update

* Change step_offset scheduler docstrings

* Mention it may be needed by some models

* More docstrings

These ones failed literal S&R because I performed it case-sensitive
which is fun.

---------

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

.github/workflows/nightly_tests.yml

@@ -12,6 +12,7 @@ env:
   PYTEST_TIMEOUT: 600
   RUN_SLOW: yes
   RUN_NIGHTLY: yes
+  SLACK_API_TOKEN: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
 
 jobs:
   run_nightly_tests:
@@ -78,7 +79,8 @@ jobs:
           python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile \
             -s -v -k "not Flax and not Onnx" \
             --make-reports=tests_${{ matrix.config.report }} \
-            tests/
+            --report-log=${{ matrix.config.report }}.log \
+            tests/ 
 
       - name: Run nightly Flax TPU tests
         if: ${{ matrix.config.framework == 'flax' }}
@@ -89,6 +91,7 @@ jobs:
           python -m pytest -n 0 \
             -s -v -k "Flax" \
             --make-reports=tests_${{ matrix.config.report }} \
+            --report-log=${{ matrix.config.report }}.log \
             tests/
 
       - name: Run nightly ONNXRuntime CUDA tests
@@ -100,6 +103,7 @@ jobs:
           python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile \
             -s -v -k "Onnx" \
             --make-reports=tests_${{ matrix.config.report }} \
+            --report-log=${{ matrix.config.report }}.log \ 
             tests/
 
       - name: Failure short reports
@@ -112,6 +116,12 @@ jobs:
         with:
           name: ${{ matrix.config.report }}_test_reports
           path: reports
+      
+      - name: Generate Report and Notify Channel
+        if: always()
+        run: |
+          pip install slack_sdk tabulate
+          python scripts/log_reports.py >> $GITHUB_STEP_SUMMARY
 
   run_nightly_tests_apple_m1:
     name: Nightly PyTorch MPS tests on MacOS
@@ -152,7 +162,9 @@ jobs:
           HF_HOME: /System/Volumes/Data/mnt/cache
           HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
         run: |
-          ${CONDA_RUN} python -m pytest -n 1 -s -v --make-reports=tests_torch_mps tests/
+          ${CONDA_RUN} python -m pytest -n 1 -s -v --make-reports=tests_torch_mps \
+            --report-log=tests_torch_mps.log \
+            tests/
 
       - name: Failure short reports
         if: ${{ failure() }}
@@ -164,3 +176,9 @@ jobs:
         with:
           name: torch_mps_test_reports
           path: reports
+
+      - name: Generate Report and Notify Channel
+        if: always()
+        run: |
+          pip install slack_sdk tabulate
+          python scripts/log_reports.py >> $GITHUB_STEP_SUMMARY

.github/workflows/notify_slack_about_release.yml

@@ -0,0 +1,23 @@
+name: Notify Slack about a release
+
+on:
+  workflow_dispatch:
+  release:
+    types: [published]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v3
+    
+    - name: Setup Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.8'
+    
+    - name: Notify Slack about the release
+      env:
+        SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}
+      run: pip install requests && python utils/notify_slack_about_release.py

.github/workflows/pypi_publish.yaml

@@ -0,0 +1,79 @@
+# Adapted from https://blog.deepjyoti30.dev/pypi-release-github-action
+
+name: PyPI release
+
+on:
+  workflow_dispatch:
+  push:
+    tags:
+      - "*"
+
+jobs:
+  find-and-checkout-latest-branch:
+    runs-on: ubuntu-latest
+    outputs:
+      latest_branch: ${{ steps.set_latest_branch.outputs.latest_branch }}
+    steps:
+      - name: Checkout Repo
+        uses: actions/checkout@v3
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.8'
+
+      - name: Fetch latest branch
+        id: fetch_latest_branch
+        run: |
+          pip install -U requests packaging
+          LATEST_BRANCH=$(python utils/fetch_latest_release_branch.py)
+          echo "Latest branch: $LATEST_BRANCH"
+          echo "latest_branch=$LATEST_BRANCH" >> $GITHUB_ENV
+          
+      - name: Set latest branch output
+        id: set_latest_branch
+        run: echo "::set-output name=latest_branch::${{ env.latest_branch }}"
+
+  release:
+    needs: find-and-checkout-latest-branch
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout Repo
+        uses: actions/checkout@v3
+        with:
+          ref: ${{ needs.find-and-checkout-latest-branch.outputs.latest_branch }}
+          
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.8"
+      
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -U setuptools wheel twine
+      
+      - name: Build the dist files
+        run: python setup.py bdist_wheel && python setup.py sdist
+      
+      - name: Publish to the test PyPI
+        env:
+          TWINE_USERNAME: ${{ secrets.TEST_PYPI_USERNAME }}
+          TWINE_PASSWORD: ${{ secrets.TEST_PYPI_PASSWORD }}
+        run: twine upload dist/* -r pypitest --repository-url=https://test.pypi.org/legacy/    
+
+      - name: Test installing diffusers and importing
+        run: |
+          pip install diffusers && pip uninstall diffusers -y
+          pip install -i https://testpypi.python.org/pypi diffusers
+          python -c "from diffusers import __version__; print(__version__)"
+          python -c "from diffusers import DiffusionPipeline; pipe = DiffusionPipeline.from_pretrained('fusing/unet-ldm-dummy-update'); pipe()"
+          python -c "from diffusers import DiffusionPipeline; pipe = DiffusionPipeline.from_pretrained('hf-internal-testing/tiny-stable-diffusion-pipe', safety_checker=None); pipe('ah suh du')"
+          python -c "from diffusers import *"
+
+      - name: Publish to PyPI
+        env:
+          TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
+          TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
+        run: twine upload dist/* -r pypi

docs/source/en/_toctree.yml

@@ -104,6 +104,8 @@
       title: Latent Consistency Model-LoRA
     - local: using-diffusers/inference_with_lcm
       title: Latent Consistency Model
+    - local: using-diffusers/inference_with_tcd_lora
+      title: Trajectory Consistency Distillation-LoRA
     - local: using-diffusers/svd
       title: Stable Video Diffusion
     title: Specific pipeline examples
@@ -304,6 +306,8 @@
       title: Latent Consistency Models
     - local: api/pipelines/latent_diffusion
       title: Latent Diffusion
+    - local: api/pipelines/ledits_pp
+      title: LEDITS++
     - local: api/pipelines/panorama
       title: MultiDiffusion
     - local: api/pipelines/musicldm

docs/source/en/api/pipelines/ledits_pp.md

@@ -0,0 +1,54 @@
+<!--Copyright 2023 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# LEDITS++
+
+LEDITS++ was proposed in [LEDITS++: Limitless Image Editing using Text-to-Image Models](https://huggingface.co/papers/2311.16711) by Manuel Brack, Felix Friedrich, Katharina Kornmeier, Linoy Tsaban, Patrick Schramowski, Kristian Kersting, Apolinário Passos.
+
+The abstract from the paper is:
+
+*Text-to-image diffusion models have recently received increasing interest for their astonishing ability to produce high-fidelity images from solely text inputs. Subsequent research efforts aim to exploit and apply their capabilities to real image editing. However, existing image-to-image methods are often inefficient, imprecise, and of limited versatility. They either require time-consuming fine-tuning, deviate unnecessarily strongly from the input image, and/or lack support for multiple, simultaneous edits. To address these issues, we introduce LEDITS++, an efficient yet versatile and precise textual image manipulation technique. LEDITS++'s novel inversion approach requires no tuning nor optimization and produces high-fidelity results with a few diffusion steps. Second, our methodology supports multiple simultaneous edits and is architecture-agnostic. Third, we use a novel implicit masking technique that limits changes to relevant image regions. We propose the novel TEdBench++ benchmark as part of our exhaustive evaluation. Our results demonstrate the capabilities of LEDITS++ and its improvements over previous methods. The project page is available at https://leditsplusplus-project.static.hf.space .*
+
+<Tip>
+
+You can find additional information about LEDITS++ on the [project page](https://leditsplusplus-project.static.hf.space/index.html) and try it out in a [demo](https://huggingface.co/spaces/editing-images/leditsplusplus).
+
+</Tip>
+
+<Tip warning={true}>
+Due to some backward compatability issues with the current diffusers implementation of [`~schedulers.DPMSolverMultistepScheduler`] this implementation of LEdits++ can no longer guarantee perfect inversion. 
+This issue is unlikely to have any noticeable effects on applied use-cases. However, we provide an alternative implementation that guarantees perfect inversion in a dedicated [GitHub repo](https://github.com/ml-research/ledits_pp). 
+</Tip>
+
+We provide two distinct pipelines based on different pre-trained models. 
+
+## LEditsPPPipelineStableDiffusion
+[[autodoc]] pipelines.ledits_pp.LEditsPPPipelineStableDiffusion
+	- all
+	- __call__
+	- invert
+
+## LEditsPPPipelineStableDiffusionXL
+[[autodoc]] pipelines.ledits_pp.LEditsPPPipelineStableDiffusionXL
+	- all
+	- __call__
+	- invert
+
+
+
+## LEditsPPDiffusionPipelineOutput
+[[autodoc]] pipelines.ledits_pp.pipeline_output.LEditsPPDiffusionPipelineOutput
+	- all
+
+## LEditsPPInversionPipelineOutput
+[[autodoc]] pipelines.ledits_pp.pipeline_output.LEditsPPInversionPipelineOutput
+	- all

docs/source/en/api/pipelines/overview.md

@@ -57,6 +57,7 @@ The table below lists all the pipelines currently available in 🤗 Diffusers an
 | [Latent Consistency Models](latent_consistency_models) | text2image |
 | [Latent Diffusion](latent_diffusion) | text2image, super-resolution |
 | [LDM3D](stable_diffusion/ldm3d_diffusion) | text2image, text-to-3D, text-to-pano, upscaling |
+| [LEDITS++](ledits_pp) | image editing |
 | [MultiDiffusion](panorama) | text2image |
 | [MusicLDM](musicldm) | text2audio |
 | [Paint by Example](paint_by_example) | inpainting |

docs/source/en/api/pipelines/semantic_stable_diffusion.md

@@ -30,6 +30,6 @@ Make sure to check out the Schedulers [guide](../../using-diffusers/schedulers)
 	- all
 	- __call__
 
-## StableDiffusionSafePipelineOutput
+## SemanticStableDiffusionPipelineOutput
 [[autodoc]] pipelines.semantic_stable_diffusion.pipeline_output.SemanticStableDiffusionPipelineOutput
 	- all

docs/source/en/using-diffusers/inference_with_tcd_lora.md

@@ -0,0 +1,438 @@
+<!--Copyright 2024 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+[[open-in-colab]]
+
+# Trajectory Consistency Distillation-LoRA
+
+Trajectory Consistency Distillation (TCD) enables a model to generate higher quality and more detailed images with fewer steps. Moreover, owing to the effective error mitigation during the distillation process, TCD demonstrates superior performance even under conditions of large inference steps.
+
+The major advantages of TCD are:
+
+- Better than Teacher: TCD demonstrates superior generative quality at both small and large inference steps and exceeds the performance of [DPM-Solver++(2S)](../../api/schedulers/multistep_dpm_solver) with Stable Diffusion XL (SDXL). There is no additional discriminator or LPIPS supervision included during TCD training.
+
+- Flexible Inference Steps: The inference steps for TCD sampling can be freely adjusted without adversely affecting the image quality.
+
+- Freely change detail level: During inference, the level of detail in the image can be adjusted with a single hyperparameter, *gamma*.
+
+> [!TIP]
+> For more technical details of TCD, please refer to the [paper](https://arxiv.org/abs/2402.19159) or official [project page](https://mhh0318.github.io/tcd/)).
+
+For large models like SDXL, TCD is trained with [LoRA](https://huggingface.co/docs/peft/conceptual_guides/adapter#low-rank-adaptation-lora) to reduce memory usage. This is also useful because you can reuse LoRAs between different finetuned models, as long as they share the same base model, without further training.
+
+
+
+This guide will show you how to perform inference with TCD-LoRAs for a variety of tasks like text-to-image and inpainting, as well as how you can easily combine TCD-LoRAs with other adapters. Choose one of the supported base model and it's corresponding TCD-LoRA checkpoint from the table below to get started.
+
+| Base model                                                                                      | TCD-LoRA checkpoint                                            |
+|-------------------------------------------------------------------------------------------------|----------------------------------------------------------------|
+| [stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5)                  | [TCD-SD15](https://huggingface.co/h1t/TCD-SD15-LoRA)           |
+| [stable-diffusion-2-1-base](https://huggingface.co/stabilityai/stable-diffusion-2-1-base)       | [TCD-SD21-base](https://huggingface.co/h1t/TCD-SD21-base-LoRA) |
+| [stable-diffusion-xl-base-1.0](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0) | [TCD-SDXL](https://huggingface.co/h1t/TCD-SDXL-LoRA)           |
+
+
+Make sure you have [PEFT](https://github.com/huggingface/peft) installed for better LoRA support.
+
+```bash
+pip install -U peft
+```
+
+## General tasks
+
+In this guide, let's use the [`StableDiffusionXLPipeline`] and the [`TCDScheduler`]. Use the [`~StableDiffusionPipeline.load_lora_weights`] method to load the SDXL-compatible TCD-LoRA weights.
+
+A few tips to keep in mind for TCD-LoRA inference are to:
+
+- Keep the `num_inference_steps` between 4 and 50
+- Set `eta` (used to control stochasticity at each step) between 0 and 1. You should use a higher `eta` when increasing the number of inference steps, but the downside is that a larger `eta` in [`TCDScheduler`] leads to blurrier images. A value of 0.3 is recommended to produce good results.
+
+<hfoptions id="tasks">
+<hfoption id="text-to-image">
+
+```python
+import torch
+from diffusers import StableDiffusionXLPipeline, TCDScheduler
+
+device = "cuda"
+base_model_id = "stabilityai/stable-diffusion-xl-base-1.0"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+
+pipe = StableDiffusionXLPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, variant="fp16").to(device)
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id)
+pipe.fuse_lora()
+
+prompt = "Painting of the orange cat Otto von Garfield, Count of Bismarck-Schönhausen, Duke of Lauenburg, Minister-President of Prussia. Depicted wearing a Prussian Pickelhaube and eating his favorite meal - lasagna."
+
+image = pipe(
+    prompt=prompt,
+    num_inference_steps=4,
+    guidance_scale=0,
+    eta=0.3, 
+    generator=torch.Generator(device=device).manual_seed(0),
+).images[0]
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/demo_image.png)
+
+</hfoption>
+
+<hfoption id="inpainting">
+
+```python
+import torch
+from diffusers import AutoPipelineForInpainting, TCDScheduler
+from diffusers.utils import load_image, make_image_grid
+
+device = "cuda"
+base_model_id = "diffusers/stable-diffusion-xl-1.0-inpainting-0.1"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+
+pipe = AutoPipelineForInpainting.from_pretrained(base_model_id, torch_dtype=torch.float16, variant="fp16").to(device)
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id)
+pipe.fuse_lora()
+
+img_url = "https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo.png"
+mask_url = "https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo_mask.png"
+
+init_image = load_image(img_url).resize((1024, 1024))
+mask_image = load_image(mask_url).resize((1024, 1024))
+
+prompt = "a tiger sitting on a park bench"
+
+image = pipe(
+  prompt=prompt,
+  image=init_image,
+  mask_image=mask_image,
+  num_inference_steps=8,
+  guidance_scale=0,
+  eta=0.3,
+  strength=0.99,  # make sure to use `strength` below 1.0
+  generator=torch.Generator(device=device).manual_seed(0),
+).images[0]
+
+grid_image = make_image_grid([init_image, mask_image, image], rows=1, cols=3)
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/inpainting_tcd.png)
+
+
+</hfoption>
+</hfoptions>
+
+## Community models
+
+TCD-LoRA also works with many community finetuned models and plugins. For example, load the [animagine-xl-3.0](https://huggingface.co/cagliostrolab/animagine-xl-3.0) checkpoint which is a community finetuned version of SDXL for generating anime images.
+
+```python
+import torch
+from diffusers import StableDiffusionXLPipeline, TCDScheduler
+
+device = "cuda"
+base_model_id = "cagliostrolab/animagine-xl-3.0"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+
+pipe = StableDiffusionXLPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, variant="fp16").to(device)
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id)
+pipe.fuse_lora()
+
+prompt = "A man, clad in a meticulously tailored military uniform, stands with unwavering resolve. The uniform boasts intricate details, and his eyes gleam with determination. Strands of vibrant, windswept hair peek out from beneath the brim of his cap."
+
+image = pipe(
+    prompt=prompt,
+    num_inference_steps=8,
+    guidance_scale=0,
+    eta=0.3, 
+    generator=torch.Generator(device=device).manual_seed(0),
+).images[0]
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/animagine_xl.png)
+
+TCD-LoRA also supports other LoRAs trained on different styles. For example, let's load the [TheLastBen/Papercut_SDXL](https://huggingface.co/TheLastBen/Papercut_SDXL) LoRA and fuse it with the TCD-LoRA with the [`~loaders.UNet2DConditionLoadersMixin.set_adapters`] method. 
+
+> [!TIP]
+> Check out the [Merge LoRAs](merge_loras) guide to learn more about efficient merging methods.
+
+```python
+import torch
+from diffusers import StableDiffusionXLPipeline
+from scheduling_tcd import TCDScheduler 
+
+device = "cuda"
+base_model_id = "stabilityai/stable-diffusion-xl-base-1.0"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+styled_lora_id = "TheLastBen/Papercut_SDXL"
+
+pipe = StableDiffusionXLPipeline.from_pretrained(base_model_id, torch_dtype=torch.float16, variant="fp16").to(device)
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id, adapter_name="tcd")
+pipe.load_lora_weights(styled_lora_id, adapter_name="style")
+pipe.set_adapters(["tcd", "style"], adapter_weights=[1.0, 1.0])
+
+prompt = "papercut of a winter mountain, snow"
+
+image = pipe(
+    prompt=prompt,
+    num_inference_steps=4,
+    guidance_scale=0,
+    eta=0.3, 
+    generator=torch.Generator(device=device).manual_seed(0),
+).images[0]
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/styled_lora.png)
+
+
+## Adapters
+
+TCD-LoRA is very versatile, and it can be combined with other adapter types like ControlNets, IP-Adapter, and AnimateDiff.
+
+<hfoptions id="adapters">
+<hfoption id="ControlNet">
+
+### Depth ControlNet
+
+```python
+import torch
+import numpy as np
+from PIL import Image
+from transformers import DPTFeatureExtractor, DPTForDepthEstimation
+from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline
+from diffusers.utils import load_image, make_image_grid
+from scheduling_tcd import TCDScheduler 
+
+device = "cuda"
+depth_estimator = DPTForDepthEstimation.from_pretrained("Intel/dpt-hybrid-midas").to(device)
+feature_extractor = DPTFeatureExtractor.from_pretrained("Intel/dpt-hybrid-midas")
+
+def get_depth_map(image):
+    image = feature_extractor(images=image, return_tensors="pt").pixel_values.to(device)
+    with torch.no_grad(), torch.autocast(device):
+        depth_map = depth_estimator(image).predicted_depth
+
+    depth_map = torch.nn.functional.interpolate(
+        depth_map.unsqueeze(1),
+        size=(1024, 1024),
+        mode="bicubic",
+        align_corners=False,
+    )
+    depth_min = torch.amin(depth_map, dim=[1, 2, 3], keepdim=True)
+    depth_max = torch.amax(depth_map, dim=[1, 2, 3], keepdim=True)
+    depth_map = (depth_map - depth_min) / (depth_max - depth_min)
+    image = torch.cat([depth_map] * 3, dim=1)
+
+    image = image.permute(0, 2, 3, 1).cpu().numpy()[0]
+    image = Image.fromarray((image * 255.0).clip(0, 255).astype(np.uint8))
+    return image
+
+base_model_id = "stabilityai/stable-diffusion-xl-base-1.0"
+controlnet_id = "diffusers/controlnet-depth-sdxl-1.0"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+
+controlnet = ControlNetModel.from_pretrained(
+    controlnet_id,
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to(device)
+pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
+    base_model_id,
+    controlnet=controlnet,
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to(device)
+pipe.enable_model_cpu_offload()
+
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id)
+pipe.fuse_lora()
+
+prompt = "stormtrooper lecture, photorealistic"
+
+image = load_image("https://huggingface.co/lllyasviel/sd-controlnet-depth/resolve/main/images/stormtrooper.png")
+depth_image = get_depth_map(image)
+
+controlnet_conditioning_scale = 0.5  # recommended for good generalization
+
+image = pipe(
+    prompt, 
+    image=depth_image, 
+    num_inference_steps=4, 
+    guidance_scale=0,
+    eta=0.3,
+    controlnet_conditioning_scale=controlnet_conditioning_scale,
+    generator=torch.Generator(device=device).manual_seed(0),
+).images[0]
+
+grid_image = make_image_grid([depth_image, image], rows=1, cols=2)
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/controlnet_depth_tcd.png)
+
+### Canny ControlNet
+```python
+import torch
+from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline
+from diffusers.utils import load_image, make_image_grid
+from scheduling_tcd import TCDScheduler 
+
+device = "cuda"
+base_model_id = "stabilityai/stable-diffusion-xl-base-1.0"
+controlnet_id = "diffusers/controlnet-canny-sdxl-1.0"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+
+controlnet = ControlNetModel.from_pretrained(
+    controlnet_id,
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to(device)
+pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
+    base_model_id,
+    controlnet=controlnet,
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to(device)
+pipe.enable_model_cpu_offload()
+
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id)
+pipe.fuse_lora()
+
+prompt = "ultrarealistic shot of a furry blue bird"
+
+canny_image = load_image("https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/bird_canny.png")
+
+controlnet_conditioning_scale = 0.5  # recommended for good generalization
+
+image = pipe(
+    prompt, 
+    image=canny_image, 
+    num_inference_steps=4, 
+    guidance_scale=0,
+    eta=0.3,
+    controlnet_conditioning_scale=controlnet_conditioning_scale,
+    generator=torch.Generator(device=device).manual_seed(0),
+).images[0]
+
+grid_image = make_image_grid([canny_image, image], rows=1, cols=2)
+```
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/controlnet_canny_tcd.png)
+
+<Tip>
+The inference parameters in this example might not work for all examples, so we recommend you to try different values for `num_inference_steps`, `guidance_scale`, `controlnet_conditioning_scale` and `cross_attention_kwargs` parameters and choose the best one. 
+</Tip>
+
+</hfoption>
+<hfoption id="IP-Adapter">
+
+This example shows how to use the TCD-LoRA with the [IP-Adapter](https://github.com/tencent-ailab/IP-Adapter/tree/main) and SDXL.
+
+```python
+import torch
+from diffusers import StableDiffusionXLPipeline
+from diffusers.utils import load_image, make_image_grid
+
+from ip_adapter import IPAdapterXL
+from scheduling_tcd import TCDScheduler 
+
+device = "cuda"
+base_model_path = "stabilityai/stable-diffusion-xl-base-1.0"
+image_encoder_path = "sdxl_models/image_encoder"
+ip_ckpt = "sdxl_models/ip-adapter_sdxl.bin"
+tcd_lora_id = "h1t/TCD-SDXL-LoRA"
+
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    base_model_path, 
+    torch_dtype=torch.float16, 
+    variant="fp16"
+)
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+pipe.load_lora_weights(tcd_lora_id)
+pipe.fuse_lora()
+
+ip_model = IPAdapterXL(pipe, image_encoder_path, ip_ckpt, device)
+
+ref_image = load_image("https://raw.githubusercontent.com/tencent-ailab/IP-Adapter/main/assets/images/woman.png").resize((512, 512))
+
+prompt = "best quality, high quality, wearing sunglasses"
+
+image = ip_model.generate(
+    pil_image=ref_image, 
+    prompt=prompt,
+    scale=0.5,
+    num_samples=1, 
+    num_inference_steps=4, 
+    guidance_scale=0,
+    eta=0.3, 
+    seed=0,
+)[0]
+
+grid_image = make_image_grid([ref_image, image], rows=1, cols=2)
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/ip_adapter.png)
+
+
+
+</hfoption>
+<hfoption id="AnimateDiff">
+
+[`AnimateDiff`] allows animating images using Stable Diffusion models. TCD-LoRA can substantially accelerate the process without degrading image quality. The quality of animation with TCD-LoRA and AnimateDiff has a more lucid outcome.
+
+```python
+import torch
+from diffusers import MotionAdapter, AnimateDiffPipeline, DDIMScheduler
+from scheduling_tcd import TCDScheduler
+from diffusers.utils import export_to_gif
+
+adapter = MotionAdapter.from_pretrained("guoyww/animatediff-motion-adapter-v1-5")
+pipe = AnimateDiffPipeline.from_pretrained(
+    "frankjoshua/toonyou_beta6",
+    motion_adapter=adapter,
+).to("cuda")
+
+# set TCDScheduler
+pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
+
+# load TCD LoRA
+pipe.load_lora_weights("h1t/TCD-SD15-LoRA", adapter_name="tcd")
+pipe.load_lora_weights("guoyww/animatediff-motion-lora-zoom-in", weight_name="diffusion_pytorch_model.safetensors", adapter_name="motion-lora")
+
+pipe.set_adapters(["tcd", "motion-lora"], adapter_weights=[1.0, 1.2])
+
+prompt = "best quality, masterpiece, 1girl, looking at viewer, blurry background, upper body, contemporary, dress"
+generator = torch.manual_seed(0)
+frames = pipe(
+    prompt=prompt,
+    num_inference_steps=5,
+    guidance_scale=0,
+    cross_attention_kwargs={"scale": 1},
+    num_frames=24,
+    eta=0.3,
+    generator=generator
+).frames[0]
+export_to_gif(frames, "animation.gif")
+```
+
+![](https://github.com/jabir-zheng/TCD/raw/main/assets/animation_example.gif)
+
+</hfoption>
+</hfoptions>

examples/community/latent_consistency_img2img.py

@@ -513,9 +513,7 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
             there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the alpha value at step 0.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen

examples/community/latent_consistency_txt2img.py

@@ -418,9 +418,7 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
             there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the alpha value at step 0.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen

examples/community/scheduling_ufogen.py

@@ -171,9 +171,7 @@ class UFOGenScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         rescale_betas_zero_snr (`bool`, defaults to `False`):
             Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and
             dark samples instead of limiting it to samples with medium brightness. Loosely related to

examples/custom_diffusion/train_custom_diffusion.py

@@ -1178,7 +1178,7 @@ def main(args):
                         grads_text_encoder = text_encoder.get_input_embeddings().weight.grad
                     # Get the index for tokens that we want to zero the grads for
                     index_grads_to_zero = torch.arange(len(tokenizer)) != modifier_token_id[0]
-                    for i in range(len(modifier_token_id[1:])):
+                    for i in range(1, len(modifier_token_id)):
                         index_grads_to_zero = index_grads_to_zero & (
                             torch.arange(len(tokenizer)) != modifier_token_id[i]
                         )

scripts/log_reports.py

@@ -0,0 +1,139 @@
+import argparse
+import json
+import os
+from datetime import date
+from pathlib import Path
+
+from slack_sdk import WebClient
+from tabulate import tabulate
+
+
+MAX_LEN_MESSAGE = 2900  # slack endpoint has a limit of 3001 characters
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--slack_channel_name", default="diffusers-ci-nightly")
+
+
+def main(slack_channel_name=None):
+    failed = []
+    passed = []
+
+    group_info = []
+
+    total_num_failed = 0
+    empty_file = False or len(list(Path().glob("*.log"))) == 0
+
+    total_empty_files = []
+
+    for log in Path().glob("*.log"):
+        section_num_failed = 0
+        i = 0
+        with open(log) as f:
+            for line in f:
+                line = json.loads(line)
+                i += 1
+                if line.get("nodeid", "") != "":
+                    test = line["nodeid"]
+                    if line.get("duration", None) is not None:
+                        duration = f'{line["duration"]:.4f}'
+                        if line.get("outcome", "") == "failed":
+                            section_num_failed += 1
+                            failed.append([test, duration, log.name.split("_")[0]])
+                            total_num_failed += 1
+                        else:
+                            passed.append([test, duration, log.name.split("_")[0]])
+            empty_file = i == 0
+        group_info.append([str(log), section_num_failed, failed])
+        total_empty_files.append(empty_file)
+        os.remove(log)
+        failed = []
+    text = (
+        "🌞 There were no failures!"
+        if not any(total_empty_files)
+        else "Something went wrong there is at least one empty file - please check GH action results."
+    )
+    no_error_payload = {
+        "type": "section",
+        "text": {
+            "type": "plain_text",
+            "text": text,
+            "emoji": True,
+        },
+    }
+
+    message = ""
+    payload = [
+        {
+            "type": "header",
+            "text": {
+                "type": "plain_text",
+                "text": "🤗 Results of the Diffusers scheduled nightly tests.",
+            },
+        },
+    ]
+    if total_num_failed > 0:
+        for i, (name, num_failed, failed_tests) in enumerate(group_info):
+            if num_failed > 0:
+                if num_failed == 1:
+                    message += f"*{name}: {num_failed} failed test*\n"
+                else:
+                    message += f"*{name}: {num_failed} failed tests*\n"
+                failed_table = []
+                for test in failed_tests:
+                    failed_table.append(test[0].split("::"))
+                failed_table = tabulate(
+                    failed_table,
+                    headers=["Test Location", "Test Case", "Test Name"],
+                    showindex="always",
+                    tablefmt="grid",
+                    maxcolwidths=[12, 12, 12],
+                )
+                message += "\n```\n" + failed_table + "\n```"
+
+            if total_empty_files[i]:
+                message += f"\n*{name}: Warning! Empty file - please check the GitHub action job *\n"
+        print(f"### {message}")
+    else:
+        payload.append(no_error_payload)
+
+    if len(message) > MAX_LEN_MESSAGE:
+        print(f"Truncating long message from {len(message)} to {MAX_LEN_MESSAGE}")
+        message = message[:MAX_LEN_MESSAGE] + "..."
+
+    if len(message) != 0:
+        md_report = {
+            "type": "section",
+            "text": {"type": "mrkdwn", "text": message},
+        }
+        payload.append(md_report)
+        action_button = {
+            "type": "section",
+            "text": {"type": "mrkdwn", "text": "*For more details:*"},
+            "accessory": {
+                "type": "button",
+                "text": {"type": "plain_text", "text": "Check Action results", "emoji": True},
+                "url": f"https://github.com/huggingface/diffusers/actions/runs/{os.environ['GITHUB_RUN_ID']}",
+            },
+        }
+        payload.append(action_button)
+
+    date_report = {
+        "type": "context",
+        "elements": [
+            {
+                "type": "plain_text",
+                "text": f"Nightly test results for {date.today()}",
+            },
+        ],
+    }
+    payload.append(date_report)
+
+    print(payload)
+
+    client = WebClient(token=os.environ.get("SLACK_API_TOKEN"))
+    client.chat_postMessage(channel=f"#{slack_channel_name}", text=message, blocks=payload)
+
+
+if __name__ == "__main__":
+    args = parser.parse_args()
+    main(args.slack_channel_name)

src/diffusers/__init__.py

@@ -253,6 +253,8 @@ else:
             "LatentConsistencyModelImg2ImgPipeline",
             "LatentConsistencyModelPipeline",
             "LDMTextToImagePipeline",
+            "LEditsPPPipelineStableDiffusion",
+            "LEditsPPPipelineStableDiffusionXL",
             "MusicLDMPipeline",
             "PaintByExamplePipeline",
             "PIAPipeline",
@@ -623,6 +625,8 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             LatentConsistencyModelImg2ImgPipeline,
             LatentConsistencyModelPipeline,
             LDMTextToImagePipeline,
+            LEditsPPPipelineStableDiffusion,
+            LEditsPPPipelineStableDiffusionXL,
             MusicLDMPipeline,
             PaintByExamplePipeline,
             PIAPipeline,

src/diffusers/loaders/single_file_utils.py

@@ -454,8 +454,7 @@ def set_image_size(pipeline_class_name, original_config, checkpoint, image_size=
     model_type = infer_model_type(original_config, checkpoint, model_type)
 
     if pipeline_class_name == "StableDiffusionUpscalePipeline":
-        image_size = original_config["model"]["params"]["unet_config"]["params"]["image_size"]
-        return image_size
+        return 512
 
     elif model_type in ["SDXL", "SDXL-Refiner", "Playground"]:
         image_size = 1024

src/diffusers/models/unets/unet_2d_condition.py

@@ -80,7 +80,7 @@ class UNet2DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin,
         in_channels (`int`, *optional*, defaults to 4): Number of channels in the input sample.
         out_channels (`int`, *optional*, defaults to 4): Number of channels in the output.
         center_input_sample (`bool`, *optional*, defaults to `False`): Whether to center the input sample.
-        flip_sin_to_cos (`bool`, *optional*, defaults to `False`):
+        flip_sin_to_cos (`bool`, *optional*, defaults to `True`):
             Whether to flip the sin to cos in the time embedding.
         freq_shift (`int`, *optional*, defaults to 0): The frequency shift to apply to the time embedding.
         down_block_types (`Tuple[str]`, *optional*, defaults to `("CrossAttnDownBlock2D", "CrossAttnDownBlock2D", "CrossAttnDownBlock2D", "DownBlock2D")`):
@@ -109,7 +109,7 @@ class UNet2DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin,
             The number of transformer blocks of type [`~models.attention.BasicTransformerBlock`]. Only relevant for
             [`~models.unet_2d_blocks.CrossAttnDownBlock2D`], [`~models.unet_2d_blocks.CrossAttnUpBlock2D`],
             [`~models.unet_2d_blocks.UNetMidBlock2DCrossAttn`].
-       reverse_transformer_layers_per_block : (`Tuple[Tuple]`, *optional*, defaults to None):
+        reverse_transformer_layers_per_block : (`Tuple[Tuple]`, *optional*, defaults to None):
             The number of transformer blocks of type [`~models.attention.BasicTransformerBlock`], in the upsampling
             blocks of the U-Net. Only relevant if `transformer_layers_per_block` is of type `Tuple[Tuple]` and for
             [`~models.unet_2d_blocks.CrossAttnDownBlock2D`], [`~models.unet_2d_blocks.CrossAttnUpBlock2D`],
@@ -147,9 +147,9 @@ class UNet2DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin,
             The second activation function to use in timestep embedding. Choose from `silu`, `mish` and `gelu`.
         time_cond_proj_dim (`int`, *optional*, defaults to `None`):
             The dimension of `cond_proj` layer in the timestep embedding.
-        conv_in_kernel (`int`, *optional*, default to `3`): The kernel size of `conv_in` layer. conv_out_kernel (`int`,
-        *optional*, default to `3`): The kernel size of `conv_out` layer. projection_class_embeddings_input_dim (`int`,
-        *optional*): The dimension of the `class_labels` input when
+        conv_in_kernel (`int`, *optional*, default to `3`): The kernel size of `conv_in` layer.
+        conv_out_kernel (`int`, *optional*, default to `3`): The kernel size of `conv_out` layer.
+        projection_class_embeddings_input_dim (`int`, *optional*): The dimension of the `class_labels` input when
             `class_embed_type="projection"`. Required when `class_embed_type="projection"`.
         class_embeddings_concat (`bool`, *optional*, defaults to `False`): Whether to concatenate the time
             embeddings with the class embeddings.

src/diffusers/pipelines/__init__.py

@@ -23,6 +23,7 @@ _import_structure = {
     "controlnet_xs": [],
     "deprecated": [],
     "latent_diffusion": [],
+    "ledits_pp": [],
     "stable_diffusion": [],
     "stable_diffusion_xl": [],
 }
@@ -171,6 +172,12 @@ else:
         "LatentConsistencyModelPipeline",
     ]
     _import_structure["latent_diffusion"].extend(["LDMTextToImagePipeline"])
+    _import_structure["ledits_pp"].extend(
+        [
+            "LEditsPPPipelineStableDiffusion",
+            "LEditsPPPipelineStableDiffusionXL",
+        ]
+    )
     _import_structure["musicldm"] = ["MusicLDMPipeline"]
     _import_structure["paint_by_example"] = ["PaintByExamplePipeline"]
     _import_structure["pia"] = ["PIAPipeline"]
@@ -424,6 +431,12 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
             LatentConsistencyModelPipeline,
         )
         from .latent_diffusion import LDMTextToImagePipeline
+        from .ledits_pp import (
+            LEditsPPDiffusionPipelineOutput,
+            LEditsPPInversionPipelineOutput,
+            LEditsPPPipelineStableDiffusion,
+            LEditsPPPipelineStableDiffusionXL,
+        )
         from .musicldm import MusicLDMPipeline
         from .paint_by_example import PaintByExamplePipeline
         from .pia import PIAPipeline

src/diffusers/pipelines/ledits_pp/__init__.py

@@ -0,0 +1,55 @@
+from typing import TYPE_CHECKING
+
+from ...utils import (
+    DIFFUSERS_SLOW_IMPORT,
+    OptionalDependencyNotAvailable,
+    _LazyModule,
+    get_objects_from_module,
+    is_torch_available,
+    is_transformers_available,
+)
+
+
+_dummy_objects = {}
+_import_structure = {}
+
+try:
+    if not (is_transformers_available() and is_torch_available()):
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    from ...utils import dummy_torch_and_transformers_objects  # noqa F403
+
+    _dummy_objects.update(get_objects_from_module(dummy_torch_and_transformers_objects))
+else:
+    _import_structure["pipeline_leditspp_stable_diffusion"] = ["LEditsPPPipelineStableDiffusion"]
+    _import_structure["pipeline_leditspp_stable_diffusion_xl"] = ["LEditsPPPipelineStableDiffusionXL"]
+
+    _import_structure["pipeline_output"] = ["LEditsPPDiffusionPipelineOutput", "LEditsPPDiffusionPipelineOutput"]
+
+if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
+    try:
+        if not (is_transformers_available() and is_torch_available()):
+            raise OptionalDependencyNotAvailable()
+
+    except OptionalDependencyNotAvailable:
+        from ...utils.dummy_torch_and_transformers_objects import *
+    else:
+        from .pipeline_leditspp_stable_diffusion import (
+            LEditsPPDiffusionPipelineOutput,
+            LEditsPPInversionPipelineOutput,
+            LEditsPPPipelineStableDiffusion,
+        )
+        from .pipeline_leditspp_stable_diffusion_xl import LEditsPPPipelineStableDiffusionXL
+
+else:
+    import sys
+
+    sys.modules[__name__] = _LazyModule(
+        __name__,
+        globals()["__file__"],
+        _import_structure,
+        module_spec=__spec__,
+    )
+
+    for name, value in _dummy_objects.items():
+        setattr(sys.modules[__name__], name, value)

src/diffusers/pipelines/ledits_pp/pipeline_output.py

@@ -0,0 +1,43 @@
+from dataclasses import dataclass
+from typing import List, Optional, Union
+
+import numpy as np
+import PIL.Image
+
+from ...utils import BaseOutput
+
+
+@dataclass
+class LEditsPPDiffusionPipelineOutput(BaseOutput):
+    """
+    Output class for LEdits++ Diffusion pipelines.
+
+    Args:
+        images (`List[PIL.Image.Image]` or `np.ndarray`)
+            List of denoised PIL images of length `batch_size` or NumPy array of shape `(batch_size, height, width,
+            num_channels)`.
+        nsfw_content_detected (`List[bool]`)
+            List indicating whether the corresponding generated image contains “not-safe-for-work” (nsfw) content or
+            `None` if safety checking could not be performed.
+    """
+
+    images: Union[List[PIL.Image.Image], np.ndarray]
+    nsfw_content_detected: Optional[List[bool]]
+
+
+@dataclass
+class LEditsPPInversionPipelineOutput(BaseOutput):
+    """
+    Output class for LEdits++ Diffusion pipelines.
+
+    Args:
+        input_images (`List[PIL.Image.Image]` or `np.ndarray`)
+            List of the cropped and resized input images as PIL images of length `batch_size` or NumPy array of shape `
+            (batch_size, height, width, num_channels)`.
+        vae_reconstruction_images (`List[PIL.Image.Image]` or `np.ndarray`)
+            List of VAE reconstruction of all input images as PIL images of length `batch_size` or NumPy array of shape `
+            (batch_size, height, width, num_channels)`.
+    """
+
+    images: Union[List[PIL.Image.Image], np.ndarray]
+    vae_reconstruction_images: Union[List[PIL.Image.Image], np.ndarray]

src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py

@@ -758,6 +758,7 @@ class StableDiffusionImg2ImgPipeline(
             init_latents = torch.cat([init_latents], dim=0)
 
         shape = init_latents.shape
+        print(shape)
         noise = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
 
         # get latents

src/diffusers/schedulers/scheduling_ddim.py

@@ -157,9 +157,7 @@ class DDIMScheduler(SchedulerMixin, ConfigMixin):
             there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the alpha value at step 0.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen

src/diffusers/schedulers/scheduling_ddim_flax.py

@@ -93,9 +93,7 @@ class FlaxDDIMScheduler(FlaxSchedulerMixin, ConfigMixin):
             step there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the value of alpha at step 0.
         steps_offset (`int`, default `0`):
-            an offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False`, to make the last step use step 0 for the previous alpha product, as done in
-            stable diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, default `epsilon`):
             indicates whether the model predicts the noise (epsilon), or the samples. One of `epsilon`, `sample`.
             `v-prediction` is not supported for this scheduler.

src/diffusers/schedulers/scheduling_ddim_inverse.py

@@ -155,9 +155,7 @@ class DDIMInverseScheduler(SchedulerMixin, ConfigMixin):
             there is no previous alpha. When this option is `True` the previous alpha product is fixed to 0, otherwise
             it uses the alpha value at step `num_train_timesteps - 1`.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use `num_train_timesteps - 1` for the previous alpha
-            product.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen

src/diffusers/schedulers/scheduling_ddim_parallel.py

@@ -159,9 +159,7 @@ class DDIMParallelScheduler(SchedulerMixin, ConfigMixin):
             step there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the value of alpha at step 0.
         steps_offset (`int`, default `0`):
-            an offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False`, to make the last step use step 0 for the previous alpha product, as done in
-            stable diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, default `epsilon`, optional):
             prediction type of the scheduler function, one of `epsilon` (predicting the noise of the diffusion
             process), `sample` (directly predicting the noisy sample`) or `v_prediction` (see section 2.4

src/diffusers/schedulers/scheduling_ddpm.py

@@ -167,9 +167,7 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         rescale_betas_zero_snr (`bool`, defaults to `False`):
             Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and
             dark samples instead of limiting it to samples with medium brightness. Loosely related to

src/diffusers/schedulers/scheduling_ddpm_parallel.py

@@ -173,9 +173,7 @@ class DDPMParallelScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2. of [Common Diffusion Noise Schedules and Sample
             Steps are Flawed](https://arxiv.org/abs/2305.08891) for more information.
         steps_offset (`int`, default `0`):
-            an offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False`, to make the last step use step 0 for the previous alpha product, as done in
-            stable diffusion.
+            An offset added to the inference steps, as required by some model families.
         rescale_betas_zero_snr (`bool`, defaults to `False`):
             Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and
             dark samples instead of limiting it to samples with medium brightness. Loosely related to

src/diffusers/schedulers/scheduling_deis_multistep.py

@@ -115,9 +115,7 @@ class DEISMultistepScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_dpmsolver_multistep.py

@@ -178,9 +178,7 @@ class DPMSolverMultistepScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         rescale_betas_zero_snr (`bool`, defaults to `False`):
             Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and
             dark samples instead of limiting it to samples with medium brightness. Loosely related to
@@ -899,6 +897,7 @@ class DPMSolverMultistepScheduler(SchedulerMixin, ConfigMixin):
         timestep: int,
         sample: torch.FloatTensor,
         generator=None,
+        variance_noise: Optional[torch.FloatTensor] = None,
         return_dict: bool = True,
     ) -> Union[SchedulerOutput, Tuple]:
         """
@@ -914,6 +913,9 @@ class DPMSolverMultistepScheduler(SchedulerMixin, ConfigMixin):
                 A current instance of a sample created by the diffusion process.
             generator (`torch.Generator`, *optional*):
                 A random number generator.
+            variance_noise (`torch.FloatTensor`):
+                Alternative to generating noise with `generator` by directly providing the noise for the variance
+                itself. Useful for methods such as [`LEdits++`].
             return_dict (`bool`):
                 Whether or not to return a [`~schedulers.scheduling_utils.SchedulerOutput`] or `tuple`.
 
@@ -948,11 +950,12 @@ class DPMSolverMultistepScheduler(SchedulerMixin, ConfigMixin):
 
         # Upcast to avoid precision issues when computing prev_sample
         sample = sample.to(torch.float32)
-
-        if self.config.algorithm_type in ["sde-dpmsolver", "sde-dpmsolver++"]:
+        if self.config.algorithm_type in ["sde-dpmsolver", "sde-dpmsolver++"] and variance_noise is None:
             noise = randn_tensor(
                 model_output.shape, generator=generator, device=model_output.device, dtype=torch.float32
             )
+        elif self.config.algorithm_type in ["sde-dpmsolver", "sde-dpmsolver++"]:
+            noise = variance_noise.to(device=model_output.device, dtype=torch.float32)
         else:
             noise = None
 

src/diffusers/schedulers/scheduling_dpmsolver_multistep_inverse.py

@@ -134,9 +134,7 @@ class DPMSolverMultistepInverseScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]
@@ -792,6 +790,7 @@ class DPMSolverMultistepInverseScheduler(SchedulerMixin, ConfigMixin):
         timestep: int,
         sample: torch.FloatTensor,
         generator=None,
+        variance_noise: Optional[torch.FloatTensor] = None,
         return_dict: bool = True,
     ) -> Union[SchedulerOutput, Tuple]:
         """
@@ -807,6 +806,9 @@ class DPMSolverMultistepInverseScheduler(SchedulerMixin, ConfigMixin):
                 A current instance of a sample created by the diffusion process.
             generator (`torch.Generator`, *optional*):
                 A random number generator.
+            variance_noise (`torch.FloatTensor`):
+                Alternative to generating noise with `generator` by directly providing the noise for the variance
+                itself. Useful for methods such as [`CycleDiffusion`].
             return_dict (`bool`):
                 Whether or not to return a [`~schedulers.scheduling_utils.SchedulerOutput`] or `tuple`.
 
@@ -837,10 +839,12 @@ class DPMSolverMultistepInverseScheduler(SchedulerMixin, ConfigMixin):
             self.model_outputs[i] = self.model_outputs[i + 1]
         self.model_outputs[-1] = model_output
 
-        if self.config.algorithm_type in ["sde-dpmsolver", "sde-dpmsolver++"]:
+        if self.config.algorithm_type in ["sde-dpmsolver", "sde-dpmsolver++"] and variance_noise is None:
             noise = randn_tensor(
                 model_output.shape, generator=generator, device=model_output.device, dtype=model_output.dtype
             )
+        elif self.config.algorithm_type in ["sde-dpmsolver", "sde-dpmsolver++"]:
+            noise = variance_noise
         else:
             noise = None
 

src/diffusers/schedulers/scheduling_dpmsolver_sde.py

@@ -153,9 +153,7 @@ class DPMSolverSDEScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_euler_ancestral_discrete.py

@@ -156,9 +156,7 @@ class EulerAncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         rescale_betas_zero_snr (`bool`, defaults to `False`):
             Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and
             dark samples instead of limiting it to samples with medium brightness. Loosely related to

src/diffusers/schedulers/scheduling_euler_discrete.py

@@ -162,9 +162,7 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         rescale_betas_zero_snr (`bool`, defaults to `False`):
             Whether to rescale the betas to have zero terminal SNR. This enables the model to generate very bright and
             dark samples instead of limiting it to samples with medium brightness. Loosely related to

src/diffusers/schedulers/scheduling_heun_discrete.py

@@ -101,9 +101,7 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_k_dpm_2_ancestral_discrete.py

@@ -99,9 +99,7 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_k_dpm_2_discrete.py

@@ -98,9 +98,7 @@ class KDPM2DiscreteScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_lcm.py

@@ -165,9 +165,7 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
             there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the alpha value at step 0.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen

src/diffusers/schedulers/scheduling_lms_discrete.py

@@ -119,9 +119,7 @@ class LMSDiscreteScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_pndm.py

@@ -104,9 +104,7 @@ class PNDMScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_pndm_flax.py

@@ -99,9 +99,7 @@ class FlaxPNDMScheduler(FlaxSchedulerMixin, ConfigMixin):
             step there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the value of alpha at step 0.
         steps_offset (`int`, default `0`):
-            an offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False`, to make the last step use step 0 for the previous alpha product, as done in
-            stable diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, default `epsilon`, optional):
             prediction type of the scheduler function, one of `epsilon` (predicting the noise of the diffusion
             process), `sample` (directly predicting the noisy sample`) or `v_prediction` (see section 2.4

src/diffusers/schedulers/scheduling_sasolver.py

@@ -131,9 +131,7 @@ class SASolverScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/schedulers/scheduling_tcd.py

@@ -166,9 +166,7 @@ class TCDScheduler(SchedulerMixin, ConfigMixin):
             there is no previous alpha. When this option is `True` the previous alpha product is fixed to `1`,
             otherwise it uses the alpha value at step 0.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen

src/diffusers/schedulers/scheduling_unipc_multistep.py

@@ -126,9 +126,7 @@ class UniPCMultistepScheduler(SchedulerMixin, ConfigMixin):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
         steps_offset (`int`, defaults to 0):
-            An offset added to the inference steps. You can use a combination of `offset=1` and
-            `set_alpha_to_one=False` to make the last step use step 0 for the previous alpha product like in Stable
-            Diffusion.
+            An offset added to the inference steps, as required by some model families.
     """
 
     _compatibles = [e.name for e in KarrasDiffusionSchedulers]

src/diffusers/utils/dummy_torch_and_transformers_objects.py

@@ -647,6 +647,36 @@ class LDMTextToImagePipeline(metaclass=DummyObject):
         requires_backends(cls, ["torch", "transformers"])
 
 
+class LEditsPPPipelineStableDiffusion(metaclass=DummyObject):
+    _backends = ["torch", "transformers"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch", "transformers"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch", "transformers"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch", "transformers"])
+
+
+class LEditsPPPipelineStableDiffusionXL(metaclass=DummyObject):
+    _backends = ["torch", "transformers"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch", "transformers"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch", "transformers"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch", "transformers"])
+
+
 class MusicLDMPipeline(metaclass=DummyObject):
     _backends = ["torch", "transformers"]
 

tests/pipelines/ledits_pp/test_ledits_pp_stable_diffusion.py

@@ -0,0 +1,244 @@
+# coding=utf-8
+# Copyright 2023 HuggingFace Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import gc
+import random
+import unittest
+
+import numpy as np
+import torch
+from PIL import Image
+from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
+
+from diffusers import (
+    AutoencoderKL,
+    DPMSolverMultistepScheduler,
+    LEditsPPPipelineStableDiffusion,
+    UNet2DConditionModel,
+)
+from diffusers.utils.testing_utils import (
+    enable_full_determinism,
+    floats_tensor,
+    load_image,
+    require_torch_gpu,
+    skip_mps,
+    slow,
+    torch_device,
+)
+
+
+enable_full_determinism()
+
+
+@skip_mps
+class LEditsPPPipelineStableDiffusionFastTests(unittest.TestCase):
+    pipeline_class = LEditsPPPipelineStableDiffusion
+
+    def get_dummy_components(self):
+        torch.manual_seed(0)
+        unet = UNet2DConditionModel(
+            block_out_channels=(32, 64, 64),
+            layers_per_block=2,
+            sample_size=32,
+            in_channels=4,
+            out_channels=4,
+            down_block_types=("DownBlock2D", "CrossAttnDownBlock2D", "CrossAttnDownBlock2D"),
+            up_block_types=("CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "UpBlock2D"),
+            cross_attention_dim=32,
+        )
+        scheduler = DPMSolverMultistepScheduler(algorithm_type="sde-dpmsolver++", solver_order=2)
+        torch.manual_seed(0)
+        vae = AutoencoderKL(
+            block_out_channels=[32, 64],
+            in_channels=3,
+            out_channels=3,
+            down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"],
+            up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"],
+            latent_channels=4,
+        )
+        torch.manual_seed(0)
+        text_encoder_config = CLIPTextConfig(
+            bos_token_id=0,
+            eos_token_id=2,
+            hidden_size=32,
+            intermediate_size=37,
+            layer_norm_eps=1e-05,
+            num_attention_heads=4,
+            num_hidden_layers=5,
+            pad_token_id=1,
+            vocab_size=1000,
+        )
+        text_encoder = CLIPTextModel(text_encoder_config)
+        tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
+
+        components = {
+            "unet": unet,
+            "scheduler": scheduler,
+            "vae": vae,
+            "text_encoder": text_encoder,
+            "tokenizer": tokenizer,
+            "safety_checker": None,
+            "feature_extractor": None,
+        }
+        return components
+
+    def get_dummy_inputs(self, device, seed=0):
+        if str(device).startswith("mps"):
+            generator = torch.manual_seed(seed)
+        else:
+            generator = torch.Generator(device=device).manual_seed(seed)
+        inputs = {
+            "generator": generator,
+            "editing_prompt": ["wearing glasses", "sunshine"],
+            "reverse_editing_direction": [False, True],
+            "edit_guidance_scale": [10.0, 5.0],
+        }
+        return inputs
+
+    def get_dummy_inversion_inputs(self, device, seed=0):
+        images = floats_tensor((2, 3, 32, 32), rng=random.Random(0)).cpu().permute(0, 2, 3, 1)
+        images = 255 * images
+        image_1 = Image.fromarray(np.uint8(images[0])).convert("RGB")
+        image_2 = Image.fromarray(np.uint8(images[1])).convert("RGB")
+
+        if str(device).startswith("mps"):
+            generator = torch.manual_seed(seed)
+        else:
+            generator = torch.Generator(device=device).manual_seed(seed)
+
+        inputs = {
+            "image": [image_1, image_2],
+            "source_prompt": "",
+            "source_guidance_scale": 3.5,
+            "num_inversion_steps": 20,
+            "skip": 0.15,
+            "generator": generator,
+        }
+        return inputs
+
+    def test_ledits_pp_inversion(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        sd_pipe = LEditsPPPipelineStableDiffusion(**components)
+        sd_pipe = sd_pipe.to(torch_device)
+        sd_pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_dummy_inversion_inputs(device)
+        inputs["image"] = inputs["image"][0]
+        sd_pipe.invert(**inputs)
+        assert sd_pipe.init_latents.shape == (
+            1,
+            4,
+            int(32 / sd_pipe.vae_scale_factor),
+            int(32 / sd_pipe.vae_scale_factor),
+        )
+
+        latent_slice = sd_pipe.init_latents[0, -1, -3:, -3:].to(device)
+        print(latent_slice.flatten())
+        expected_slice = np.array([-0.9084, -0.0367, 0.2940, 0.0839, 0.6890, 0.2651, -0.7104, 2.1090, -0.7822])
+        assert np.abs(latent_slice.flatten() - expected_slice).max() < 1e-3
+
+    def test_ledits_pp_inversion_batch(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        sd_pipe = LEditsPPPipelineStableDiffusion(**components)
+        sd_pipe = sd_pipe.to(torch_device)
+        sd_pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_dummy_inversion_inputs(device)
+        sd_pipe.invert(**inputs)
+        assert sd_pipe.init_latents.shape == (
+            2,
+            4,
+            int(32 / sd_pipe.vae_scale_factor),
+            int(32 / sd_pipe.vae_scale_factor),
+        )
+
+        latent_slice = sd_pipe.init_latents[0, -1, -3:, -3:].to(device)
+        print(latent_slice.flatten())
+        expected_slice = np.array([0.2528, 0.1458, -0.2166, 0.4565, -0.5657, -1.0286, -0.9961, 0.5933, 1.1173])
+        assert np.abs(latent_slice.flatten() - expected_slice).max() < 1e-3
+
+        latent_slice = sd_pipe.init_latents[1, -1, -3:, -3:].to(device)
+        print(latent_slice.flatten())
+        expected_slice = np.array([-0.0796, 2.0583, 0.5501, 0.5358, 0.0282, -0.2803, -1.0470, 0.7023, -0.0072])
+        assert np.abs(latent_slice.flatten() - expected_slice).max() < 1e-3
+
+    def test_ledits_pp_warmup_steps(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        pipe = LEditsPPPipelineStableDiffusion(**components)
+        pipe = pipe.to(torch_device)
+        pipe.set_progress_bar_config(disable=None)
+
+        inversion_inputs = self.get_dummy_inversion_inputs(device)
+        pipe.invert(**inversion_inputs)
+
+        inputs = self.get_dummy_inputs(device)
+
+        inputs["edit_warmup_steps"] = [0, 5]
+        pipe(**inputs).images
+
+        inputs["edit_warmup_steps"] = [5, 0]
+        pipe(**inputs).images
+
+        inputs["edit_warmup_steps"] = [5, 10]
+        pipe(**inputs).images
+
+        inputs["edit_warmup_steps"] = [10, 5]
+        pipe(**inputs).images
+
+
+@slow
+@require_torch_gpu
+class LEditsPPPipelineStableDiffusionSlowTests(unittest.TestCase):
+    def tearDown(self):
+        super().tearDown()
+        gc.collect()
+        torch.cuda.empty_cache()
+
+    @classmethod
+    def setUpClass(cls):
+        raw_image = load_image(
+            "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/pix2pix/cat_6.png"
+        )
+        raw_image = raw_image.convert("RGB").resize((512, 512))
+        cls.raw_image = raw_image
+
+    def test_ledits_pp_editing(self):
+        pipe = LEditsPPPipelineStableDiffusion.from_pretrained(
+            "runwayml/stable-diffusion-v1-5", safety_checker=None, torch_dtype=torch.float16
+        )
+        pipe = pipe.to(torch_device)
+        pipe.set_progress_bar_config(disable=None)
+
+        generator = torch.manual_seed(0)
+        _ = pipe.invert(image=self.raw_image, generator=generator)
+        generator = torch.manual_seed(0)
+        inputs = {
+            "generator": generator,
+            "editing_prompt": ["cat", "dog"],
+            "reverse_editing_direction": [True, False],
+            "edit_guidance_scale": [5.0, 5.0],
+            "edit_threshold": [0.8, 0.8],
+        }
+        reconstruction = pipe(**inputs, output_type="np").images[0]
+
+        output_slice = reconstruction[150:153, 140:143, -1]
+        output_slice = output_slice.flatten()
+        expected_slice = np.array(
+            [0.9453125, 0.93310547, 0.84521484, 0.94628906, 0.9111328, 0.80859375, 0.93847656, 0.9042969, 0.8144531]
+        )
+        assert np.abs(output_slice - expected_slice).max() < 1e-2

tests/pipelines/ledits_pp/test_ledits_pp_stable_diffusion_xl.py

@@ -0,0 +1,289 @@
+# coding=utf-8
+# Copyright 2023 HuggingFace Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import random
+import unittest
+
+import numpy as np
+import torch
+from PIL import Image
+from transformers import (
+    CLIPImageProcessor,
+    CLIPTextConfig,
+    CLIPTextModel,
+    CLIPTextModelWithProjection,
+    CLIPTokenizer,
+    CLIPVisionConfig,
+    CLIPVisionModelWithProjection,
+)
+
+from diffusers import (
+    AutoencoderKL,
+    DPMSolverMultistepScheduler,
+    LEditsPPPipelineStableDiffusionXL,
+    UNet2DConditionModel,
+)
+
+# from diffusers.image_processor import VaeImageProcessor
+from diffusers.utils.testing_utils import (
+    enable_full_determinism,
+    floats_tensor,
+    load_image,
+    require_torch_gpu,
+    skip_mps,
+    slow,
+    torch_device,
+)
+
+
+enable_full_determinism()
+
+
+@skip_mps
+class LEditsPPPipelineStableDiffusionXLFastTests(unittest.TestCase):
+    pipeline_class = LEditsPPPipelineStableDiffusionXL
+
+    def get_dummy_components(self, skip_first_text_encoder=False, time_cond_proj_dim=None):
+        torch.manual_seed(0)
+        unet = UNet2DConditionModel(
+            block_out_channels=(32, 64),
+            layers_per_block=2,
+            sample_size=32,
+            in_channels=4,
+            out_channels=4,
+            time_cond_proj_dim=time_cond_proj_dim,
+            down_block_types=("DownBlock2D", "CrossAttnDownBlock2D"),
+            up_block_types=("CrossAttnUpBlock2D", "UpBlock2D"),
+            # SD2-specific config below
+            attention_head_dim=(2, 4),
+            use_linear_projection=True,
+            addition_embed_type="text_time",
+            addition_time_embed_dim=8,
+            transformer_layers_per_block=(1, 2),
+            projection_class_embeddings_input_dim=80,  # 6 * 8 + 32
+            cross_attention_dim=64 if not skip_first_text_encoder else 32,
+        )
+        scheduler = DPMSolverMultistepScheduler(algorithm_type="sde-dpmsolver++", solver_order=2)
+        torch.manual_seed(0)
+        vae = AutoencoderKL(
+            block_out_channels=[32, 64],
+            in_channels=3,
+            out_channels=3,
+            down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"],
+            up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"],
+            latent_channels=4,
+            sample_size=128,
+        )
+        torch.manual_seed(0)
+        image_encoder_config = CLIPVisionConfig(
+            hidden_size=32,
+            image_size=224,
+            projection_dim=32,
+            intermediate_size=37,
+            num_attention_heads=4,
+            num_channels=3,
+            num_hidden_layers=5,
+            patch_size=14,
+        )
+        image_encoder = CLIPVisionModelWithProjection(image_encoder_config)
+
+        feature_extractor = CLIPImageProcessor(
+            crop_size=224,
+            do_center_crop=True,
+            do_normalize=True,
+            do_resize=True,
+            image_mean=[0.48145466, 0.4578275, 0.40821073],
+            image_std=[0.26862954, 0.26130258, 0.27577711],
+            resample=3,
+            size=224,
+        )
+
+        torch.manual_seed(0)
+        text_encoder_config = CLIPTextConfig(
+            bos_token_id=0,
+            eos_token_id=2,
+            hidden_size=32,
+            intermediate_size=37,
+            layer_norm_eps=1e-05,
+            num_attention_heads=4,
+            num_hidden_layers=5,
+            pad_token_id=1,
+            vocab_size=1000,
+            # SD2-specific config below
+            hidden_act="gelu",
+            projection_dim=32,
+        )
+        text_encoder = CLIPTextModel(text_encoder_config)
+        tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
+
+        text_encoder_2 = CLIPTextModelWithProjection(text_encoder_config)
+        tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
+
+        components = {
+            "unet": unet,
+            "scheduler": scheduler,
+            "vae": vae,
+            "text_encoder": text_encoder if not skip_first_text_encoder else None,
+            "tokenizer": tokenizer if not skip_first_text_encoder else None,
+            "text_encoder_2": text_encoder_2,
+            "tokenizer_2": tokenizer_2,
+            "image_encoder": image_encoder,
+            "feature_extractor": feature_extractor,
+        }
+        return components
+
+    def get_dummy_inputs(self, device, seed=0):
+        if str(device).startswith("mps"):
+            generator = torch.manual_seed(seed)
+        else:
+            generator = torch.Generator(device=device).manual_seed(seed)
+        inputs = {
+            "generator": generator,
+            "editing_prompt": ["wearing glasses", "sunshine"],
+            "reverse_editing_direction": [False, True],
+            "edit_guidance_scale": [10.0, 5.0],
+        }
+        return inputs
+
+    def get_dummy_inversion_inputs(self, device, seed=0):
+        images = floats_tensor((2, 3, 32, 32), rng=random.Random(0)).cpu().permute(0, 2, 3, 1)
+        images = 255 * images
+        image_1 = Image.fromarray(np.uint8(images[0])).convert("RGB")
+        image_2 = Image.fromarray(np.uint8(images[1])).convert("RGB")
+
+        if str(device).startswith("mps"):
+            generator = torch.manual_seed(seed)
+        else:
+            generator = torch.Generator(device=device).manual_seed(seed)
+
+        inputs = {
+            "image": [image_1, image_2],
+            "source_prompt": "",
+            "source_guidance_scale": 3.5,
+            "num_inversion_steps": 20,
+            "skip": 0.15,
+            "generator": generator,
+        }
+        return inputs
+
+    def test_ledits_pp_inversion(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        sd_pipe = LEditsPPPipelineStableDiffusionXL(**components)
+        sd_pipe = sd_pipe.to(torch_device)
+        sd_pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_dummy_inversion_inputs(device)
+        inputs["image"] = inputs["image"][0]
+        sd_pipe.invert(**inputs)
+        assert sd_pipe.init_latents.shape == (
+            1,
+            4,
+            int(32 / sd_pipe.vae_scale_factor),
+            int(32 / sd_pipe.vae_scale_factor),
+        )
+
+        latent_slice = sd_pipe.init_latents[0, -1, -3:, -3:].to(device)
+        expected_slice = np.array([-0.9084, -0.0367, 0.2940, 0.0839, 0.6890, 0.2651, -0.7103, 2.1090, -0.7821])
+        assert np.abs(latent_slice.flatten() - expected_slice).max() < 1e-3
+
+    def test_ledits_pp_inversion_batch(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        sd_pipe = LEditsPPPipelineStableDiffusionXL(**components)
+        sd_pipe = sd_pipe.to(torch_device)
+        sd_pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_dummy_inversion_inputs(device)
+        sd_pipe.invert(**inputs)
+        assert sd_pipe.init_latents.shape == (
+            2,
+            4,
+            int(32 / sd_pipe.vae_scale_factor),
+            int(32 / sd_pipe.vae_scale_factor),
+        )
+
+        latent_slice = sd_pipe.init_latents[0, -1, -3:, -3:].to(device)
+        print(latent_slice.flatten())
+        expected_slice = np.array([0.2528, 0.1458, -0.2166, 0.4565, -0.5656, -1.0286, -0.9961, 0.5933, 1.1172])
+        assert np.abs(latent_slice.flatten() - expected_slice).max() < 1e-3
+
+        latent_slice = sd_pipe.init_latents[1, -1, -3:, -3:].to(device)
+        print(latent_slice.flatten())
+        expected_slice = np.array([-0.0796, 2.0583, 0.5500, 0.5358, 0.0282, -0.2803, -1.0470, 0.7024, -0.0072])
+        print(latent_slice.flatten())
+        assert np.abs(latent_slice.flatten() - expected_slice).max() < 1e-3
+
+    def test_ledits_pp_warmup_steps(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        pipe = LEditsPPPipelineStableDiffusionXL(**components)
+        pipe = pipe.to(torch_device)
+        pipe.set_progress_bar_config(disable=None)
+
+        inversion_inputs = self.get_dummy_inversion_inputs(device)
+        inversion_inputs["image"] = inversion_inputs["image"][0]
+        pipe.invert(**inversion_inputs)
+
+        inputs = self.get_dummy_inputs(device)
+
+        inputs["edit_warmup_steps"] = [0, 5]
+        pipe(**inputs).images
+
+        inputs["edit_warmup_steps"] = [5, 0]
+        pipe(**inputs).images
+
+        inputs["edit_warmup_steps"] = [5, 10]
+        pipe(**inputs).images
+
+        inputs["edit_warmup_steps"] = [10, 5]
+        pipe(**inputs).images
+
+
+@slow
+@require_torch_gpu
+class LEditsPPPipelineStableDiffusionXLSlowTests(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        raw_image = load_image(
+            "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/pix2pix/cat_6.png"
+        )
+        raw_image = raw_image.convert("RGB").resize((512, 512))
+        cls.raw_image = raw_image
+
+    def test_ledits_pp_edit(self):
+        pipe = LEditsPPPipelineStableDiffusionXL.from_pretrained(
+            "stabilityai/stable-diffusion-xl-base-1.0", safety_checker=None, add_watermarker=None
+        )
+        pipe = pipe.to(torch_device)
+        pipe.set_progress_bar_config(disable=None)
+
+        generator = torch.manual_seed(0)
+        _ = pipe.invert(image=self.raw_image, generator=generator, num_zero_noise_steps=0)
+        inputs = {
+            "generator": generator,
+            "editing_prompt": ["cat", "dog"],
+            "reverse_editing_direction": [True, False],
+            "edit_guidance_scale": [2.0, 4.0],
+            "edit_threshold": [0.8, 0.8],
+        }
+        reconstruction = pipe(**inputs, output_type="np").images[0]
+
+        output_slice = reconstruction[150:153, 140:143, -1]
+        output_slice = output_slice.flatten()
+        expected_slice = np.array(
+            [0.56419, 0.44121838, 0.2765603, 0.5708484, 0.42763475, 0.30945742, 0.5387106, 0.4735807, 0.3547244]
+        )
+        assert np.abs(output_slice - expected_slice).max() < 1e-3

tests/pipelines/stable_diffusion_2/test_stable_diffusion_upscale.py

@@ -522,7 +522,7 @@ class StableDiffusionUpscalePipelineIntegrationTests(unittest.TestCase):
         ckpt_path = (
             "https://huggingface.co/stabilityai/stable-diffusion-x4-upscaler/blob/main/x4-upscaler-ema.safetensors"
         )
-        single_file_pipe = StableDiffusionUpscalePipeline.from_single_file(ckpt_path, load_safety_checker=True)
+        single_file_pipe = StableDiffusionUpscalePipeline.from_single_file(ckpt_path)
 
         for param_name, param_value in single_file_pipe.text_encoder.config.to_dict().items():
             if param_name in ["torch_dtype", "architectures", "_name_or_path"]:
@@ -540,13 +540,12 @@ class StableDiffusionUpscalePipelineIntegrationTests(unittest.TestCase):
         for param_name, param_value in single_file_pipe.vae.config.items():
             if param_name in PARAMS_TO_IGNORE:
                 continue
+
+            # The sample_size parameter for the VAE is incorrectly configured on the hub
+            # It must be 512, but it is 256 on the hub
+            if param_name == "sample_size":
+                pipe.vae.config[param_name] = param_value
+
             assert (
                 pipe.vae.config[param_name] == param_value
             ), f"{param_name} differs between single file loading and pretrained loading"
-
-        for param_name, param_value in single_file_pipe.safety_checker.config.to_dict().items():
-            if param_name in PARAMS_TO_IGNORE:
-                continue
-            assert (
-                pipe.safety_checker.config.to_dict()[param_name] == param_value
-            ), f"{param_name} differs between single file loading and pretrained loading"

utils/fetch_latest_release_branch.py

@@ -0,0 +1,68 @@
+# coding=utf-8
+# Copyright 2024 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import requests
+from packaging.version import parse
+
+
+# GitHub repository details
+USER = "huggingface"
+REPO = "diffusers"
+
+
+def fetch_all_branches(user, repo):
+    branches = []  # List to store all branches
+    page = 1  # Start from first page
+    while True:
+        # Make a request to the GitHub API for the branches
+        response = requests.get(f"https://api.github.com/repos/{user}/{repo}/branches", params={"page": page})
+
+        # Check if the request was successful
+        if response.status_code == 200:
+            # Add the branches from the current page to the list
+            branches.extend([branch["name"] for branch in response.json()])
+
+            # Check if there is a 'next' link for pagination
+            if "next" in response.links:
+                page += 1  # Move to the next page
+            else:
+                break  # Exit loop if there is no next page
+        else:
+            print("Failed to retrieve branches:", response.status_code)
+            break
+
+    return branches
+
+
+def main():
+    # Fetch all branches
+    branches = fetch_all_branches(USER, REPO)
+
+    # Filter branches.
+    # print(f"Total branches: {len(branches)}")
+    filtered_branches = []
+    for branch in branches:
+        if branch.startswith("v") and ("-release" in branch or "-patch" in branch):
+            filtered_branches.append(branch)
+            # print(f"Filtered: {branch}")
+
+    sorted_branches = sorted(filtered_branches, key=lambda x: parse(x.split("-")[0][1:]), reverse=True)
+    latest_branch = sorted_branches[0]
+    # print(f"Latest branch: {latest_branch}")
+    return latest_branch
+
+
+if __name__ == "__main__":
+    print(main())

utils/notify_slack_about_release.py

@@ -0,0 +1,80 @@
+# coding=utf-8
+# Copyright 2024 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+
+import requests
+
+
+# Configuration
+LIBRARY_NAME = "diffusers"
+GITHUB_REPO = "huggingface/diffusers"
+SLACK_WEBHOOK_URL = os.getenv("SLACK_WEBHOOK_URL")
+
+
+def check_pypi_for_latest_release(library_name):
+    """Check PyPI for the latest release of the library."""
+    response = requests.get(f"https://pypi.org/pypi/{library_name}/json")
+    if response.status_code == 200:
+        data = response.json()
+        return data["info"]["version"]
+    else:
+        print("Failed to fetch library details from PyPI.")
+        return None
+
+
+def get_github_release_info(github_repo):
+    """Fetch the latest release info from GitHub."""
+    url = f"https://api.github.com/repos/{github_repo}/releases/latest"
+    response = requests.get(url)
+
+    if response.status_code == 200:
+        data = response.json()
+        return {"tag_name": data["tag_name"], "url": data["html_url"], "release_time": data["published_at"]}
+
+    else:
+        print("Failed to fetch release info from GitHub.")
+        return None
+
+
+def notify_slack(webhook_url, library_name, version, release_info):
+    """Send a notification to a Slack channel."""
+    message = (
+        f"🚀 New release for {library_name} available: version **{version}** 🎉\n"
+        f"📜 Release Notes: {release_info['url']}\n"
+        f"⏱️ Release time: {release_info['release_time']}"
+    )
+    payload = {"text": message}
+    response = requests.post(webhook_url, json=payload)
+
+    if response.status_code == 200:
+        print("Notification sent to Slack successfully.")
+    else:
+        print("Failed to send notification to Slack.")
+
+
+def main():
+    latest_version = check_pypi_for_latest_release(LIBRARY_NAME)
+    release_info = get_github_release_info(GITHUB_REPO)
+    parsed_version = release_info["tag_name"].replace("v", "")
+
+    if latest_version and release_info and latest_version == parsed_version:
+        notify_slack(SLACK_WEBHOOK_URL, LIBRARY_NAME, latest_version, release_info)
+    else:
+        raise ValueError("There were some problems.")
+
+
+if __name__ == "__main__":
+    main()