * Support ControlNet v1.1 shuffle properly (huggingface#3340) * add inferring_controlnet_cond_batch * Revert "add inferring_controlnet_cond_batch" This reverts commit abe8d63. * set guess_mode to True whenever global_pool_conditions is True Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * nit * add integration test --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * [Tests] better determinism (huggingface#3374) * enable deterministic pytorch and cuda operations. * disable manual seeding. * make style && make quality for unet_2d tests. * enable determinism for the unet2dconditional model. * add CUBLAS_WORKSPACE_CONFIG for better reproducibility. * relax tolerance (very weird issue, though). * revert to torch manual_seed() where needed. * relax more tolerance. * better placement of the cuda variable and relax more tolerance. * enable determinism for 3d condition model. * relax tolerance. * add: determinism to alt_diffusion. * relax tolerance for alt diffusion. * dance diffusion. * dance diffusion is flaky. * test_dict_tuple_outputs_equivalent edit. * fix two more tests. * fix more ddim tests. * fix: argument. * change to diff in place of difference. * fix: test_save_load call. * test_save_load_float16 call. * fix: expected_max_diff * fix: paint by example. * relax tolerance. * add determinism to 1d unet model. * torch 2.0 regressions seem to be brutal * determinism to vae. * add reason to skipping. * up tolerance. * determinism to vq. * determinism to cuda. * determinism to the generic test pipeline file. * refactor general pipelines testing a bit. * determinism to alt diffusion i2i * up tolerance for alt diff i2i and audio diff * up tolerance. * determinism to audioldm * increase tolerance for audioldm lms. * increase tolerance for paint by paint. * increase tolerance for repaint. * determinism to cycle diffusion and sd 1. * relax tol for cycle diffusion 🚲 * relax tol for sd 1.0 * relax tol for controlnet. * determinism to img var. * relax tol for img variation. * tolerance to i2i sd * make style * determinism to inpaint. * relax tolerance for inpaiting. * determinism for inpainting legacy * relax tolerance. * determinism to instruct pix2pix * determinism to model editing. * model editing tolerance. * panorama determinism * determinism to pix2pix zero. * determinism to sag. * sd 2. determinism * sd. tolerance * disallow tf32 matmul. * relax tolerance is all you need. * make style and determinism to sd 2 depth * relax tolerance for depth. * tolerance to diffedit. * tolerance to sd 2 inpaint. * up tolerance. * determinism in upscaling. * tolerance in upscaler. * more tolerance relaxation. * determinism to v pred. * up tol for v_pred * unclip determinism * determinism to unclip img2img * determinism to text to video. * determinism to last set of tests * up tol. * vq cumsum doesn't have a deterministic kernel * relax tol * relax tol * [docs] Add transformers to install (huggingface#3388) add transformers to install * [deepspeed] partial ZeRO-3 support (huggingface#3076) * [deepspeed] partial ZeRO-3 support * cleanup * improve deepspeed fixes * Improve * make style --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Add omegaconf for tests (huggingface#3400) Add omegaconfg * Fix various bugs with LoRA Dreambooth and Dreambooth script (huggingface#3353) * Improve checkpointing lora * fix more * Improve doc string * Update src/diffusers/loaders.py * make stytle * Apply suggestions from code review * Update src/diffusers/loaders.py * Apply suggestions from code review * Apply suggestions from code review * better * Fix all * Fix multi-GPU dreambooth * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Fix all * make style * make style --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Fix docker file (huggingface#3402) * up * up * fix: deepseepd_plugin retrieval from accelerate state (huggingface#3410) * [Docs] Add `sigmoid` beta_scheduler to docstrings of relevant Schedulers (huggingface#3399) * Add `sigmoid` beta scheduler to `DDPMScheduler` docstring * Add `sigmoid` beta scheduler to `RePaintScheduler` docstring --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Don't install accelerate and transformers from source (huggingface#3415) * Don't install transformers and accelerate from source (huggingface#3414) * Improve fast tests (huggingface#3416) Update pr_tests.yml * attention refactor: the trilogy (huggingface#3387) * Replace `AttentionBlock` with `Attention` * use _from_deprecated_attn_block check re: @patrickvonplaten * [Docs] update the PT 2.0 optimization doc with latest findings (huggingface#3370) * add: benchmarking stats for A100 and V100. * Apply suggestions from code review Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * address patrick's comments. * add: rtx 4090 stats * ⚔ benchmark reports done * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * 3313 pr link. * add: plots. Co-authored-by: Pedro <pedro@huggingface.co> * fix formattimg * update number percent. --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Fix style rendering (huggingface#3433) * Fix style rendering. * Fix typo * unCLIP scheduler do not use note (huggingface#3417) * Replace deprecated command with environment file (huggingface#3409) Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * fix warning message pipeline loading (huggingface#3446) * add stable diffusion tensorrt img2img pipeline (huggingface#3419) * add stable diffusion tensorrt img2img pipeline Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * update docstrings Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> --------- Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> * Refactor controlnet and add img2img and inpaint (huggingface#3386) * refactor controlnet and add img2img and inpaint * First draft to get pipelines to work * make style * Fix more * Fix more * More tests * Fix more * Make inpainting work * make style and more tests * Apply suggestions from code review * up * make style * Fix imports * Fix more * Fix more * Improve examples * add test * Make sure import is correctly deprecated * Make sure everything works in compile mode * make sure authorship is correctly attributed * [Scheduler] DPM-Solver (++) Inverse Scheduler (huggingface#3335) * Add DPM-Solver Multistep Inverse Scheduler * Add draft tests for DiffEdit * Add inverse sde-dpmsolver steps to tune image diversity from inverted latents * Fix tests --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * [Docs] Fix incomplete docstring for resnet.py (huggingface#3438) Fix incomplete docstrings for resnet.py * fix tiled vae blend extent range (huggingface#3384) fix tiled vae bleand extent range * Small update to "Next steps" section (huggingface#3443) Small update to "Next steps" section: - PyTorch 2 is recommended. - Updated improvement figures. * Allow arbitrary aspect ratio in IFSuperResolutionPipeline (huggingface#3298) * Update pipeline_if_superresolution.py Allow arbitrary aspect ratio in IFSuperResolutionPipeline by using the input image shape * IFSuperResolutionPipeline: allow the user to override the height and width through the arguments * update IFSuperResolutionPipeline width/height doc string to match StableDiffusionInpaintPipeline conventions --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Adding 'strength' parameter to StableDiffusionInpaintingPipeline (huggingface#3424) * Added explanation of 'strength' parameter * Added get_timesteps function which relies on new strength parameter * Added `strength` parameter which defaults to 1. * Swapped ordering so `noise_timestep` can be calculated before masking the image this is required when you aren't applying 100% noise to the masked region, e.g. strength < 1. * Added strength to check_inputs, throws error if out of range * Changed `prepare_latents` to initialise latents w.r.t strength inspired from the stable diffusion img2img pipeline, init latents are initialised by converting the init image into a VAE latent and adding noise (based upon the strength parameter passed in), e.g. random when strength = 1, or the init image at strength = 0. * WIP: Added a unit test for the new strength parameter in the StableDiffusionInpaintingPipeline still need to add correct regression values * Created a is_strength_max to initialise from pure random noise * Updated unit tests w.r.t new strength parameter + fixed new strength unit test * renamed parameter to avoid confusion with variable of same name * Updated regression values for new strength test - now passes * removed 'copied from' comment as this method is now different and divergent from the cpy * Update src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Ensure backwards compatibility for prepare_mask_and_masked_image created a return_image boolean and initialised to false * Ensure backwards compatibility for prepare_latents * Fixed copy check typo * Fixes w.r.t backward compibility changes * make style * keep function argument ordering same for backwards compatibility in callees with copied from statements * make fix-copies --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> Co-authored-by: William Berman <WLBberman@gmail.com> * [WIP] Bugfix - Pipeline.from_pretrained is broken when the pipeline is partially downloaded (huggingface#3448) Added bugfix using f strings. * Fix gradient checkpointing bugs in freezing part of models (requires_grad=False) (huggingface#3404) * gradient checkpointing bug fix * bug fix; changes for reviews * reformat * reformat --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Make dreambooth lora more robust to orig unet (huggingface#3462) * Make dreambooth lora more robust to orig unet * up * Reduce peak VRAM by releasing large attention tensors (as soon as they're unnecessary) (huggingface#3463) Release large tensors in attention (as soon as they're no longer required). Reduces peak VRAM by nearly 2 GB for 1024x1024 (even after slicing), and the savings scale up with image size. * Add min snr to text2img lora training script (huggingface#3459) add min snr to text2img lora training script * Add inpaint lora scale support (huggingface#3460) * add inpaint lora scale support * add inpaint lora scale test --------- Co-authored-by: yueyang.hyy <yueyang.hyy@alibaba-inc.com> * [From ckpt] Fix from_ckpt (huggingface#3466) * Correct from_ckpt * make style * Update full dreambooth script to work with IF (huggingface#3425) * Add IF dreambooth docs (huggingface#3470) * parameterize pass single args through tuple (huggingface#3477) * attend and excite tests disable determinism on the class level (huggingface#3478) * dreambooth docs torch.compile note (huggingface#3471) * dreambooth docs torch.compile note * Update examples/dreambooth/README.md Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * Update examples/dreambooth/README.md Co-authored-by: Pedro Cuenca <pedro@huggingface.co> --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * add: if entry in the dreambooth training docs. (huggingface#3472) * [docs] Textual inversion inference (huggingface#3473) * add textual inversion inference to docs * add to toctree --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> * [docs] Distributed inference (huggingface#3376) * distributed inference * move to inference section * apply feedback * update with split_between_processes * apply feedback * [{Up,Down}sample1d] explicit view kernel size as number elements in flattened indices (huggingface#3479) explicit view kernel size as number elements in flattened indices * mps & onnx tests rework (huggingface#3449) * Remove ONNX tests from PR. They are already a part of push_tests.yml. * Remove mps tests from PRs. They are already performed on push. * Fix workflow name for fast push tests. * Extract mps tests to a workflow. For better control/filtering. * Remove --extra-index-url from mps tests * Increase tolerance of mps test This test passes in my Mac (Ventura 13.3) but fails in the CI hardware (Ventura 13.2). I ran the local tests following the same steps that exist in the CI workflow. * Temporarily run mps tests on pr So we can test. * Revert "Temporarily run mps tests on pr" Tests passed, go back to running on push. --------- Signed-off-by: Asfiya Baig <asfiyab@nvidia.com> Co-authored-by: Ilia Larchenko <41329713+IliaLarchenko@users.noreply.github.com> Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> Co-authored-by: YiYi Xu <yixu310@gmail.com> Co-authored-by: yiyixuxu <yixu@yis-macbook-pro.lan> Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> Co-authored-by: Horace He <horacehe2007@yahoo.com> Co-authored-by: Umar <55330742+mu94-csl@users.noreply.github.com> Co-authored-by: Mylo <36931363+gitmylo@users.noreply.github.com> Co-authored-by: Markus Pobitzer <markuspobitzer@gmail.com> Co-authored-by: Cheng Lu <lucheng.lc15@gmail.com> Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> Co-authored-by: Isamu Isozaki <isamu.website@gmail.com> Co-authored-by: Cesar Aybar <csaybar@gmail.com> Co-authored-by: Will Rice <will@spokestack.io> Co-authored-by: Adrià Arrufat <1671644+arrufat@users.noreply.github.com> Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> Co-authored-by: At-sushi <dkahw210@kyoto.zaq.ne.jp> Co-authored-by: Lucca Zenóbio <luccazen@gmail.com> Co-authored-by: Lysandre Debut <lysandre@huggingface.co> Co-authored-by: Isotr0py <41363108+Isotr0py@users.noreply.github.com> Co-authored-by: pdoane <pdoane2@gmail.com> Co-authored-by: Will Berman <wlbberman@gmail.com> Co-authored-by: yiyixuxu <yixu310@gmail,com> Co-authored-by: Rupert Menneer <71332436+rupertmenneer@users.noreply.github.com> Co-authored-by: sudowind <wfpkueecs@163.com> Co-authored-by: Takuma Mori <takuma104@gmail.com> Co-authored-by: Stas Bekman <stas00@users.noreply.github.com> Co-authored-by: Pedro Cuenca <pedro@huggingface.co> Co-authored-by: Laureηt <laurentfainsin@protonmail.com> Co-authored-by: Jongwoo Han <jongwooo.han@gmail.com> Co-authored-by: asfiyab-nvidia <117682710+asfiyab-nvidia@users.noreply.github.com> Co-authored-by: clarencechen <clarencechenct@gmail.com> Co-authored-by: Laureηt <laurent@fainsin.bzh> Co-authored-by: superlabs-dev <133080491+superlabs-dev@users.noreply.github.com> Co-authored-by: Dev Aggarwal <devxpy@gmail.com> Co-authored-by: Vimarsh Chaturvedi <vimarsh.c@gmail.com> Co-authored-by: 7eu7d7 <31194890+7eu7d7@users.noreply.github.com> Co-authored-by: cmdr2 <shashank.shekhar.global@gmail.com> Co-authored-by: wfng92 <43742196+wfng92@users.noreply.github.com> Co-authored-by: Glaceon-Hyy <ffheyy0017@gmail.com> Co-authored-by: yueyang.hyy <yueyang.hyy@alibaba-inc.com>
Configuration menu - View commit details
-
Copy full SHA for f61028f - Browse repository at this point
Copy the full SHA f61028fView commit details -
[Community] reference only control (huggingface#3435)
* add reference only control * add reference only control * add reference only control * fix lint * fix lint * reference adain * bugfix EulerAncestralDiscreteScheduler * fix style fidelity rule * fix default output size * del unused line * fix deterministic
Configuration menu - View commit details
-
Copy full SHA for 30329a2 - Browse repository at this point
Copy the full SHA 30329a2View commit details -
Support for cross-attention bias / mask (huggingface#2634)
* Cross-attention masks prefer qualified symbol, fix accidental Optional prefer qualified symbol in AttentionProcessor prefer qualified symbol in embeddings.py qualified symbol in transformed_2d qualify FloatTensor in unet_2d_blocks move new transformer_2d params attention_mask, encoder_attention_mask to the end of the section which is assumed (e.g. by functions such as checkpoint()) to have a stable positional param interface. regard return_dict as a special-case which is assumed to be injected separately from positional params (e.g. by create_custom_forward()). move new encoder_attention_mask param to end of CrossAttn block interfaces and Unet2DCondition interface, to maintain positional param interface. regenerate modeling_text_unet.py remove unused import unet_2d_condition encoder_attention_mask docs Co-authored-by: Pedro Cuenca <pedro@huggingface.co> versatile_diffusion/modeling_text_unet.py encoder_attention_mask docs Co-authored-by: Pedro Cuenca <pedro@huggingface.co> transformer_2d encoder_attention_mask docs Co-authored-by: Pedro Cuenca <pedro@huggingface.co> unet_2d_blocks.py: add parameter name comments Co-authored-by: Pedro Cuenca <pedro@huggingface.co> revert description. bool-to-bias treatment happens in unet_2d_condition only. comment parameter names fix copies, style * encoder_attention_mask for SimpleCrossAttnDownBlock2D, SimpleCrossAttnUpBlock2D * encoder_attention_mask for UNetMidBlock2DSimpleCrossAttn * support attention_mask, encoder_attention_mask in KCrossAttnDownBlock2D, KCrossAttnUpBlock2D, KAttentionBlock. fix binding of attention_mask, cross_attention_kwargs params in KCrossAttnDownBlock2D, KCrossAttnUpBlock2D checkpoint invocations. * fix mistake made during merge conflict resolution * regenerate versatile_diffusion * pass time embedding into checkpointed attention invocation * always assume encoder_attention_mask is a mask (i.e. not a bias). * style, fix-copies * add tests for cross-attention masks * add test for padding of attention mask * explain mask's query_tokens dim. fix explanation about broadcasting over channels; we actually broadcast over query tokens * support both masks and biases in Transformer2DModel#forward. document behaviour * fix-copies * delete attention_mask docs on the basis I never tested self-attention masking myself. not comfortable explaining it, since I don't actually understand how a self-attn mask can work in its current form: the key length will be different in every ResBlock (we don't downsample the mask when we downsample the image). * review feedback: the standard Unet blocks shouldn't pass temb to attn (only to resnet). remove from KCrossAttnDownBlock2D,KCrossAttnUpBlock2D#forward. * remove encoder_attention_mask param from SimpleCrossAttn{Up,Down}Block2D,UNetMidBlock2DSimpleCrossAttn, and mask-choice in those blocks' #forward, on the basis that they only do one type of attention, so the consumer can pass whichever type of attention_mask is appropriate. * put attention mask padding back to how it was (since the SD use-case it enabled wasn't important, and it breaks the original unclip use-case). disable the test which was added. * fix-copies * style * fix-copies * put encoder_attention_mask param back into Simple block forward interfaces, to ensure consistency of forward interface. * restore passing of emb to KAttentionBlock#forward, on the basis that removal caused test failures. restore also the passing of emb to checkpointed calls to KAttentionBlock#forward. * make simple unet2d blocks use encoder_attention_mask, but only when attention_mask is None. this should fix UnCLIP compatibility. * fix copies
Configuration menu - View commit details
-
Copy full SHA for ca87f4d - Browse repository at this point
Copy the full SHA ca87f4dView commit details -
do not scale the initial global step by gradient accumulation steps w…
…hen loading from checkpoint (huggingface#3506)
Configuration menu - View commit details
-
Copy full SHA for cdf38f1 - Browse repository at this point
Copy the full SHA cdf38f1View commit details -
Configuration menu - View commit details
-
Copy full SHA for 51f0951 - Browse repository at this point
Copy the full SHA 51f0951View commit details -
Configuration menu - View commit details
-
Copy full SHA for 4ccb2b5 - Browse repository at this point
Copy the full SHA 4ccb2b5View commit details -
Revert "Clean up code and make slow tests pass."
This reverts commit ec7fb87.
Configuration menu - View commit details
-
Copy full SHA for 97e8eef - Browse repository at this point
Copy the full SHA 97e8eefView commit details -
Batched load of textual inversions (huggingface#3277)
* Batched load of textual inversions - Only call resize_token_embeddings once per batch as it is the most expensive operation - Allow pretrained_model_name_or_path and token to be an optional list - Remove Dict from type annotation pretrained_model_name_or_path as it was not supported in this function - Add comment that single files (e.g. .pt/.safetensors) are supported - Add comment for token parameter - Convert token override log message from warning to info * Update src/diffusers/loaders.py Check for duplicate tokens Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * Update condition for None tokens --------- Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 9f84416 - Browse repository at this point
Copy the full SHA 9f84416View commit details
Commits on May 22, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 9f7247c - Browse repository at this point
Copy the full SHA 9f7247cView commit details
Commits on May 23, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 302fde9 - Browse repository at this point
Copy the full SHA 302fde9View commit details -
Configuration menu - View commit details
-
Copy full SHA for 6326cb7 - Browse repository at this point
Copy the full SHA 6326cb7View commit details -
Re-add contributor note and refactored fast tests fixed latents code …
…to remove CPU specific logic.
Configuration menu - View commit details
-
Copy full SHA for 6d0f321 - Browse repository at this point
Copy the full SHA 6d0f321View commit details -
Configuration menu - View commit details
-
Copy full SHA for 73504c4 - Browse repository at this point
Copy the full SHA 73504c4View commit details
Commits on May 24, 2023
-
- Updated the checkpoint ids to the new ids where appropriate - Refactored the UniDiffuserTextDecoder methods to return only tensors (and made other changes to support this) - Cleaned up the code following suggestions by patrickvonplaten
Configuration menu - View commit details
-
Copy full SHA for 0ed1857 - Browse repository at this point
Copy the full SHA 0ed1857View commit details -
Configuration menu - View commit details
-
Copy full SHA for d53026d - Browse repository at this point
Copy the full SHA d53026dView commit details -
Remove padding logic from UniDiffuserTextDecoder.generate_beam since …
…the inputs are already padded to a consistent length.
Configuration menu - View commit details
-
Copy full SHA for 0adb0a8 - Browse repository at this point
Copy the full SHA 0adb0a8View commit details -
Update checkpoint id for small test v1 checkpoint to hf-internal-test…
…ing/unidiffuser-test-v1.
Configuration menu - View commit details
-
Copy full SHA for 43b8894 - Browse repository at this point
Copy the full SHA 43b8894View commit details -
Configuration menu - View commit details
-
Copy full SHA for a5a9dac - Browse repository at this point
Copy the full SHA a5a9dacView commit details
Commits on May 25, 2023
-
Configuration menu - View commit details
-
Copy full SHA for d4b11aa - Browse repository at this point
Copy the full SHA d4b11aaView commit details -
Move ImageTextPipelineOutput documentation from /api/pipelines/unidif…
…fuser.mdx to /api/diffusion_pipeline.mdx.
Configuration menu - View commit details
-
Copy full SHA for 98ce17d - Browse repository at this point
Copy the full SHA 98ce17dView commit details
Commits on May 26, 2023
-
Configuration menu - View commit details
-
Copy full SHA for f8c325a - Browse repository at this point
Copy the full SHA f8c325aView commit details -
Configuration menu - View commit details
-
Copy full SHA for b4feac8 - Browse repository at this point
Copy the full SHA b4feac8View commit details -
Configuration menu - View commit details
-
Copy full SHA for 4f21661 - Browse repository at this point
Copy the full SHA 4f21661View commit details -
Configuration menu - View commit details
-
Copy full SHA for 07d68d7 - Browse repository at this point
Copy the full SHA 07d68d7View commit details