fix DDPMScheduler.set_timesteps() when num_inference_steps not a divisor of num_train_timesteps

.github/workflows/nightly_tests.yml

@@ -62,6 +62,7 @@ jobs:
         run: |
           python -m pip install -e .[quality,test]
           python -m pip install -U git+https://github.com/huggingface/transformers
+          python -m pip install git+https://github.com/huggingface/accelerate
 
       - name: Environment
         run: |
@@ -134,6 +135,7 @@ jobs:
           ${CONDA_RUN} python -m pip install --upgrade pip
           ${CONDA_RUN} python -m pip install -e .[quality,test]
           ${CONDA_RUN} python -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu
+          ${CONDA_RUN} python -m pip install git+https://github.com/huggingface/accelerate
 
       - name: Environment
         shell: arch -arch arm64 bash {0}
@@ -157,4 +159,4 @@ jobs:
         uses: actions/upload-artifact@v2
         with:
           name: torch_mps_test_reports
-          path: reports
+          path: reports

.github/workflows/pr_tests.yml

@@ -60,6 +60,7 @@ jobs:
         apt-get update && apt-get install libsndfile1-dev -y
         python -m pip install -e .[quality,test]
         python -m pip install -U git+https://github.com/huggingface/transformers
+        python -m pip install git+https://github.com/huggingface/accelerate
 
     - name: Environment
       run: |
@@ -126,6 +127,7 @@ jobs:
         ${CONDA_RUN} python -m pip install --upgrade pip
         ${CONDA_RUN} python -m pip install -e .[quality,test]
         ${CONDA_RUN} python -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu
+        ${CONDA_RUN} python -m pip install git+https://github.com/huggingface/accelerate
         ${CONDA_RUN} python -m pip install -U git+https://github.com/huggingface/transformers
 
     - name: Environment

.github/workflows/push_tests.yml

@@ -62,6 +62,7 @@ jobs:
       run: |
         python -m pip install -e .[quality,test]
         python -m pip install -U git+https://github.com/huggingface/transformers
+        python -m pip install git+https://github.com/huggingface/accelerate
 
     - name: Environment
       run: |
@@ -130,6 +131,7 @@ jobs:
     - name: Install dependencies
       run: |
         python -m pip install -e .[quality,test,training]
+        python -m pip install git+https://github.com/huggingface/accelerate
         python -m pip install -U git+https://github.com/huggingface/transformers
 
     - name: Environment
@@ -151,4 +153,4 @@ jobs:
       uses: actions/upload-artifact@v2
       with:
         name: examples_test_reports
-        path: reports
+        path: reports

docs/README.md

@@ -155,9 +155,9 @@ adds a link to its documentation with this syntax: \[\`XXXClass\`\] or \[\`funct
 function to be in the main package.
 
 If you want to create a link to some internal class or function, you need to
-provide its path. For instance: \[\`pipeline_utils.ImagePipelineOutput\`\]. This will be converted into a link with
-`pipeline_utils.ImagePipelineOutput` in the description. To get rid of the path and only keep the name of the object you are
-linking to in the description, add a ~: \[\`~pipeline_utils.ImagePipelineOutput\`\] will generate a link with `ImagePipelineOutput` in the description.
+provide its path. For instance: \[\`pipelines.ImagePipelineOutput\`\]. This will be converted into a link with
+`pipelines.ImagePipelineOutput` in the description. To get rid of the path and only keep the name of the object you are
+linking to in the description, add a ~: \[\`~pipelines.ImagePipelineOutput\`\] will generate a link with `ImagePipelineOutput` in the description.
 
 The same works for methods so you can either use \[\`XXXClass.method\`\] or \[~\`XXXClass.method\`\].
 

docs/source/api/diffusion_pipeline.mdx

@@ -30,13 +30,17 @@ Any pipeline object can be saved locally with [`~DiffusionPipeline.save_pretrain
 
 ## DiffusionPipeline
 [[autodoc]] DiffusionPipeline
-	- from_pretrained
-	- save_pretrained
-	- to
+	- all
+	- __call__
 	- device
-	- components
+	- to
 
 ## ImagePipelineOutput
 By default diffusion pipelines return an object of class
 
-[[autodoc]] pipeline_utils.ImagePipelineOutput
+[[autodoc]] pipelines.ImagePipelineOutput
+
+## AudioPipelineOutput
+By default diffusion pipelines return an object of class
+
+[[autodoc]] pipelines.AudioPipelineOutput

docs/source/api/models.mdx

@@ -41,13 +41,13 @@ The models are built on the base class ['ModelMixin'] that is a `torch.nn.module
 [[autodoc]] models.vae.DecoderOutput
 
 ## VQEncoderOutput
-[[autodoc]] models.vae.VQEncoderOutput
+[[autodoc]] models.vq_model.VQEncoderOutput
 
 ## VQModel
 [[autodoc]] VQModel
 
 ## AutoencoderKLOutput
-[[autodoc]] models.vae.AutoencoderKLOutput
+[[autodoc]] models.autoencoder_kl.AutoencoderKLOutput
 
 ## AutoencoderKL
 [[autodoc]] AutoencoderKL

docs/source/api/outputs.mdx

@@ -25,7 +25,7 @@ pipeline = DDIMPipeline.from_pretrained("google/ddpm-cifar10-32")
 outputs = pipeline()
 ```
 
-The `outputs` object is a [`~pipeline_utils.ImagePipelineOutput`], as we can see in the
+The `outputs` object is a [`~pipelines.ImagePipelineOutput`], as we can see in the
 documentation of that class below, it means it has an image attribute.
 
 You can access each attribute as you would usually do, and if that attribute has not been returned by the model, you will get `None`:

docs/source/api/pipelines/alt_diffusion.mdx

@@ -69,15 +69,15 @@ If you want to use all possible use cases in a single `DiffusionPipeline` we rec
 
 ## AltDiffusionPipelineOutput
 [[autodoc]] pipelines.alt_diffusion.AltDiffusionPipelineOutput
+	- all
+	- __call__
 
 ## AltDiffusionPipeline
 [[autodoc]] AltDiffusionPipeline
+	- all
 	- __call__
-	- enable_attention_slicing
-	- disable_attention_slicing
 
 ## AltDiffusionImg2ImgPipeline
 [[autodoc]] AltDiffusionImg2ImgPipeline
+	- all
 	- __call__
-	- enable_attention_slicing
-	- disable_attention_slicing

docs/source/api/pipelines/audio_diffusion.mdx

@@ -91,12 +91,8 @@ display(Audio(output.audios[0], rate=pipe.mel.get_sample_rate()))
 
 ## AudioDiffusionPipeline
 [[autodoc]] AudioDiffusionPipeline
-    - __call__
-    - encode
-    - slerp
-
+	- all
+	- __call__
 
 ## Mel
 [[autodoc]] Mel
-    - audio_slice_to_image
-    - image_to_audio

docs/source/api/pipelines/cycle_diffusion.mdx

@@ -96,4 +96,5 @@ image.save("black_to_blue.png")
 
 ## CycleDiffusionPipeline
 [[autodoc]] CycleDiffusionPipeline
+	- all
 	- __call__

docs/source/api/pipelines/dance_diffusion.mdx

@@ -30,4 +30,5 @@ The original codebase of this implementation can be found [here](https://github.
 
 ## DanceDiffusionPipeline
 [[autodoc]] DanceDiffusionPipeline
-    - __call__
+	- all
+	- __call__

docs/source/api/pipelines/ddim.mdx

@@ -32,4 +32,5 @@ For questions, feel free to contact the author on [tsong.me](https://tsong.me/).
 
 ## DDIMPipeline
 [[autodoc]] DDIMPipeline
-    - __call__
+	- all
+	- __call__

docs/source/api/pipelines/ddpm.mdx

@@ -33,4 +33,5 @@ The original codebase of this paper can be found [here](https://github.com/hojon
 
 # DDPMPipeline
 [[autodoc]] DDPMPipeline
-    - __call__
+	- all
+	- __call__

docs/source/api/pipelines/latent_diffusion.mdx

@@ -40,8 +40,10 @@ The original codebase can be found [here](https://github.com/CompVis/latent-diff
 
 ## LDMTextToImagePipeline
 [[autodoc]] LDMTextToImagePipeline
-    - __call__
+	- all
+	- __call__
 
 ## LDMSuperResolutionPipeline
 [[autodoc]] LDMSuperResolutionPipeline
-    - __call__
+	- all
+	- __call__

docs/source/api/pipelines/latent_diffusion_uncond.mdx

@@ -38,4 +38,5 @@ The original codebase can be found [here](https://github.com/CompVis/latent-diff
 
 ## LDMPipeline
 [[autodoc]] LDMPipeline
-    - __call__
+	- all
+	- __call__

docs/source/api/pipelines/paint_by_example.mdx

@@ -69,5 +69,6 @@ image
 ```
 
 ## PaintByExamplePipeline
-[[autodoc]] pipelines.paint_by_example.pipeline_paint_by_example.PaintByExamplePipeline
-    - __call__
+[[autodoc]] PaintByExamplePipeline
+	- all
+	- __call__

docs/source/api/pipelines/pndm.mdx

@@ -30,6 +30,6 @@ The original codebase can be found [here](https://github.com/luping-liu/PNDM).
 
 
 ## PNDMPipeline
-[[autodoc]] pipelines.pndm.pipeline_pndm.PNDMPipeline
-    - __call__
-
+[[autodoc]] PNDMPipeline
+	- all
+	- __call__

docs/source/api/pipelines/repaint.mdx

@@ -72,6 +72,6 @@ inpainted_image = output.images[0]
 ```
 
 ## RePaintPipeline
-[[autodoc]] pipelines.repaint.pipeline_repaint.RePaintPipeline
-    - __call__
-
+[[autodoc]] RePaintPipeline
+	- all
+	- __call__

docs/source/api/pipelines/score_sde_ve.mdx

@@ -32,5 +32,5 @@ This pipeline implements the Variance Expanding (VE) variant of the method.
 
 ## ScoreSdeVePipeline
 [[autodoc]] ScoreSdeVePipeline
-    - __call__
-
+	- all
+	- __call__

docs/source/api/pipelines/stable_diffusion.mdx

@@ -73,16 +73,18 @@ If you want to use all possible use cases in a single `DiffusionPipeline` you ca
 
 ## StableDiffusionPipeline
 [[autodoc]] StableDiffusionPipeline
+	- all
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 
+
+
 ## StableDiffusionImg2ImgPipeline
 [[autodoc]] StableDiffusionImg2ImgPipeline
+	- all
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
@@ -91,6 +93,7 @@ If you want to use all possible use cases in a single `DiffusionPipeline` you ca
 
 ## StableDiffusionInpaintPipeline
 [[autodoc]] StableDiffusionInpaintPipeline
+	- all
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
@@ -99,6 +102,7 @@ If you want to use all possible use cases in a single `DiffusionPipeline` you ca
 
 ## StableDiffusionDepth2ImgPipeline
 [[autodoc]] StableDiffusionDepth2ImgPipeline
+	- all
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
@@ -107,15 +111,16 @@ If you want to use all possible use cases in a single `DiffusionPipeline` you ca
 
 ## StableDiffusionImageVariationPipeline
 [[autodoc]] StableDiffusionImageVariationPipeline
+	- all
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 
-
 ## StableDiffusionUpscalePipeline
 [[autodoc]] StableDiffusionUpscalePipeline
+	- all
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing

docs/source/api/pipelines/stable_diffusion_safe.mdx

@@ -81,10 +81,10 @@ To use a different scheduler, you can either change it via the [`ConfigMixin.fro
 
 ## StableDiffusionSafePipelineOutput
 [[autodoc]] pipelines.stable_diffusion_safe.StableDiffusionSafePipelineOutput
+	- all
+	- __call__
 
 ## StableDiffusionPipelineSafe
 [[autodoc]] StableDiffusionPipelineSafe
+	- all
 	- __call__
-	- enable_attention_slicing
-	- disable_attention_slicing
-

docs/source/api/pipelines/stochastic_karras_ve.mdx

@@ -32,4 +32,5 @@ This pipeline implements the Stochastic sampling tailored to the Variance-Expand
 
 ## KarrasVePipeline
 [[autodoc]] KarrasVePipeline
-    - __call__
+	- all
+	- __call__

docs/source/api/pipelines/unclip.mdx

@@ -24,8 +24,14 @@ The unCLIP model in diffusers comes from kakaobrain's karlo and the original cod
 | Pipeline | Tasks | Colab
 |---|---|:---:|
 | [pipeline_unclip.py](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/unclip/pipeline_unclip.py) | *Text-to-Image Generation* | - |
+| [pipeline_unclip_image_variation.py](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/unclip/pipeline_unclip_image_variation.py) | *Image-Guided Image Generation* | - |
 
 
 ## UnCLIPPipeline
-[[autodoc]] pipelines.unclip.pipeline_unclip.UnCLIPPipeline
-    - __call__
+[[autodoc]] UnCLIPPipeline
+	- all
+	- __call__
+
+[[autodoc]] UnCLIPImageVariationPipeline
+	- all
+	- __call__

docs/source/api/pipelines/versatile_diffusion.mdx

@@ -56,18 +56,15 @@ To use a different scheduler, you can either change it via the [`ConfigMixin.fro
 
 ## VersatileDiffusionTextToImagePipeline
 [[autodoc]] VersatileDiffusionTextToImagePipeline
+	- all
 	- __call__
-	- enable_attention_slicing
-	- disable_attention_slicing
 
 ## VersatileDiffusionImageVariationPipeline
 [[autodoc]] VersatileDiffusionImageVariationPipeline
+	- all
 	- __call__
-	- enable_attention_slicing
-	- disable_attention_slicing
 
 ## VersatileDiffusionDualGuidedPipeline
 [[autodoc]] VersatileDiffusionDualGuidedPipeline
+	- all
 	- __call__
-	- enable_attention_slicing
-	- disable_attention_slicing

docs/source/api/pipelines/vq_diffusion.mdx

@@ -30,5 +30,6 @@ The original codebase can be found [here](https://github.com/microsoft/VQ-Diffus
 
 
 ## VQDiffusionPipeline
-[[autodoc]] pipelines.vq_diffusion.pipeline_vq_diffusion.VQDiffusionPipeline
-    - __call__
+[[autodoc]] VQDiffusionPipeline
+	- all
+	- __call__