Change checkpoint key used to identify CLIP models in single file checkpoints (#8319)

update

.github/workflows/benchmark.yml

@@ -39,7 +39,7 @@ jobs:
           python utils/print_env.py
       - name: Diffusers Benchmarking
         env:
-            HUGGING_FACE_HUB_TOKEN: ${{ secrets.DIFFUSERS_BOT_TOKEN }}
+            HF_TOKEN: ${{ secrets.DIFFUSERS_BOT_TOKEN }}
             BASE_PATH: benchmark_outputs
         run: |
           export TOTAL_GPU_MEMORY=$(python -c "import torch; print(torch.cuda.get_device_properties(0).total_memory / (1024**3))")

.github/workflows/build_docker_images.yml

@@ -25,17 +25,17 @@ jobs:
     steps:
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v1
-      
+
       - name: Check out code
         uses: actions/checkout@v3
-      
+
       - name: Find Changed Dockerfiles
         id: file_changes
         uses: jitterbit/get-changed-files@v1
         with:
           format: 'space-delimited'
           token: ${{ secrets.GITHUB_TOKEN }}
-      
+
       - name: Build Changed Docker Images
         run: |
           CHANGED_FILES="${{ steps.file_changes.outputs.all }}"
@@ -52,7 +52,7 @@ jobs:
   build-and-push-docker-images:
     runs-on: [ self-hosted, intel-cpu, 8-cpu, ci ]
     if: github.event_name != 'pull_request'
-    
+
     permissions:
       contents: read
       packages: write
@@ -69,6 +69,7 @@ jobs:
           - diffusers-flax-tpu
           - diffusers-onnxruntime-cpu
           - diffusers-onnxruntime-cuda
+          - diffusers-doc-builder
 
     steps:
       - name: Checkout repository
@@ -90,24 +91,11 @@ jobs:
 
       - name: Post to a Slack channel
         id: slack
-        uses: slackapi/slack-github-action@6c661ce58804a1a20f6dc5fbee7f0381b469e001
+        uses: huggingface/hf-workflows/.github/actions/post-slack@main
         with:
           # Slack channel id, channel name, or user id to post message.
           # See also: https://api.slack.com/methods/chat.postMessage#channels
-          channel-id: ${{ env.CI_SLACK_CHANNEL }}
-          # For posting a rich message using Block Kit
-          payload: |
-            {
-              "text": "${{ matrix.image-name }} Docker Image build result: ${{ job.status }}\n${{ github.event.head_commit.url }}",
-              "blocks": [
-                {
-                  "type": "section",
-                  "text": {
-                    "type": "mrkdwn",
-                    "text": "${{ matrix.image-name }} Docker Image build result: ${{ job.status }}\n${{ github.event.head_commit.url }}"
-                  }
-                }
-              ]
-            }
-        env:
-          SLACK_BOT_TOKEN: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}
+          slack_channel: ${{ env.CI_SLACK_CHANNEL }}
+          title: "🤗 Results of the ${{ matrix.image-name }} Docker Image build"
+          status: ${{ job.status }}
+          slack_token: ${{ secrets.SLACK_CIFEEDBACK_BOT_TOKEN }}

.github/workflows/build_documentation.yml

@@ -21,7 +21,7 @@ jobs:
       package: diffusers
       notebook_folder: diffusers_doc
       languages: en ko zh ja pt
-
+      custom_container: diffusers/diffusers-doc-builder
     secrets:
       token: ${{ secrets.HUGGINGFACE_PUSH }}
       hf_token: ${{ secrets.HF_DOC_BUILD_PUSH }}

.github/workflows/build_pr_documentation.yml

@@ -20,3 +20,4 @@ jobs:
       install_libgl1: true
       package: diffusers
       languages: en ko zh ja pt
+      custom_container: diffusers/diffusers-doc-builder

.github/workflows/nightly_tests.yml

@@ -81,7 +81,7 @@ jobs:
 
       - name: Nightly PyTorch CUDA checkpoint (pipelines) tests
         env:
-          HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
           # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
           CUBLAS_WORKSPACE_CONFIG: :16:8
         run: |
@@ -141,7 +141,7 @@ jobs:
     - name: Run nightly PyTorch CUDA tests for non-pipeline modules
       if: ${{ matrix.module != 'examples'}}
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
         # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
         CUBLAS_WORKSPACE_CONFIG: :16:8
       run: |
@@ -154,7 +154,7 @@ jobs:
     - name: Run nightly example tests with Torch
       if: ${{ matrix.module == 'examples' }}
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
         # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
         CUBLAS_WORKSPACE_CONFIG: :16:8
       run: |
@@ -211,7 +211,7 @@ jobs:
 
     - name: Run nightly LoRA tests with PEFT and Torch
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
         # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
         CUBLAS_WORKSPACE_CONFIG: :16:8
       run: |
@@ -269,7 +269,7 @@ jobs:
 
     - name: Run nightly Flax TPU tests
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m pytest -n 0 \
           -s -v -k "Flax" \
@@ -324,7 +324,7 @@ jobs:
 
     - name: Run nightly ONNXRuntime CUDA tests
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile \
           -s -v -k "Onnx" \
@@ -390,7 +390,7 @@ jobs:
         shell: arch -arch arm64 bash {0}
         env:
           HF_HOME: /System/Volumes/Data/mnt/cache
-          HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
         run: |
           ${CONDA_RUN} python -m pytest -n 1 -s -v --make-reports=tests_torch_mps \
             --report-log=tests_torch_mps.log \

.github/workflows/pr_tests.yml

@@ -156,7 +156,7 @@ jobs:
       if: ${{ matrix.config.framework == 'pytorch_examples' }}
       run: |
         python -m venv /opt/venv && export PATH="/opt/venv/bin:$PATH"
-        python -m uv pip install peft
+        python -m uv pip install peft timm
         python -m pytest -n 4 --max-worker-restart=0 --dist=loadfile \
           --make-reports=tests_${{ matrix.config.report }} \
           examples

.github/workflows/push_tests.yml

@@ -56,6 +56,7 @@ jobs:
     needs: setup_torch_cuda_pipeline_matrix
     strategy:
       fail-fast: false
+      max-parallel: 8
       matrix:
         module: ${{ fromJson(needs.setup_torch_cuda_pipeline_matrix.outputs.pipeline_test_matrix) }}
     runs-on: [single-gpu, nvidia-gpu, t4, ci]
@@ -86,7 +87,7 @@ jobs:
           python utils/print_env.py
       - name: Slow PyTorch CUDA checkpoint tests on Ubuntu
         env:
-          HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
           # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
           CUBLAS_WORKSPACE_CONFIG: :16:8
         run: |
@@ -124,7 +125,7 @@ jobs:
         shell: bash
     strategy:
       matrix:
-        module: [models, schedulers, lora, others]
+        module: [models, schedulers, lora, others, single_file]
     steps:
     - name: Checkout diffusers
       uses: actions/checkout@v3
@@ -143,7 +144,7 @@ jobs:
 
     - name: Run slow PyTorch CUDA tests
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
         # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
         CUBLAS_WORKSPACE_CONFIG: :16:8
       run: |
@@ -193,7 +194,7 @@ jobs:
 
     - name: Run slow PEFT CUDA tests
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
         # https://pytorch.org/docs/stable/notes/randomness.html#avoiding-nondeterministic-algorithms
         CUBLAS_WORKSPACE_CONFIG: :16:8
       run: |
@@ -242,7 +243,7 @@ jobs:
 
     - name: Run slow Flax TPU tests
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m pytest -n 0 \
           -s -v -k "Flax" \
@@ -289,7 +290,7 @@ jobs:
 
     - name: Run slow ONNXRuntime CUDA tests
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile \
           -s -v -k "Onnx" \
@@ -336,7 +337,7 @@ jobs:
         python utils/print_env.py
     - name: Run example tests on GPU
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile -s -v -k "compile" --make-reports=tests_torch_compile_cuda tests/
     - name: Failure short reports
@@ -377,7 +378,7 @@ jobs:
         python utils/print_env.py
     - name: Run example tests on GPU
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile -s -v -k "xformers" --make-reports=tests_torch_xformers_cuda tests/
     - name: Failure short reports
@@ -422,9 +423,10 @@ jobs:
 
     - name: Run example tests on GPU
       env:
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         python -m venv /opt/venv && export PATH="/opt/venv/bin:$PATH"
+        python -m uv pip install timm
         python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile -s -v --make-reports=examples_torch_cuda examples/
 
     - name: Failure short reports

.github/workflows/push_tests_fast.yml

@@ -107,7 +107,7 @@ jobs:
       if: ${{ matrix.config.framework == 'pytorch_examples' }}
       run: |
         python -m venv /opt/venv && export PATH="/opt/venv/bin:$PATH"
-        python -m uv pip install peft
+        python -m uv pip install peft timm
         python -m pytest -n 4 --max-worker-restart=0 --dist=loadfile \
           --make-reports=tests_${{ matrix.config.report }} \
           examples

.github/workflows/push_tests_mps.yml

@@ -23,7 +23,7 @@ concurrency:
 jobs:
   run_fast_tests_apple_m1:
     name: Fast PyTorch MPS tests on MacOS
-    runs-on: [ self-hosted, apple-m1 ]
+    runs-on: macos-13-xlarge
 
     steps:
     - name: Checkout diffusers
@@ -59,7 +59,7 @@ jobs:
       shell: arch -arch arm64 bash {0}
       env:
         HF_HOME: /System/Volumes/Data/mnt/cache
-        HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
+        HF_TOKEN: ${{ secrets.HF_TOKEN }}
       run: |
         ${CONDA_RUN} python -m pytest -n 0 -s -v --make-reports=tests_torch_mps tests/
 

.github/workflows/run_tests_from_a_pr.yml

@@ -0,0 +1,73 @@
+name: Check running SLOW tests from a PR (only GPU)
+
+on:
+  workflow_dispatch:
+    inputs:
+      docker_image:
+        default: 'diffusers/diffusers-pytorch-cuda'
+        description: 'Name of the Docker image'
+        required: true
+      branch: 
+        description: 'PR Branch to test on'
+        required: true
+      test:
+        description: 'Tests to run (e.g.: `tests/models`).'
+        required: true
+
+env:
+  DIFFUSERS_IS_CI: yes
+  IS_GITHUB_CI: "1"
+  HF_HOME: /mnt/cache
+  OMP_NUM_THREADS: 8
+  MKL_NUM_THREADS: 8
+  PYTEST_TIMEOUT: 600
+  RUN_SLOW: yes
+
+jobs:
+  run_tests:
+    name: "Run a test on our runner from a PR"
+    runs-on: [single-gpu, nvidia-gpu, t4, ci]
+    container:
+      image: ${{ github.event.inputs.docker_image }}
+      options: --gpus 0 --privileged --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+
+    steps:
+      - name: Validate test files input
+        id: validate_test_files
+        env: 
+          PY_TEST: ${{ github.event.inputs.test }}
+        run: |
+          if [[ ! "$PY_TEST" =~ ^tests/ ]]; then
+            echo "Error: The input string must start with 'tests/'."
+            exit 1
+          fi
+          
+          if [[ ! "$PY_TEST" =~ ^tests/(models|pipelines) ]]; then
+            echo "Error: The input string must contain either 'models' or 'pipelines' after 'tests/'."
+            exit 1
+          fi
+          
+          if [[ "$PY_TEST" == *";"* ]]; then
+            echo "Error: The input string must not contain ';'."
+            exit 1
+          fi
+          echo "$PY_TEST"
+
+      - name: Checkout PR branch
+        uses: actions/checkout@v4
+        with:
+          ref: ${{ github.event.inputs.branch }}
+          repository: ${{ github.event.pull_request.head.repo.full_name }}
+
+
+      - name: Install pytest 
+        run: | 
+          python -m venv /opt/venv && export PATH="/opt/venv/bin:$PATH"
+          python -m uv pip install -e [quality,test]
+          python -m uv pip install peft
+      
+      - name: Run tests
+        env: 
+            PY_TEST: ${{ github.event.inputs.test }}
+        run: |
+          pytest "$PY_TEST"

.github/workflows/update_metadata.yml

@@ -25,6 +25,6 @@ jobs:
 
       - name: Update metadata
         env:
-          HUGGING_FACE_HUB_TOKEN: ${{ secrets.SAYAK_HF_TOKEN }}
+          HF_TOKEN: ${{ secrets.SAYAK_HF_TOKEN }}
         run: |
           python utils/update_metadata.py --commit_sha ${{ github.sha }}

CONTRIBUTING.md

@@ -355,7 +355,7 @@ You will need basic `git` proficiency to be able to contribute to
 manual. Type `git --help` in a shell and enjoy. If you prefer books, [Pro
 Git](https://git-scm.com/book/en/v2) is a very good reference.
 
-Follow these steps to start contributing ([supported Python versions](https://github.com/huggingface/diffusers/blob/main/setup.py#L265)):
+Follow these steps to start contributing ([supported Python versions](https://github.com/huggingface/diffusers/blob/42f25d601a910dceadaee6c44345896b4cfa9928/setup.py#L270)):
 
 1. Fork the [repository](https://github.com/huggingface/diffusers) by
 clicking on the 'Fork' button on the repository's page. This creates a copy of the code

README.md

@@ -77,7 +77,7 @@ Please refer to the [How to use Stable Diffusion in Apple Silicon](https://huggi
 
 ## Quickstart
 
-Generating outputs is super easy with 🤗 Diffusers. To generate an image from text, use the `from_pretrained` method to load any pretrained diffusion model (browse the [Hub](https://huggingface.co/models?library=diffusers&sort=downloads) for 22000+ checkpoints):
+Generating outputs is super easy with 🤗 Diffusers. To generate an image from text, use the `from_pretrained` method to load any pretrained diffusion model (browse the [Hub](https://huggingface.co/models?library=diffusers&sort=downloads) for 25.000+ checkpoints):
 
 ```python
 from diffusers import DiffusionPipeline
@@ -219,7 +219,7 @@ Also, say 👋 in our public Discord channel <a href="https://discord.gg/G7tWnz9
 - https://github.com/deep-floyd/IF
 - https://github.com/bentoml/BentoML
 - https://github.com/bmaltais/kohya_ss
-- +9000 other amazing GitHub repositories 💪
+- +11.000 other amazing GitHub repositories 💪
 
 Thank you for using us ❤️.
 

docker/diffusers-doc-builder/Dockerfile

@@ -0,0 +1,51 @@
+FROM ubuntu:20.04
+LABEL maintainer="Hugging Face"
+LABEL repository="diffusers"
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
+                   build-essential \
+                   git \
+                   git-lfs \
+                   curl \
+                   ca-certificates \
+                   libsndfile1-dev \
+                   python3.10 \
+                   python3-pip \
+                   libgl1 \
+                   zip \
+                   python3.10-venv && \
+    rm -rf /var/lib/apt/lists
+
+# make sure to use venv
+RUN python3.10 -m venv /opt/venv
+ENV PATH="/opt/venv/bin:$PATH"
+
+# pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)
+RUN python3.10 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
+    python3.10 -m uv pip install --no-cache-dir \
+        torch \
+        torchvision \
+        torchaudio \
+        invisible_watermark \
+        --extra-index-url https://download.pytorch.org/whl/cpu && \
+    python3.10 -m uv pip install --no-cache-dir \
+        accelerate \
+        datasets \
+        hf-doc-builder \
+        huggingface-hub \
+        Jinja2 \
+        librosa \
+        numpy \
+        scipy \
+        tensorboard \
+        transformers \
+        matplotlib \
+        setuptools==69.5.1
+
+CMD ["/bin/bash"]

docker/diffusers-flax-cpu/Dockerfile

@@ -4,22 +4,25 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
-                   build-essential \
-                   git \
-                   git-lfs \
-                   curl \
-                   ca-certificates \
-                   libsndfile1-dev \
-                   libgl1 \
-                   python3.8 \
-                   python3-pip \
-                   python3.8-venv && \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
+        build-essential \
+        git \
+        git-lfs \
+        curl \
+        ca-certificates \
+        libsndfile1-dev \
+        libgl1 \
+        python3.10 \
+        python3-pip \
+        python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)

docker/diffusers-flax-tpu/Dockerfile

@@ -4,8 +4,11 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
                    build-essential \
                    git \
                    git-lfs \
@@ -13,13 +16,13 @@ RUN apt update && \
                    ca-certificates \
                    libsndfile1-dev \
                    libgl1 \
-                   python3.8 \
+                   python3.10 \
                    python3-pip \
-                   python3.8-venv && \
+                   python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)

docker/diffusers-onnxruntime-cpu/Dockerfile

@@ -4,8 +4,11 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
                    build-essential \
                    git \
                    git-lfs \
@@ -13,13 +16,13 @@ RUN apt update && \
                    ca-certificates \
                    libsndfile1-dev \
                    libgl1 \
-                   python3.8 \
+                   python3.10 \
                    python3-pip \
-                   python3.8-venv && \
+                   python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)

docker/diffusers-onnxruntime-cuda/Dockerfile

@@ -4,8 +4,11 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
                    build-essential \
                    git \
                    git-lfs \
@@ -13,24 +16,24 @@ RUN apt update && \
                    ca-certificates \
                    libsndfile1-dev \
                    libgl1 \
-                   python3.8 \
+                   python3.10 \
                    python3-pip \
-                   python3.8-venv && \
+                   python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)
-RUN python3 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
-    python3 -m uv pip install --no-cache-dir \
+RUN python3.10 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
+    python3.10 -m uv pip install --no-cache-dir \
         torch \
         torchvision \
         torchaudio \
         "onnxruntime-gpu>=1.13.1" \
         --extra-index-url https://download.pytorch.org/whl/cu117 && \
-    python3 -m uv pip install --no-cache-dir \
+    python3.10 -m uv pip install --no-cache-dir \
         accelerate \
         datasets \
         hf-doc-builder \

docker/diffusers-pytorch-compile-cuda/Dockerfile

@@ -4,8 +4,11 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
     build-essential \
     git \
     git-lfs \
@@ -13,24 +16,23 @@ RUN apt update && \
     ca-certificates \
     libsndfile1-dev \
     libgl1 \
-    python3.9 \
-    python3.9-dev \
+    python3.10 \
     python3-pip \
-    python3.9-venv && \
+    python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3.9 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)
-RUN python3.9 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
-    python3.9 -m uv pip install --no-cache-dir \
+RUN python3.10 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
+    python3.10 -m uv pip install --no-cache-dir \
     torch \
     torchvision \
     torchaudio \
     invisible_watermark && \
-    python3.9 -m pip install --no-cache-dir \
+    python3.10 -m pip install --no-cache-dir \
     accelerate \
     datasets \
     hf-doc-builder \

docker/diffusers-pytorch-cpu/Dockerfile

@@ -4,33 +4,36 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
                    build-essential \
                    git \
                    git-lfs \
                    curl \
                    ca-certificates \
                    libsndfile1-dev \
-                   python3.8 \
+                   python3.10 \
                    python3-pip \
                    libgl1 \
-                   python3.8-venv && \
+                   python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)
-RUN python3 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
-    python3 -m uv pip install --no-cache-dir \
+RUN python3.10 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
+    python3.10 -m uv pip install --no-cache-dir \
         torch \
         torchvision \
         torchaudio \
         invisible_watermark \
         --extra-index-url https://download.pytorch.org/whl/cpu && \
-    python3 -m uv pip install --no-cache-dir \
+    python3.10 -m uv pip install --no-cache-dir \
         accelerate \
         datasets \
         hf-doc-builder \

docker/diffusers-pytorch-cuda/Dockerfile

@@ -4,8 +4,11 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
     build-essential \
     git \
     git-lfs \
@@ -13,23 +16,23 @@ RUN apt update && \
     ca-certificates \
     libsndfile1-dev \
     libgl1 \
-    python3.8 \
+    python3.10 \
     python3-pip \
-    python3.8-venv && \
+    python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)
-RUN python3 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
-    python3 -m uv pip install --no-cache-dir \
+RUN python3.10 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
+    python3.10 -m uv pip install --no-cache-dir \
     torch \
     torchvision \
     torchaudio \
     invisible_watermark && \
-    python3 -m pip install --no-cache-dir \
+    python3.10 -m pip install --no-cache-dir \
     accelerate \
     datasets \
     hf-doc-builder \

docker/diffusers-pytorch-xformers-cuda/Dockerfile

@@ -4,8 +4,11 @@ LABEL repository="diffusers"
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y bash \
+RUN apt-get -y update \
+    && apt-get install -y software-properties-common \
+    && add-apt-repository ppa:deadsnakes/ppa
+
+RUN apt install -y bash \
                    build-essential \
                    git \
                    git-lfs \
@@ -13,23 +16,23 @@ RUN apt update && \
                    ca-certificates \
                    libsndfile1-dev \
                    libgl1 \
-                   python3.8 \
+                   python3.10 \
                    python3-pip \
-                   python3.8-venv && \
+                   python3.10-venv && \
     rm -rf /var/lib/apt/lists
 
 # make sure to use venv
-RUN python3 -m venv /opt/venv
+RUN python3.10 -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
 
 # pre-install the heavy dependencies (these can later be overridden by the deps from setup.py)
-RUN python3 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
-    python3 -m pip install --no-cache-dir \
+RUN python3.10 -m pip install --no-cache-dir --upgrade pip uv==0.1.11 && \
+    python3.10 -m pip install --no-cache-dir \
         torch \
         torchvision \
         torchaudio \
         invisible_watermark && \
-    python3 -m uv pip install --no-cache-dir \
+    python3.10 -m uv pip install --no-cache-dir \
         accelerate \
         datasets \
         hf-doc-builder \

docs/README.md

@@ -242,10 +242,10 @@ Here's an example of a tuple return, comprising several objects:
 
 ```
     Returns:
-        `tuple(torch.FloatTensor)` comprising various elements depending on the configuration ([`BertConfig`]) and inputs:
-        - ** loss** (*optional*, returned when `masked_lm_labels` is provided) `torch.FloatTensor` of shape `(1,)` --
+        `tuple(torch.Tensor)` comprising various elements depending on the configuration ([`BertConfig`]) and inputs:
+        - ** loss** (*optional*, returned when `masked_lm_labels` is provided) `torch.Tensor` of shape `(1,)` --
           Total loss is the sum of the masked language modeling loss and the next sequence prediction (classification) loss.
-        - **prediction_scores** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) --
+        - **prediction_scores** (`torch.Tensor` of shape `(batch_size, sequence_length, config.vocab_size)`) --
           Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
 ```
 

docs/source/en/_toctree.yml

@@ -93,6 +93,8 @@
     title: Trajectory Consistency Distillation-LoRA
   - local: using-diffusers/svd
     title: Stable Video Diffusion
+  - local: using-diffusers/marigold_usage
+    title: Marigold Computer Vision
   title: Specific pipeline examples
 - sections:
   - local: training/overview
@@ -232,6 +234,12 @@
       title: ConsistencyDecoderVAE
     - local: api/models/transformer2d
       title: Transformer2D
+    - local: api/models/pixart_transformer2d
+      title: PixArtTransformer2D
+    - local: api/models/dit_transformer2d
+      title: DiTTransformer2D
+    - local: api/models/hunyuan_transformer_2d
+      title: HunyuanDiT2DModel
     - local: api/models/transformer_temporal
       title: Transformer Temporal
     - local: api/models/prior_transformer
@@ -279,6 +287,8 @@
       title: DiffEdit
     - local: api/pipelines/dit
       title: DiT
+    - local: api/pipelines/hunyuandit
+      title: Hunyuan-DiT
     - local: api/pipelines/i2vgenxl
       title: I2VGen-XL
     - local: api/pipelines/pix2pix
@@ -295,6 +305,8 @@
       title: Latent Diffusion
     - local: api/pipelines/ledits_pp
       title: LEDITS++
+    - local: api/pipelines/marigold
+      title: Marigold
     - local: api/pipelines/panorama
       title: MultiDiffusion
     - local: api/pipelines/musicldm
@@ -305,6 +317,8 @@
       title: Personalized Image Animator (PIA)
     - local: api/pipelines/pixart
       title: PixArt-α
+    - local: api/pipelines/pixart_sigma
+      title: PixArt-Σ
     - local: api/pipelines/self_attention_guidance
       title: Self-Attention Guidance
     - local: api/pipelines/semantic_stable_diffusion
@@ -439,6 +453,8 @@
       title: Utilities
     - local: api/image_processor
       title: VAE Image Processor
+    - local: api/video_processor
+      title: Video Processor
     title: Internal classes
     isExpanded: false
   title: API

docs/source/en/api/image_processor.md

@@ -25,3 +25,11 @@ All pipelines with [`VaeImageProcessor`] accept PIL Image, PyTorch tensor, or Nu
 The [`VaeImageProcessorLDM3D`] accepts RGB and depth inputs and returns RGB and depth outputs.
 
 [[autodoc]] image_processor.VaeImageProcessorLDM3D
+
+## PixArtImageProcessor
+
+[[autodoc]] image_processor.PixArtImageProcessor
+
+## IPAdapterMaskProcessor
+
+[[autodoc]] image_processor.IPAdapterMaskProcessor

docs/source/en/api/loaders/single_file.md

@@ -10,13 +10,134 @@ an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express o
 specific language governing permissions and limitations under the License.
 -->
 
-# Single files
+# Loading Pipelines and Models via `from_single_file`
 
-Diffusers supports loading pretrained pipeline (or model) weights stored in a single file, such as a `ckpt` or `safetensors` file. These single file types are typically produced from community trained models. There are three classes for loading single file weights:
+The `from_single_file` method allows you to load supported pipelines using a single checkpoint file as opposed to Diffusers' multiple folders format. This is useful if you are working with Stable Diffusion Web UI's (such as A1111) that rely on a single file format to distribute all the components of a model.
 
-- [`FromSingleFileMixin`] supports loading pretrained pipeline weights stored in a single file, which can either be a `ckpt` or `safetensors` file.
-- [`FromOriginalVAEMixin`] supports loading a pretrained [`AutoencoderKL`] from pretrained ControlNet weights stored in a single file, which can either be a `ckpt` or `safetensors` file.
-- [`FromOriginalControlnetMixin`] supports loading pretrained ControlNet weights stored in a single file, which can either be a `ckpt` or `safetensors` file.
+The `from_single_file` method also supports loading models in their originally distributed format. This means that supported models that have been finetuned with other services can be loaded directly into Diffusers model objects and pipelines.
+
+## Pipelines that currently support `from_single_file` loading
+
+- [`StableDiffusionPipeline`]
+- [`StableDiffusionImg2ImgPipeline`]
+- [`StableDiffusionInpaintPipeline`]
+- [`StableDiffusionControlNetPipeline`]
+- [`StableDiffusionControlNetImg2ImgPipeline`]
+- [`StableDiffusionControlNetInpaintPipeline`]
+- [`StableDiffusionUpscalePipeline`]
+- [`StableDiffusionXLPipeline`]
+- [`StableDiffusionXLImg2ImgPipeline`]
+- [`StableDiffusionXLInpaintPipeline`]
+- [`StableDiffusionXLInstructPix2PixPipeline`]
+- [`StableDiffusionXLControlNetPipeline`]
+- [`StableDiffusionXLKDiffusionPipeline`]
+- [`LatentConsistencyModelPipeline`]
+- [`LatentConsistencyModelImg2ImgPipeline`]
+- [`StableDiffusionControlNetXSPipeline`]
+- [`StableDiffusionXLControlNetXSPipeline`]
+- [`LEditsPPPipelineStableDiffusion`]
+- [`LEditsPPPipelineStableDiffusionXL`]
+- [`PIAPipeline`]
+
+## Models that currently support `from_single_file` loading
+
+- [`UNet2DConditionModel`]
+- [`StableCascadeUNet`]
+- [`AutoencoderKL`]
+- [`ControlNetModel`]
+
+## Usage Examples
+
+## Loading a Pipeline using `from_single_file`
+
+```python
+from diffusers import StableDiffusionXLPipeline
+
+ckpt_path = "https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/sd_xl_base_1.0_0.9vae.safetensors"
+pipe = StableDiffusionXLPipeline.from_single_file(ckpt_path)
+```
+
+## Setting components in a Pipeline using `from_single_file`
+
+Set components of a pipeline by passing them directly to the `from_single_file` method. For example, here we are swapping out the pipeline's default scheduler with the `DDIMScheduler`.
+
+```python
+from diffusers import StableDiffusionXLPipeline, DDIMScheduler
+
+ckpt_path = "https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/sd_xl_base_1.0_0.9vae.safetensors"
+
+scheduler = DDIMScheduler()
+pipe = StableDiffusionXLPipeline.from_single_file(ckpt_path, scheduler=scheduler)
+
+```
+
+Here we are passing in a ControlNet model to the `StableDiffusionControlNetPipeline`.
+
+```python
+from diffusers import StableDiffusionControlNetPipeline, ControlNetModel
+
+ckpt_path = "https://huggingface.co/runwayml/stable-diffusion-v1-5/blob/main/v1-5-pruned-emaonly.safetensors"
+
+controlnet = ControlNetModel.from_pretrained("lllyasviel/control_v11p_sd15_canny")
+pipe = StableDiffusionControlNetPipeline.from_single_file(ckpt_path, controlnet=controlnet)
+
+```
+
+## Loading a Model using `from_single_file`
+
+```python
+from diffusers import StableCascadeUNet
+
+ckpt_path = "https://huggingface.co/stabilityai/stable-cascade/blob/main/stage_b_lite.safetensors"
+model = StableCascadeUNet.from_single_file(ckpt_path)
+
+```
+
+## Using a Diffusers model repository to configure single file loading
+
+Under the hood, `from_single_file` will try to automatically determine a model repository to use to configure the components of a pipeline. You can also explicitly set the model repository to configure the pipeline with the `config` argument.
+
+```python
+from diffusers import StableDiffusionXLPipeline
+
+ckpt_path = "https://huggingface.co/segmind/SSD-1B/blob/main/SSD-1B.safetensors"
+repo_id = "segmind/SSD-1B"
+
+pipe = StableDiffusionXLPipeline.from_single_file(ckpt_path, config=repo_id)
+
+```
+
+In the example above, since we explicitly passed `repo_id="segmind/SSD-1B"` to the `config` argument, it will use this [configuration file](https://huggingface.co/segmind/SSD-1B/blob/main/unet/config.json) from the `unet` subfolder in `"segmind/SSD-1B"` to configure the `unet` component of the pipeline; Similarly, it will use the `config.json` file from `vae` subfolder to configure the `vae` model, `config.json` file from `text_encoder` folder to configure `text_encoder` and so on.
+
+<Tip>
+
+Most of the time you do not need to explicitly set a `config` argument. `from_single_file` will automatically map the checkpoint to the appropriate model repository. However, this option can be useful in cases where model components in the checkpoint might have been changed from what was originally distributed, or in cases where a checkpoint file might not have the necessary metadata to correctly determine the configuration to use for the pipeline.
+
+</Tip>
+
+## Override configuration options when using single file loading
+
+Override the default model or pipeline configuration options by providing the relevant arguments directly to the `from_single_file` method. Any argument supported by the model or pipeline class can be configured in this way:
+
+### Setting a pipeline configuration option
+
+```python
+from diffusers import StableDiffusionXLInstructPix2PixPipeline
+
+ckpt_path = "https://huggingface.co/stabilityai/cosxl/blob/main/cosxl_edit.safetensors"
+pipe = StableDiffusionXLInstructPix2PixPipeline.from_single_file(ckpt_path, config="diffusers/sdxl-instructpix2pix-768", is_cosxl_edit=True)
+
+```
+
+### Setting a model configuration option
+
+```python
+from diffusers import UNet2DConditionModel
+
+ckpt_path = "https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/sd_xl_base_1.0_0.9vae.safetensors"
+model = UNet2DConditionModel.from_single_file(ckpt_path, upcast_attention=True)
+
+```
 
 <Tip>
 
@@ -24,14 +145,116 @@ To learn more about how to load single file weights, see the [Load different Sta
 
 </Tip>
 
+## Working with local files
+
+As of `diffusers>=0.28.0` the `from_single_file` method will attempt to configure a pipeline or model by first inferring the model type from the keys in the checkpoint file. This inferred model type is then used to determine the appropriate model repository on the Hugging Face Hub to configure the model or pipeline.
+
+For example, any single file checkpoint based on the Stable Diffusion XL base model will use the [`stabilityai/stable-diffusion-xl-base-1.0`](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0) model repository to configure the pipeline.
+
+If you are working in an environment with restricted internet access, it is recommended that you download the config files and checkpoints for the model to your preferred directory and pass the local paths to the `pretrained_model_link_or_path` and `config` arguments of the `from_single_file` method.
+
+```python
+from huggingface_hub import hf_hub_download, snapshot_download
+
+my_local_checkpoint_path = hf_hub_download(
+    repo_id="segmind/SSD-1B",
+    filename="SSD-1B.safetensors"
+)
+
+my_local_config_path = snapshot_download(
+    repo_id="segmind/SSD-1B",
+    allowed_patterns=["*.json", "**/*.json", "*.txt", "**/*.txt"]
+)
+
+pipe = StableDiffusionXLPipeline.from_single_file(my_local_checkpoint_path, config=my_local_config_path, local_files_only=True)
+
+```
+
+By default this will download the checkpoints and config files to the [Hugging Face Hub cache directory](https://huggingface.co/docs/huggingface_hub/en/guides/manage-cache). You can also specify a local directory to download the files to by passing the `local_dir` argument to the `hf_hub_download` and `snapshot_download` functions.
+
+```python
+from huggingface_hub import hf_hub_download, snapshot_download
+
+my_local_checkpoint_path = hf_hub_download(
+    repo_id="segmind/SSD-1B",
+    filename="SSD-1B.safetensors"
+    local_dir="my_local_checkpoints"
+)
+
+my_local_config_path = snapshot_download(
+    repo_id="segmind/SSD-1B",
+    allowed_patterns=["*.json", "**/*.json", "*.txt", "**/*.txt"]
+    local_dir="my_local_config"
+)
+
+pipe = StableDiffusionXLPipeline.from_single_file(my_local_checkpoint_path, config=my_local_config_path, local_files_only=True)
+
+```
+
+## Working with local files on file systems that do not support symlinking
+
+By default the `from_single_file` method relies on the `huggingface_hub` caching mechanism to fetch and store checkpoints and config files for models and pipelines. If you are working with a file system that does not support symlinking, it is recommended that you first download the checkpoint file to a local directory and disable symlinking by passing the `local_dir_use_symlink=False` argument to the `hf_hub_download` and `snapshot_download` functions.
+
+```python
+from huggingface_hub import hf_hub_download, snapshot_download
+
+my_local_checkpoint_path = hf_hub_download(
+    repo_id="segmind/SSD-1B",
+    filename="SSD-1B.safetensors"
+    local_dir="my_local_checkpoints",
+    local_dir_use_symlinks=False
+)
+print("My local checkpoint: ", my_local_checkpoint_path)
+
+my_local_config_path = snapshot_download(
+    repo_id="segmind/SSD-1B",
+    allowed_patterns=["*.json", "**/*.json", "*.txt", "**/*.txt"]
+    local_dir_use_symlinks=False,
+)
+print("My local config: ", my_local_config_path)
+
+```
+
+Then pass the local paths to the `pretrained_model_link_or_path` and `config` arguments of the `from_single_file` method.
+
+```python
+pipe = StableDiffusionXLPipeline.from_single_file(my_local_checkpoint_path, config=my_local_config_path, local_files_only=True)
+
+```
+
+<Tip>
+
+As of `huggingface_hub>=0.23.0` the `local_dir_use_symlinks` argument isn't necessary for the `hf_hub_download` and `snapshot_download` functions.
+
+</Tip>
+
+## Using the original configuration file of a model
+
+If you would like to configure the model components in a pipeline using the orignal YAML configuration file, you can pass a local path or url to the original configuration file via the `original_config` argument.
+
+```python
+from diffusers import StableDiffusionXLPipeline
+
+ckpt_path = "https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/sd_xl_base_1.0_0.9vae.safetensors"
+repo_id = "stabilityai/stable-diffusion-xl-base-1.0"
+original_config = "https://raw.githubusercontent.com/Stability-AI/generative-models/main/configs/inference/sd_xl_base.yaml"
+
+pipe = StableDiffusionXLPipeline.from_single_file(ckpt_path, original_config=original_config)
+```
+
+<Tip>
+
+When using `original_config` with `local_files_only=True`, Diffusers will attempt to infer the components of the pipeline based on the type signatures of pipeline class, rather than attempting to fetch the configuration files from a model repository on the Hugging Face Hub. This is to prevent backward breaking changes in existing code that might not be able to connect to the internet to fetch the necessary configuration files.
+
+This is not as reliable as providing a path to a local model repository using the `config` argument and might lead to errors when configuring the pipeline. To avoid this, please run the pipeline with `local_files_only=False` once to download the appropriate pipeline configuration files to the local cache.
+
+</Tip>
+
+
 ## FromSingleFileMixin
 
 [[autodoc]] loaders.single_file.FromSingleFileMixin
 
-## FromOriginalVAEMixin
+## FromOriginalModelMixin
 
-[[autodoc]] loaders.autoencoder.FromOriginalVAEMixin
-
-## FromOriginalControlnetMixin
-
-[[autodoc]] loaders.controlnet.FromOriginalControlNetMixin
+[[autodoc]] loaders.single_file_model.FromOriginalModelMixin

docs/source/en/api/models/dit_transformer2d.md

@@ -0,0 +1,19 @@
+<!--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.
+-->
+
+# DiTTransformer2D
+
+A Transformer model for image-like data from [DiT](https://huggingface.co/papers/2212.09748).
+
+## DiTTransformer2DModel
+
+[[autodoc]] DiTTransformer2DModel

docs/source/en/api/models/hunyuan_transformer_2d.md

@@ -0,0 +1,20 @@
+<!--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.
+-->
+
+# HunyuanDiT2DModel
+
+A Diffusion Transformer model for 2D data from [Hunyuan-DiT](https://github.com/Tencent/HunyuanDiT).
+
+## HunyuanDiT2DModel
+
+[[autodoc]] HunyuanDiT2DModel
+

docs/source/en/api/models/pixart_transformer2d.md

@@ -0,0 +1,19 @@
+<!--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.
+-->
+
+# PixArtTransformer2D
+
+A Transformer model for image-like data from [PixArt-Alpha](https://huggingface.co/papers/2310.00426) and [PixArt-Sigma](https://huggingface.co/papers/2403.04692). 
+
+## PixArtTransformer2DModel
+
+[[autodoc]] PixArtTransformer2DModel

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

@@ -101,6 +101,53 @@ AnimateDiff tends to work better with finetuned Stable Diffusion models. If you
 
 </Tip>
 
+### AnimateDiffSDXLPipeline
+
+AnimateDiff can also be used with SDXL models. This is currently an experimental feature as only a beta release of the motion adapter checkpoint is available.
+
+```python
+import torch
+from diffusers.models import MotionAdapter
+from diffusers import AnimateDiffSDXLPipeline, DDIMScheduler
+from diffusers.utils import export_to_gif
+
+adapter = MotionAdapter.from_pretrained("guoyww/animatediff-motion-adapter-sdxl-beta", torch_dtype=torch.float16)
+
+model_id = "stabilityai/stable-diffusion-xl-base-1.0"
+scheduler = DDIMScheduler.from_pretrained(
+    model_id,
+    subfolder="scheduler",
+    clip_sample=False,
+    timestep_spacing="linspace",
+    beta_schedule="linear",
+    steps_offset=1,
+)
+pipe = AnimateDiffSDXLPipeline.from_pretrained(
+    model_id,
+    motion_adapter=adapter,
+    scheduler=scheduler,
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to("cuda")
+
+# enable memory savings
+pipe.enable_vae_slicing()
+pipe.enable_vae_tiling()
+
+output = pipe(
+    prompt="a panda surfing in the ocean, realistic, high quality",
+    negative_prompt="low quality, worst quality",
+    num_inference_steps=20,
+    guidance_scale=8,
+    width=1024,
+    height=1024,
+    num_frames=16,
+)
+
+frames = output.frames[0]
+export_to_gif(frames, "animation.gif")
+```
+
 ### AnimateDiffVideoToVideoPipeline
 
 AnimateDiff can also be used to generate visually similar videos or enable style/character/background or other edits starting from an initial video, allowing you to seamlessly explore creative possibilities.
@@ -522,6 +569,12 @@ export_to_gif(frames, "animatelcm-motion-lora.gif")
   - all
   - __call__
 
+## AnimateDiffSDXLPipeline
+
+[[autodoc]] AnimateDiffSDXLPipeline
+  - all
+  - __call__
+
 ## AnimateDiffVideoToVideoPipeline
 
 [[autodoc]] AnimateDiffVideoToVideoPipeline

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

@@ -0,0 +1,37 @@
+<!--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.
+-->
+
+# Hunyuan-DiT
+![chinese elements understanding](https://github.com/gnobitab/diffusers-hunyuan/assets/1157982/39b99036-c3cb-4f16-bb1a-40ec25eda573)
+
+[Hunyuan-DiT : A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding](https://arxiv.org/abs/2405.08748)] from Tencent Hunyuan.
+
+The abstract from the paper is:
+
+*We present Hunyuan-DiT, a text-to-image diffusion transformer with fine-grained understanding of both English and Chinese. To construct Hunyuan-DiT, we carefully design the transformer structure, text encoder, and positional encoding. We also build from scratch a whole data pipeline to update and evaluate data for iterative model optimization. For fine-grained language understanding, we train a Multimodal Large Language Model to refine the captions of the images. Finally, Hunyuan-DiT can perform multi-turn multimodal dialogue with users, generating and refining images according to the context. Through our holistic human evaluation protocol with more than 50 professional human evaluators, Hunyuan-DiT sets a new state-of-the-art in Chinese-to-image generation compared with other open-source models.*
+
+
+You can find the original codebase at [Tencent/HunyuanDiT](https://github.com/Tencent/HunyuanDiT) and all the available checkpoints at [Tencent-Hunyuan](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT).
+
+**Highlights**: HunyuanDiT supports Chinese/English-to-image, multi-resolution generation.
+
+HunyuanDiT has the following components:
+* It uses a diffusion transformer as the backbone
+* It combines two text encoders, a bilingual CLIP and a multilingual T5 encoder
+
+
+## HunyuanDiTPipeline
+
+[[autodoc]] HunyuanDiTPipeline
+	- all
+	- __call__
+	

docs/source/en/api/pipelines/i2vgenxl.md

@@ -47,6 +47,7 @@ Sample output with I2VGenXL:
 * Unlike SVD, it additionally accepts text prompts as inputs.
 * It can generate higher resolution videos.
 * When using the [`DDIMScheduler`] (which is default for this pipeline), less than 50 steps for inference leads to bad results.
+* This implementation is 1-stage variant of I2VGenXL. The main figure in the [I2VGen-XL](https://arxiv.org/abs/2311.04145) paper shows a 2-stage variant, however, 1-stage variant works well. See [this discussion](https://github.com/huggingface/diffusers/discussions/7952) for more details.
 
 ## I2VGenXLPipeline
 [[autodoc]] I2VGenXLPipeline

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

@@ -0,0 +1,76 @@
+<!--Copyright 2024 Marigold authors and 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.
+-->
+
+# Marigold Pipelines for Computer Vision Tasks
+
+![marigold](https://marigoldmonodepth.github.io/images/teaser_collage_compressed.jpg)
+
+Marigold was proposed in [Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation](https://huggingface.co/papers/2312.02145), a CVPR 2024 Oral paper by [Bingxin Ke](http://www.kebingxin.com/), [Anton Obukhov](https://www.obukhov.ai/), [Shengyu Huang](https://shengyuh.github.io/), [Nando Metzger](https://nandometzger.github.io/), [Rodrigo Caye Daudt](https://rcdaudt.github.io/), and [Konrad Schindler](https://scholar.google.com/citations?user=FZuNgqIAAAAJ&hl=en). 
+The idea is to repurpose the rich generative prior of Text-to-Image Latent Diffusion Models (LDMs) for traditional computer vision tasks. 
+Initially, this idea was explored to fine-tune Stable Diffusion for Monocular Depth Estimation, as shown in the teaser above. 
+Later, 
+- [Tianfu Wang](https://tianfwang.github.io/) trained the first Latent Consistency Model (LCM) of Marigold, which unlocked fast single-step inference;
+- [Kevin Qu](https://www.linkedin.com/in/kevin-qu-b3417621b/?locale=en_US) extended the approach to Surface Normals Estimation;
+- [Anton Obukhov](https://www.obukhov.ai/) contributed the pipelines and documentation into diffusers (enabled and supported by [YiYi Xu](https://yiyixuxu.github.io/) and [Sayak Paul](https://sayak.dev/)).
+
+The abstract from the paper is:
+
+*Monocular depth estimation is a fundamental computer vision task. Recovering 3D depth from a single image is geometrically ill-posed and requires scene understanding, so it is not surprising that the rise of deep learning has led to a breakthrough. The impressive progress of monocular depth estimators has mirrored the growth in model capacity, from relatively modest CNNs to large Transformer architectures. Still, monocular depth estimators tend to struggle when presented with images with unfamiliar content and layout, since their knowledge of the visual world is restricted by the data seen during training, and challenged by zero-shot generalization to new domains. This motivates us to explore whether the extensive priors captured in recent generative diffusion models can enable better, more generalizable depth estimation. We introduce Marigold, a method for affine-invariant monocular depth estimation that is derived from Stable Diffusion and retains its rich prior knowledge. The estimator can be fine-tuned in a couple of days on a single GPU using only synthetic training data. It delivers state-of-the-art performance across a wide range of datasets, including over 20% performance gains in specific cases. Project page: https://marigoldmonodepth.github.io.*
+
+## Available Pipelines
+
+Each pipeline supports one Computer Vision task, which takes an input RGB image as input and produces a *prediction* of the modality of interest, such as a depth map of the input image. 
+Currently, the following tasks are implemented:
+
+| Pipeline                                                                                                                                    | Predicted Modalities                                                                                             |                                                                       Demos                                                                        |
+|---------------------------------------------------------------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------:|
+| [MarigoldDepthPipeline](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/marigold/pipeline_marigold_depth.py)     | [Depth](https://en.wikipedia.org/wiki/Depth_map), [Disparity](https://en.wikipedia.org/wiki/Binocular_disparity) | [Fast Demo (LCM)](https://huggingface.co/spaces/prs-eth/marigold-lcm), [Slow Original Demo (DDIM)](https://huggingface.co/spaces/prs-eth/marigold) |
+| [MarigoldNormalsPipeline](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/marigold/pipeline_marigold_normals.py) | [Surface normals](https://en.wikipedia.org/wiki/Normal_mapping)                                                  |                                   [Fast Demo (LCM)](https://huggingface.co/spaces/prs-eth/marigold-normals-lcm)                                    |
+
+
+## Available Checkpoints
+
+The original checkpoints can be found under the [PRS-ETH](https://huggingface.co/prs-eth/) Hugging Face organization. 
+
+<Tip>
+
+Make sure to check out the Schedulers [guide](../../using-diffusers/schedulers) to learn how to explore the tradeoff between scheduler speed and quality, and see the [reuse components across pipelines](../../using-diffusers/loading#reuse-components-across-pipelines) section to learn how to efficiently load the same components into multiple pipelines. Also, to know more about reducing the memory usage of this pipeline, refer to the ["Reduce memory usage"] section [here](../../using-diffusers/svd#reduce-memory-usage).
+
+</Tip>
+
+<Tip warning={true}>
+
+Marigold pipelines were designed and tested only with `DDIMScheduler` and `LCMScheduler`. 
+Depending on the scheduler, the number of inference steps required to get reliable predictions varies, and there is no universal value that works best across schedulers.
+Because of that, the default value of `num_inference_steps` in the `__call__` method of the pipeline is set to `None` (see the API reference). 
+Unless set explicitly, its value will be taken from the checkpoint configuration `model_index.json`. 
+This is done to ensure high-quality predictions when calling the pipeline with just the `image` argument. 
+
+</Tip>
+
+See also Marigold [usage examples](marigold_usage).
+
+## MarigoldDepthPipeline
+[[autodoc]] MarigoldDepthPipeline
+	- all
+	- __call__
+
+## MarigoldNormalsPipeline
+[[autodoc]] MarigoldNormalsPipeline
+	- all
+	- __call__
+
+## MarigoldDepthOutput
+[[autodoc]] pipelines.marigold.pipeline_marigold_depth.MarigoldDepthOutput
+
+## MarigoldNormalsOutput
+[[autodoc]] pipelines.marigold.pipeline_marigold_normals.MarigoldNormalsOutput

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

@@ -31,7 +31,7 @@ Some notes about this pipeline:
 
 <Tip>
 
-Make sure to check out the Schedulers [guide](../../using-diffusers/schedulers) to learn how to explore the tradeoff between scheduler speed and quality, and see the [reuse components across pipelines](../../using-diffusers/loading#reuse-components-across-pipelines) section to learn how to efficiently load the same components into multiple pipelines.
+Make sure to check out the Schedulers [guide](../../using-diffusers/schedulers.md) to learn how to explore the tradeoff between scheduler speed and quality, and see the [reuse components across pipelines](../../using-diffusers/loading.md#reuse-a-pipeline) section to learn how to efficiently load the same components into multiple pipelines.
 
 </Tip>
 

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

@@ -0,0 +1,151 @@
+<!--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.
+-->
+
+# PixArt-Σ
+
+![](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/pixart/header_collage_sigma.jpg)
+
+[PixArt-Σ: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generation](https://huggingface.co/papers/2403.04692) is Junsong Chen, Jincheng Yu, Chongjian Ge, Lewei Yao, Enze Xie, Yue Wu, Zhongdao Wang, James Kwok, Ping Luo, Huchuan Lu, and Zhenguo Li.
+
+The abstract from the paper is:
+
+*In this paper, we introduce PixArt-Σ, a Diffusion Transformer model (DiT) capable of directly generating images at 4K resolution. PixArt-Σ represents a significant advancement over its predecessor, PixArt-α, offering images of markedly higher fidelity and improved alignment with text prompts. A key feature of PixArt-Σ is its training efficiency. Leveraging the foundational pre-training of PixArt-α, it evolves from the ‘weaker’ baseline to a ‘stronger’ model via incorporating higher quality data, a process we term “weak-to-strong training”. The advancements in PixArt-Σ are twofold: (1) High-Quality Training Data: PixArt-Σ incorporates superior-quality image data, paired with more precise and detailed image captions. (2) Efficient Token Compression: we propose a novel attention module within the DiT framework that compresses both keys and values, significantly improving efficiency and facilitating ultra-high-resolution image generation. Thanks to these improvements, PixArt-Σ achieves superior image quality and user prompt adherence capabilities with significantly smaller model size (0.6B parameters) than existing text-to-image diffusion models, such as SDXL (2.6B parameters) and SD Cascade (5.1B parameters). Moreover, PixArt-Σ’s capability to generate 4K images supports the creation of high-resolution posters and wallpapers, efficiently bolstering the production of highquality visual content in industries such as film and gaming.*
+
+You can find the original codebase at [PixArt-alpha/PixArt-sigma](https://github.com/PixArt-alpha/PixArt-sigma) and all the available checkpoints at [PixArt-alpha](https://huggingface.co/PixArt-alpha).
+
+Some notes about this pipeline:
+
+* It uses a Transformer backbone (instead of a UNet) for denoising. As such it has a similar architecture as [DiT](https://hf.co/docs/transformers/model_doc/dit).
+* It was trained using text conditions computed from T5. This aspect makes the pipeline better at following complex text prompts with intricate details.
+* It is good at producing high-resolution images at different aspect ratios. To get the best results, the authors recommend some size brackets which can be found [here](https://github.com/PixArt-alpha/PixArt-sigma/blob/master/diffusion/data/datasets/utils.py).
+* It rivals the quality of state-of-the-art text-to-image generation systems (as of this writing) such as PixArt-α, Stable Diffusion XL, Playground V2.0 and DALL-E 3, while being more efficient than them.
+* It shows the ability of generating super high resolution images, such as 2048px or even 4K.
+* It shows that text-to-image models can grow from a weak model to a stronger one through several improvements (VAEs, datasets, and so on.)
+
+<Tip>
+
+Make sure to check out the Schedulers [guide](../../using-diffusers/schedulers) to learn how to explore the tradeoff between scheduler speed and quality, and see the [reuse components across pipelines](../../using-diffusers/loading#reuse-a-pipeline) section to learn how to efficiently load the same components into multiple pipelines.
+
+</Tip>
+
+## Inference with under 8GB GPU VRAM
+
+Run the [`PixArtSigmaPipeline`] with under 8GB GPU VRAM by loading the text encoder in 8-bit precision. Let's walk through a full-fledged example. 
+
+First, install the [bitsandbytes](https://github.com/TimDettmers/bitsandbytes) library:
+
+```bash
+pip install -U bitsandbytes
+```
+
+Then load the text encoder in 8-bit:
+
+```python
+from transformers import T5EncoderModel
+from diffusers import PixArtSigmaPipeline
+import torch
+
+text_encoder = T5EncoderModel.from_pretrained(
+    "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS",
+    subfolder="text_encoder",
+    load_in_8bit=True,
+    device_map="auto",
+
+)
+pipe = PixArtSigmaPipeline.from_pretrained(
+    "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS",
+    text_encoder=text_encoder,
+    transformer=None,
+    device_map="balanced"
+)
+```
+
+Now, use the `pipe` to encode a prompt:
+
+```python
+with torch.no_grad():
+    prompt = "cute cat"
+    prompt_embeds, prompt_attention_mask, negative_embeds, negative_prompt_attention_mask = pipe.encode_prompt(prompt)
+```
+
+Since text embeddings have been computed, remove the `text_encoder` and `pipe` from the memory, and free up som GPU VRAM:
+
+```python
+import gc 
+
+def flush():
+    gc.collect()
+    torch.cuda.empty_cache()
+
+del text_encoder
+del pipe
+flush()
+```
+
+Then compute the latents with the prompt embeddings as inputs:
+
+```python
+pipe = PixArtSigmaPipeline.from_pretrained(
+    "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS",
+    text_encoder=None,
+    torch_dtype=torch.float16,
+).to("cuda")
+
+latents = pipe(
+    negative_prompt=None, 
+    prompt_embeds=prompt_embeds,
+    negative_prompt_embeds=negative_embeds,
+    prompt_attention_mask=prompt_attention_mask,
+    negative_prompt_attention_mask=negative_prompt_attention_mask,
+    num_images_per_prompt=1,
+    output_type="latent",
+).images
+
+del pipe.transformer
+flush()
+```
+
+<Tip>
+
+Notice that while initializing `pipe`, you're setting `text_encoder` to `None` so that it's not loaded.
+
+</Tip>
+
+Once the latents are computed, pass it off to the VAE to decode into a real image:
+
+```python
+with torch.no_grad():
+    image = pipe.vae.decode(latents / pipe.vae.config.scaling_factor, return_dict=False)[0]
+image = pipe.image_processor.postprocess(image, output_type="pil")[0]
+image.save("cat.png")
+```
+
+By deleting components you aren't using and flushing the GPU VRAM, you should be able to run [`PixArtSigmaPipeline`] with under 8GB GPU VRAM.
+
+![](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/pixart/8bits_cat.png)
+
+If you want a report of your memory-usage, run this [script](https://gist.github.com/sayakpaul/3ae0f847001d342af27018a96f467e4e).
+
+<Tip warning={true}>
+
+Text embeddings computed in 8-bit can impact the quality of the generated images because of the information loss in the representation space caused by the reduced precision. It's recommended to compare the outputs with and without 8-bit.
+
+</Tip>
+
+While loading the `text_encoder`, you set `load_in_8bit` to `True`. You could also specify `load_in_4bit` to bring your memory requirements down even further to under 7GB.
+
+## PixArtSigmaPipeline
+
+[[autodoc]] PixArtSigmaPipeline
+	- all
+	- __call__
+	

docs/source/en/api/video_processor.md

@@ -0,0 +1,21 @@
+<!--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.
+-->
+
+# Video Processor
+
+The [`VideoProcessor`] provides a unified API for video pipelines to prepare inputs for VAE encoding and post-processing outputs once they're decoded. The class inherits [`VaeImageProcessor`] so it includes transformations such as resizing, normalization, and conversion between PIL Image, PyTorch, and NumPy arrays.
+
+## VideoProcessor
+
+[[autodoc]] video_processor.VideoProcessor.preprocess_video
+
+[[autodoc]] video_processor.VideoProcessor.postprocess_video

docs/source/en/installation.md

@@ -112,7 +112,7 @@ pip install -e ".[flax]"
 
 These commands will link the folder you cloned the repository to and your Python library paths.
 Python will now look inside the folder you cloned to in addition to the normal library paths.
-For example, if your Python packages are typically installed in `~/anaconda3/envs/main/lib/python3.8/site-packages/`, Python will also search the `~/diffusers/` folder you cloned to.
+For example, if your Python packages are typically installed in `~/anaconda3/envs/main/lib/python3.10/site-packages/`, Python will also search the `~/diffusers/` folder you cloned to.
 
 <Tip warning={true}>
 

docs/source/en/optimization/fp16.md

@@ -12,7 +12,7 @@ specific language governing permissions and limitations under the License.
 
 # Speed up inference
 
-There are several ways to optimize Diffusers for inference speed, such as reducing the computational burden by lowering the data precision or using a lightweight distilled model. There are also memory-efficient attention implementations, [xFormers](xformers) and [scaled dot product attetntion](https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html) in PyTorch 2.0, that reduce memory usage which also indirectly speeds up inference. Different speed optimizations can be stacked together to get the fastest inference times.
+There are several ways to optimize Diffusers for inference speed, such as reducing the computational burden by lowering the data precision or using a lightweight distilled model. There are also memory-efficient attention implementations, [xFormers](xformers) and [scaled dot product attention](https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html) in PyTorch 2.0, that reduce memory usage which also indirectly speeds up inference. Different speed optimizations can be stacked together to get the fastest inference times.
 
 > [!TIP]
 > Optimizing for inference speed or reduced memory usage can lead to improved performance in the other category, so you should try to optimize for both whenever you can. This guide focuses on inference speed, but you can learn more about lowering memory usage in the [Reduce memory usage](memory) guide.

docs/source/en/optimization/memory.md

@@ -261,7 +261,7 @@ from dataclasses import dataclass
 
 @dataclass
 class UNet2DConditionOutput:
-    sample: torch.FloatTensor
+    sample: torch.Tensor
 
 
 pipe = StableDiffusionPipeline.from_pretrained(

docs/source/en/optimization/tgate.md

@@ -6,7 +6,7 @@ Before you begin, make sure you install T-GATE.
 
 ```bash
 pip install tgate
-pip install -U pytorch diffusers transformers accelerate DeepCache
+pip install -U torch diffusers transformers accelerate DeepCache
 ```
 
 
@@ -46,12 +46,12 @@ pipe = TgatePixArtLoader(
 
 image = pipe.tgate(
        "An alpaca made of colorful building blocks, cyberpunk.",
-        gate_step=gate_step,
+       gate_step=gate_step,
        num_inference_steps=inference_step,
 ).images[0]
 ```
 </hfoption>
-<hfoption id="Stable Diffusion XL"> 
+<hfoption id="Stable Diffusion XL">
 
 Accelerate `StableDiffusionXLPipeline` with T-GATE:
 
@@ -78,9 +78,9 @@ pipe = TgateSDXLLoader(
 ).to("cuda")
 
 image = pipe.tgate(
-        "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k.",
-        gate_step=gate_step,
-        num_inference_steps=inference_step
+       "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k.",
+       gate_step=gate_step,
+       num_inference_steps=inference_step
 ).images[0]
 ```
 </hfoption>
@@ -111,9 +111,9 @@ pipe = TgateSDXLDeepCacheLoader(
 ).to("cuda")
 
 image = pipe.tgate(
-        "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k.",
-        gate_step=gate_step,
-        num_inference_steps=inference_step
+       "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k.",
+       gate_step=gate_step,
+       num_inference_steps=inference_step
 ).images[0]
 ```
 </hfoption>
@@ -151,9 +151,9 @@ pipe = TgateSDXLLoader(
 ).to("cuda")
 
 image = pipe.tgate(
-        "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k.",
-        gate_step=gate_step,
-        num_inference_steps=inference_step
+       "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k.",
+       gate_step=gate_step,
+       num_inference_steps=inference_step
 ).images[0]
 ```
 </hfoption>

docs/source/en/training/kandinsky.md

@@ -205,7 +205,7 @@ model_pred = unet(noisy_latents, timesteps, None, added_cond_kwargs=added_cond_k
 
 Once you’ve made all your changes or you’re okay with the default configuration, you’re ready to launch the training script! 🚀
 
-You'll train on the [Pokémon BLIP captions](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) dataset to generate your own Pokémon, but you can also create and train on your own dataset by following the [Create a dataset for training](create_dataset) guide. Set the environment variable `DATASET_NAME` to the name of the dataset on the Hub or if you're training on your own files, set the environment variable `TRAIN_DIR` to a path to your dataset.
+You'll train on the [Naruto BLIP captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) dataset to generate your own Naruto characters, but you can also create and train on your own dataset by following the [Create a dataset for training](create_dataset) guide. Set the environment variable `DATASET_NAME` to the name of the dataset on the Hub or if you're training on your own files, set the environment variable `TRAIN_DIR` to a path to your dataset.
 
 If you’re training on more than one GPU, add the `--multi_gpu` parameter to the `accelerate launch` command.
 
@@ -219,7 +219,7 @@ To monitor training progress with Weights & Biases, add the `--report_to=wandb`
 <hfoption id="prior model">
 
 ```bash
-export DATASET_NAME="lambdalabs/pokemon-blip-captions"
+export DATASET_NAME="lambdalabs/naruto-blip-captions"
 
 accelerate launch --mixed_precision="fp16"  train_text_to_image_prior.py \
   --dataset_name=$DATASET_NAME \
@@ -232,17 +232,17 @@ accelerate launch --mixed_precision="fp16"  train_text_to_image_prior.py \
   --checkpoints_total_limit=3 \
   --lr_scheduler="constant" \
   --lr_warmup_steps=0 \
-  --validation_prompts="A robot pokemon, 4k photo" \
+  --validation_prompts="A robot naruto, 4k photo" \
   --report_to="wandb" \
   --push_to_hub \
-  --output_dir="kandi2-prior-pokemon-model"
+  --output_dir="kandi2-prior-naruto-model"
 ```
 
 </hfoption>
 <hfoption id="decoder model">
 
 ```bash
-export DATASET_NAME="lambdalabs/pokemon-blip-captions"
+export DATASET_NAME="lambdalabs/naruto-blip-captions"
 
 accelerate launch --mixed_precision="fp16"  train_text_to_image_decoder.py \
   --dataset_name=$DATASET_NAME \
@@ -256,10 +256,10 @@ accelerate launch --mixed_precision="fp16"  train_text_to_image_decoder.py \
   --checkpoints_total_limit=3 \
   --lr_scheduler="constant" \
   --lr_warmup_steps=0 \
-  --validation_prompts="A robot pokemon, 4k photo" \
+  --validation_prompts="A robot naruto, 4k photo" \
   --report_to="wandb" \
   --push_to_hub \
-  --output_dir="kandi2-decoder-pokemon-model"
+  --output_dir="kandi2-decoder-naruto-model"
 ```
 
 </hfoption>
@@ -279,7 +279,7 @@ prior_components = {"prior_" + k: v for k,v in prior_pipeline.components.items()
 pipeline = AutoPipelineForText2Image.from_pretrained("kandinsky-community/kandinsky-2-2-decoder", **prior_components, torch_dtype=torch.float16)
 
 pipe.enable_model_cpu_offload()
-prompt="A robot pokemon, 4k photo"
+prompt="A robot naruto, 4k photo"
 image = pipeline(prompt=prompt, negative_prompt=negative_prompt).images[0]
 ```
 
@@ -299,7 +299,7 @@ import torch
 pipeline = AutoPipelineForText2Image.from_pretrained("path/to/saved/model", torch_dtype=torch.float16)
 pipeline.enable_model_cpu_offload()
 
-prompt="A robot pokemon, 4k photo"
+prompt="A robot naruto, 4k photo"
 image = pipeline(prompt=prompt).images[0]
 ```
 
@@ -313,7 +313,7 @@ unet = UNet2DConditionModel.from_pretrained("path/to/saved/model" + "/checkpoint
 pipeline = AutoPipelineForText2Image.from_pretrained("kandinsky-community/kandinsky-2-2-decoder", unet=unet, torch_dtype=torch.float16)
 pipeline.enable_model_cpu_offload()
 
-image = pipeline(prompt="A robot pokemon, 4k photo").images[0]
+image = pipeline(prompt="A robot naruto, 4k photo").images[0]
 ```
 
 </hfoption>

docs/source/en/training/lora.md

@@ -170,7 +170,7 @@ Aside from setting up the LoRA layers, the training script is more or less the s
 
 Once you've made all your changes or you're okay with the default configuration, you're ready to launch the training script! 🚀
 
-Let's train on the [Pokémon BLIP captions](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) dataset to generate our own Pokémon. Set the environment variables `MODEL_NAME` and `DATASET_NAME` to the model and dataset respectively. You should also specify where to save the model in `OUTPUT_DIR`, and the name of the model to save to on the Hub with `HUB_MODEL_ID`. The script creates and saves the following files to your repository:
+Let's train on the [Naruto BLIP captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) dataset to generate your own Naruto characters. Set the environment variables `MODEL_NAME` and `DATASET_NAME` to the model and dataset respectively. You should also specify where to save the model in `OUTPUT_DIR`, and the name of the model to save to on the Hub with `HUB_MODEL_ID`. The script creates and saves the following files to your repository:
 
 - saved model checkpoints
 - `pytorch_lora_weights.safetensors` (the trained LoRA weights)
@@ -185,9 +185,9 @@ A full training run takes ~5 hours on a 2080 Ti GPU with 11GB of VRAM.
 
 ```bash
 export MODEL_NAME="runwayml/stable-diffusion-v1-5"
-export OUTPUT_DIR="/sddata/finetune/lora/pokemon"
-export HUB_MODEL_ID="pokemon-lora"
-export DATASET_NAME="lambdalabs/pokemon-blip-captions"
+export OUTPUT_DIR="/sddata/finetune/lora/naruto"
+export HUB_MODEL_ID="naruto-lora"
+export DATASET_NAME="lambdalabs/naruto-blip-captions"
 
 accelerate launch --mixed_precision="fp16"  train_text_to_image_lora.py \
   --pretrained_model_name_or_path=$MODEL_NAME \
@@ -208,7 +208,7 @@ accelerate launch --mixed_precision="fp16"  train_text_to_image_lora.py \
   --hub_model_id=${HUB_MODEL_ID} \
   --report_to=wandb \
   --checkpointing_steps=500 \
-  --validation_prompt="A pokemon with blue eyes." \
+  --validation_prompt="A naruto with blue eyes." \
   --seed=1337
 ```
 
@@ -220,7 +220,7 @@ import torch
 
 pipeline = AutoPipelineForText2Image.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16).to("cuda")
 pipeline.load_lora_weights("path/to/lora/model", weight_name="pytorch_lora_weights.safetensors")
-image = pipeline("A pokemon with blue eyes").images[0]
+image = pipeline("A naruto with blue eyes").images[0]
 ```
 
 ## Next steps

docs/source/en/training/sdxl.md

@@ -176,7 +176,7 @@ If you want to learn more about how the training loop works, check out the [Unde
 
 Once you’ve made all your changes or you’re okay with the default configuration, you’re ready to launch the training script! 🚀
 
-Let’s train on the [Pokémon BLIP captions](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) dataset to generate your own Pokémon. Set the environment variables `MODEL_NAME` and `DATASET_NAME` to the model and the dataset (either from the Hub or a local path). You should also specify a VAE other than the SDXL VAE (either from the Hub or a local path) with `VAE_NAME` to avoid numerical instabilities.
+Let’s train on the [Naruto BLIP captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) dataset to generate your own Naruto characters. Set the environment variables `MODEL_NAME` and `DATASET_NAME` to the model and the dataset (either from the Hub or a local path). You should also specify a VAE other than the SDXL VAE (either from the Hub or a local path) with `VAE_NAME` to avoid numerical instabilities.
 
 <Tip>
 
@@ -187,7 +187,7 @@ To monitor training progress with Weights & Biases, add the `--report_to=wandb`
 ```bash
 export MODEL_NAME="stabilityai/stable-diffusion-xl-base-1.0"
 export VAE_NAME="madebyollin/sdxl-vae-fp16-fix"
-export DATASET_NAME="lambdalabs/pokemon-blip-captions"
+export DATASET_NAME="lambdalabs/naruto-blip-captions"
 
 accelerate launch train_text_to_image_sdxl.py \
   --pretrained_model_name_or_path=$MODEL_NAME \
@@ -211,7 +211,7 @@ accelerate launch train_text_to_image_sdxl.py \
   --validation_prompt="a cute Sundar Pichai creature" \
   --validation_epochs 5 \
   --checkpointing_steps=5000 \
-  --output_dir="sdxl-pokemon-model" \
+  --output_dir="sdxl-naruto-model" \
   --push_to_hub
 ```
 
@@ -226,9 +226,9 @@ import torch
 
 pipeline = DiffusionPipeline.from_pretrained("path/to/your/model", torch_dtype=torch.float16).to("cuda")
 
-prompt = "A pokemon with green eyes and red legs."
+prompt = "A naruto with green eyes and red legs."
 image = pipeline(prompt, num_inference_steps=30, guidance_scale=7.5).images[0]
-image.save("pokemon.png")
+image.save("naruto.png")
 ```
 
 </hfoption>
@@ -244,11 +244,11 @@ import torch_xla.core.xla_model as xm
 device = xm.xla_device()
 pipeline = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0").to(device)
 
-prompt = "A pokemon with green eyes and red legs."
+prompt = "A naruto with green eyes and red legs."
 start = time()
 image = pipeline(prompt, num_inference_steps=inference_steps).images[0]
 print(f'Compilation time is {time()-start} sec')
-image.save("pokemon.png")
+image.save("naruto.png")
 
 start = time()
 image = pipeline(prompt, num_inference_steps=inference_steps).images[0]

docs/source/en/training/text2image.md

@@ -158,7 +158,7 @@ Once you've made all your changes or you're okay with the default configuration,
 <hfoptions id="training-inference">
 <hfoption id="PyTorch">
 
-Let's train on the [Pokémon BLIP captions](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) dataset to generate your own Pokémon. Set the environment variables `MODEL_NAME` and `dataset_name` to the model and the dataset (either from the Hub or a local path). If you're training on more than one GPU, add the `--multi_gpu` parameter to the `accelerate launch` command.
+Let's train on the [Naruto BLIP captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) dataset to generate your own Naruto characters. Set the environment variables `MODEL_NAME` and `dataset_name` to the model and the dataset (either from the Hub or a local path). If you're training on more than one GPU, add the `--multi_gpu` parameter to the `accelerate launch` command.
 
 <Tip>
 
@@ -168,7 +168,7 @@ To train on a local dataset, set the `TRAIN_DIR` and `OUTPUT_DIR` environment va
 
 ```bash
 export MODEL_NAME="runwayml/stable-diffusion-v1-5"
-export dataset_name="lambdalabs/pokemon-blip-captions"
+export dataset_name="lambdalabs/naruto-blip-captions"
 
 accelerate launch --mixed_precision="fp16"  train_text_to_image.py \
   --pretrained_model_name_or_path=$MODEL_NAME \
@@ -183,7 +183,7 @@ accelerate launch --mixed_precision="fp16"  train_text_to_image.py \
   --max_grad_norm=1 \
   --enable_xformers_memory_efficient_attention
   --lr_scheduler="constant" --lr_warmup_steps=0 \
-  --output_dir="sd-pokemon-model" \
+  --output_dir="sd-naruto-model" \
   --push_to_hub
 ```
 
@@ -202,7 +202,7 @@ To train on a local dataset, set the `TRAIN_DIR` and `OUTPUT_DIR` environment va
 
 ```bash
 export MODEL_NAME="runwayml/stable-diffusion-v1-5"
-export dataset_name="lambdalabs/pokemon-blip-captions"
+export dataset_name="lambdalabs/naruto-blip-captions"
 
 python train_text_to_image_flax.py \
   --pretrained_model_name_or_path=$MODEL_NAME \
@@ -212,7 +212,7 @@ python train_text_to_image_flax.py \
   --max_train_steps=15000 \
   --learning_rate=1e-05 \
   --max_grad_norm=1 \
-  --output_dir="sd-pokemon-model" \
+  --output_dir="sd-naruto-model" \
   --push_to_hub
 ```
 
@@ -231,7 +231,7 @@ import torch
 pipeline = StableDiffusionPipeline.from_pretrained("path/to/saved_model", torch_dtype=torch.float16, use_safetensors=True).to("cuda")
 
 image = pipeline(prompt="yoda").images[0]
-image.save("yoda-pokemon.png")
+image.save("yoda-naruto.png")
 ```
 
 </hfoption>
@@ -246,7 +246,7 @@ from diffusers import FlaxStableDiffusionPipeline
 
 pipeline, params = FlaxStableDiffusionPipeline.from_pretrained("path/to/saved_model", dtype=jax.numpy.bfloat16)
 
-prompt = "yoda pokemon"
+prompt = "yoda naruto"
 prng_seed = jax.random.PRNGKey(0)
 num_inference_steps = 50
 
@@ -261,7 +261,7 @@ prompt_ids = shard(prompt_ids)
 
 images = pipeline(prompt_ids, params, prng_seed, num_inference_steps, jit=True).images
 images = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:])))
-image.save("yoda-pokemon.png")
+image.save("yoda-naruto.png")
 ```
 
 </hfoption>

docs/source/en/training/wuerstchen.md

@@ -131,7 +131,7 @@ If you want to learn more about how the training loop works, check out the [Unde
 
 Once you’ve made all your changes or you’re okay with the default configuration, you’re ready to launch the training script! 🚀
 
-Set the `DATASET_NAME` environment variable to the dataset name from the Hub. This guide uses the [Pokémon BLIP captions](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) dataset, but you can create and train on your own datasets as well (see the [Create a dataset for training](create_dataset) guide).
+Set the `DATASET_NAME` environment variable to the dataset name from the Hub. This guide uses the [Naruto BLIP captions](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) dataset, but you can create and train on your own datasets as well (see the [Create a dataset for training](create_dataset) guide).
 
 <Tip>
 
@@ -140,7 +140,7 @@ To monitor training progress with Weights & Biases, add the `--report_to=wandb`
 </Tip>
 
 ```bash
-export DATASET_NAME="lambdalabs/pokemon-blip-captions"
+export DATASET_NAME="lambdalabs/naruto-blip-captions"
 
 accelerate launch  train_text_to_image_prior.py \
   --mixed_precision="fp16" \
@@ -156,10 +156,10 @@ accelerate launch  train_text_to_image_prior.py \
   --checkpoints_total_limit=3 \
   --lr_scheduler="constant" \
   --lr_warmup_steps=0 \
-  --validation_prompts="A robot pokemon, 4k photo" \
+  --validation_prompts="A robot naruto, 4k photo" \
   --report_to="wandb" \
   --push_to_hub \
-  --output_dir="wuerstchen-prior-pokemon-model"
+  --output_dir="wuerstchen-prior-naruto-model"
 ```
 
 Once training is complete, you can use your newly trained model for inference!
@@ -171,7 +171,7 @@ from diffusers.pipelines.wuerstchen import DEFAULT_STAGE_C_TIMESTEPS
 
 pipeline = AutoPipelineForText2Image.from_pretrained("path/to/saved/model", torch_dtype=torch.float16).to("cuda")
 
-caption = "A cute bird pokemon holding a shield"
+caption = "A cute bird naruto holding a shield"
 images = pipeline(
     caption,
     width=1024,

docs/source/en/tutorials/basic_training.md

@@ -260,7 +260,7 @@ Then, you'll need a way to evaluate the model. For evaluation, you can use the [
 ...     # The default pipeline output type is `List[PIL.Image]`
 ...     images = pipeline(
 ...         batch_size=config.eval_batch_size,
-...         generator=torch.manual_seed(config.seed),
+...         generator=torch.Generator(device='cpu').manual_seed(config.seed), # Use a separate torch generator to avoid rewinding the random state of the main training loop
 ...     ).images
 
 ...     # Make a grid out of the images

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

@@ -19,13 +19,74 @@ The denoising loop of a pipeline can be modified with custom defined functions u
 
 This guide will demonstrate how callbacks work by a few features you can implement with them.
 
+## Official callbacks
+
+We provide a list of callbacks you can plug into an existing pipeline and modify the denoising loop. This is the current list of official callbacks:
+
+- `SDCFGCutoffCallback`: Disables the CFG after a certain number of steps for all SD 1.5 pipelines, including text-to-image, image-to-image, inpaint, and controlnet.
+- `SDXLCFGCutoffCallback`: Disables the CFG after a certain number of steps for all SDXL pipelines, including text-to-image, image-to-image, inpaint, and controlnet.
+- `IPAdapterScaleCutoffCallback`: Disables the IP Adapter after a certain number of steps for all pipelines supporting IP-Adapter.
+
+> [!TIP]
+> If you want to add a new official callback, feel free to open a [feature request](https://github.com/huggingface/diffusers/issues/new/choose) or [submit a PR](https://huggingface.co/docs/diffusers/main/en/conceptual/contribution#how-to-open-a-pr).
+
+To set up a callback, you need to specify the number of denoising steps after which the callback comes into effect. You can do so by using either one of these two arguments
+
+- `cutoff_step_ratio`: Float number with the ratio of the steps.
+- `cutoff_step_index`: Integer number with the exact number of the step.
+
+```python
+import torch
+
+from diffusers import DPMSolverMultistepScheduler, StableDiffusionXLPipeline
+from diffusers.callbacks import SDXLCFGCutoffCallback
+
+
+callback = SDXLCFGCutoffCallback(cutoff_step_ratio=0.4)
+# can also be used with cutoff_step_index
+# callback = SDXLCFGCutoffCallback(cutoff_step_ratio=None, cutoff_step_index=10)
+
+pipeline = StableDiffusionXLPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-xl-base-1.0",
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to("cuda")
+pipeline.scheduler = DPMSolverMultistepScheduler.from_config(pipeline.scheduler.config, use_karras_sigmas=True)
+
+prompt = "a sports car at the road, best quality, high quality, high detail, 8k resolution"
+
+generator = torch.Generator(device="cpu").manual_seed(2628670641)
+
+out = pipeline(
+    prompt=prompt,
+    negative_prompt="",
+    guidance_scale=6.5,
+    num_inference_steps=25,
+    generator=generator,
+    callback_on_step_end=callback,
+)
+
+out.images[0].save("official_callback.png")
+```
+
+<div class="flex gap-4">
+  <div>
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/without_cfg_callback.png" alt="generated image of a sports car at the road" />
+    <figcaption class="mt-2 text-center text-sm text-gray-500">without SDXLCFGCutoffCallback</figcaption>
+  </div>
+  <div>
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/with_cfg_callback.png" alt="generated image of a a sports car at the road with cfg callback" />
+    <figcaption class="mt-2 text-center text-sm text-gray-500">with SDXLCFGCutoffCallback</figcaption>
+  </div>
+</div>
+
 ## Dynamic classifier-free guidance
 
 Dynamic classifier-free guidance (CFG) is a feature that allows you to disable CFG after a certain number of inference steps which can help you save compute with minimal cost to performance. The callback function for this should have the following arguments:
 
-* `pipeline` (or the pipeline instance) provides access to important properties such as `num_timesteps` and `guidance_scale`. You can modify these properties by updating the underlying attributes. For this example, you'll disable CFG by setting `pipeline._guidance_scale=0.0`.
-* `step_index` and `timestep` tell you where you are in the denoising loop. Use `step_index` to turn off CFG after reaching 40% of `num_timesteps`.
-* `callback_kwargs` is a dict that contains tensor variables you can modify during the denoising loop. It only includes variables specified in the `callback_on_step_end_tensor_inputs` argument, which is passed to the pipeline's `__call__` method. Different pipelines may use different sets of variables, so please check a pipeline's `_callback_tensor_inputs` attribute for the list of variables you can modify. Some common variables include `latents` and `prompt_embeds`. For this function, change the batch size of `prompt_embeds` after setting `guidance_scale=0.0` in order for it to work properly.
+- `pipeline` (or the pipeline instance) provides access to important properties such as `num_timesteps` and `guidance_scale`. You can modify these properties by updating the underlying attributes. For this example, you'll disable CFG by setting `pipeline._guidance_scale=0.0`.
+- `step_index` and `timestep` tell you where you are in the denoising loop. Use `step_index` to turn off CFG after reaching 40% of `num_timesteps`.
+- `callback_kwargs` is a dict that contains tensor variables you can modify during the denoising loop. It only includes variables specified in the `callback_on_step_end_tensor_inputs` argument, which is passed to the pipeline's `__call__` method. Different pipelines may use different sets of variables, so please check a pipeline's `_callback_tensor_inputs` attribute for the list of variables you can modify. Some common variables include `latents` and `prompt_embeds`. For this function, change the batch size of `prompt_embeds` after setting `guidance_scale=0.0` in order for it to work properly.
 
 Your callback function should look something like this:
 

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

@@ -78,7 +78,7 @@ image = pipe(
     prompt=prompt,
     num_inference_steps=4,
     guidance_scale=0,
-    eta=0.3, 
+    eta=0.3,
     generator=torch.Generator(device=device).manual_seed(0),
 ).images[0]
 ```
@@ -156,14 +156,14 @@ image = pipe(
     prompt=prompt,
     num_inference_steps=8,
     guidance_scale=0,
-    eta=0.3, 
+    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. 
+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.
@@ -171,7 +171,7 @@ TCD-LoRA also supports other LoRAs trained on different styles. For example, let
 ```python
 import torch
 from diffusers import StableDiffusionXLPipeline
-from scheduling_tcd import TCDScheduler 
+from scheduling_tcd import TCDScheduler
 
 device = "cuda"
 base_model_id = "stabilityai/stable-diffusion-xl-base-1.0"
@@ -191,7 +191,7 @@ image = pipe(
     prompt=prompt,
     num_inference_steps=4,
     guidance_scale=0,
-    eta=0.3, 
+    eta=0.3,
     generator=torch.Generator(device=device).manual_seed(0),
 ).images[0]
 ```
@@ -215,7 +215,7 @@ 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 
+from scheduling_tcd import TCDScheduler
 
 device = "cuda"
 depth_estimator = DPTForDepthEstimation.from_pretrained("Intel/dpt-hybrid-midas").to(device)
@@ -249,13 +249,13 @@ 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)
@@ -271,9 +271,9 @@ 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, 
+    prompt,
+    image=depth_image,
+    num_inference_steps=4,
     guidance_scale=0,
     eta=0.3,
     controlnet_conditioning_scale=controlnet_conditioning_scale,
@@ -290,7 +290,7 @@ grid_image = make_image_grid([depth_image, image], rows=1, cols=2)
 import torch
 from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline
 from diffusers.utils import load_image, make_image_grid
-from scheduling_tcd import TCDScheduler 
+from scheduling_tcd import TCDScheduler
 
 device = "cuda"
 base_model_id = "stabilityai/stable-diffusion-xl-base-1.0"
@@ -301,13 +301,13 @@ 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)
@@ -322,9 +322,9 @@ canny_image = load_image("https://huggingface.co/datasets/hf-internal-testing/di
 controlnet_conditioning_scale = 0.5  # recommended for good generalization
 
 image = pipe(
-    prompt, 
-    image=canny_image, 
-    num_inference_steps=4, 
+    prompt,
+    image=canny_image,
+    num_inference_steps=4,
     guidance_scale=0,
     eta=0.3,
     controlnet_conditioning_scale=controlnet_conditioning_scale,
@@ -336,7 +336,7 @@ 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. 
+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>
@@ -350,7 +350,7 @@ from diffusers import StableDiffusionXLPipeline
 from diffusers.utils import load_image, make_image_grid
 
 from ip_adapter import IPAdapterXL
-from scheduling_tcd import TCDScheduler 
+from scheduling_tcd import TCDScheduler
 
 device = "cuda"
 base_model_path = "stabilityai/stable-diffusion-xl-base-1.0"
@@ -359,8 +359,8 @@ 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, 
+    base_model_path,
+    torch_dtype=torch.float16,
     variant="fp16"
 )
 pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config)
@@ -375,13 +375,13 @@ ref_image = load_image("https://raw.githubusercontent.com/tencent-ailab/IP-Adapt
 prompt = "best quality, high quality, wearing sunglasses"
 
 image = ip_model.generate(
-    pil_image=ref_image, 
+    pil_image=ref_image,
     prompt=prompt,
     scale=0.5,
-    num_samples=1, 
-    num_inference_steps=4, 
+    num_samples=1,
+    num_inference_steps=4,
     guidance_scale=0,
-    eta=0.3, 
+    eta=0.3,
     seed=0,
 )[0]
 

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

@@ -230,7 +230,7 @@ from diffusers.utils import load_image, make_image_grid
 
 pipeline = AutoPipelineForInpainting.from_pretrained(
     "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, variant="fp16"
-).to("cuda")
+)
 pipeline.enable_model_cpu_offload()
 # remove following line if xFormers is not installed or you have PyTorch 2.0 or higher installed
 pipeline.enable_xformers_memory_efficient_attention()
@@ -255,7 +255,7 @@ from diffusers.utils import load_image, make_image_grid
 
 pipeline = AutoPipelineForInpainting.from_pretrained(
     "runwayml/stable-diffusion-inpainting", torch_dtype=torch.float16, variant="fp16"
-).to("cuda")
+)
 pipeline.enable_model_cpu_offload()
 # remove following line if xFormers is not installed or you have PyTorch 2.0 or higher installed
 pipeline.enable_xformers_memory_efficient_attention()
@@ -296,7 +296,7 @@ from diffusers.utils import load_image, make_image_grid
 
 pipeline = AutoPipelineForInpainting.from_pretrained(
     "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16, variant="fp16"
-).to("cuda")
+)
 pipeline.enable_model_cpu_offload()
 # remove following line if xFormers is not installed or you have PyTorch 2.0 or higher installed
 pipeline.enable_xformers_memory_efficient_attention()
@@ -319,7 +319,7 @@ from diffusers.utils import load_image, make_image_grid
 
 pipeline = AutoPipelineForInpainting.from_pretrained(
     "runwayml/stable-diffusion-inpainting", torch_dtype=torch.float16, variant="fp16"
-).to("cuda")
+)
 pipeline.enable_model_cpu_offload()
 # remove following line if xFormers is not installed or you have PyTorch 2.0 or higher installed
 pipeline.enable_xformers_memory_efficient_attention()

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

@@ -0,0 +1,466 @@
+<!--Copyright 2024 Marigold authors and 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.
+-->
+
+# Marigold Pipelines for Computer Vision Tasks
+
+[Marigold](../api/pipelines/marigold) is a novel diffusion-based dense prediction approach, and a set of pipelines for various computer vision tasks, such as monocular depth estimation.
+
+This guide will show you how to use Marigold to obtain fast and high-quality predictions for images and videos. 
+
+Each pipeline supports one Computer Vision task, which takes an input RGB image as input and produces a *prediction* of the modality of interest, such as a depth map of the input image. 
+Currently, the following tasks are implemented:
+
+| Pipeline                                                                                                                                    | Predicted Modalities                                                                                             |                                                                       Demos                                                                        |
+|---------------------------------------------------------------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------:|
+| [MarigoldDepthPipeline](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/marigold/pipeline_marigold_depth.py)     | [Depth](https://en.wikipedia.org/wiki/Depth_map), [Disparity](https://en.wikipedia.org/wiki/Binocular_disparity) | [Fast Demo (LCM)](https://huggingface.co/spaces/prs-eth/marigold-lcm), [Slow Original Demo (DDIM)](https://huggingface.co/spaces/prs-eth/marigold) |
+| [MarigoldNormalsPipeline](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/marigold/pipeline_marigold_normals.py) | [Surface normals](https://en.wikipedia.org/wiki/Normal_mapping)                                                  |                                   [Fast Demo (LCM)](https://huggingface.co/spaces/prs-eth/marigold-normals-lcm)                                    |
+
+The original checkpoints can be found under the [PRS-ETH](https://huggingface.co/prs-eth/) Hugging Face organization. 
+These checkpoints are meant to work with diffusers pipelines and the [original codebase](https://github.com/prs-eth/marigold). 
+The original code can also be used to train new checkpoints.
+
+| Checkpoint                                                                                    | Modality | Comment                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
+|-----------------------------------------------------------------------------------------------|----------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| [prs-eth/marigold-v1-0](https://huggingface.co/prs-eth/marigold-v1-0)                         | Depth    | The first Marigold Depth checkpoint, which predicts *affine-invariant depth* maps. The performance of this checkpoint in benchmarks was studied in the original [paper](https://huggingface.co/papers/2312.02145). Designed to be used with the `DDIMScheduler` at inference, it requires at least 10 steps to get reliable predictions. Affine-invariant depth prediction has a range of values in each pixel between 0 (near plane) and 1 (far plane); both planes are chosen by the model as part of the inference process. See the `MarigoldImageProcessor` reference for visualization utilities. |
+| [prs-eth/marigold-depth-lcm-v1-0](https://huggingface.co/prs-eth/marigold-depth-lcm-v1-0)     | Depth    | The fast Marigold Depth checkpoint, fine-tuned from `prs-eth/marigold-v1-0`. Designed to be used with the `LCMScheduler` at inference, it requires as little as 1 step to get reliable predictions. The prediction reliability saturates at 4 steps and declines after that.                                                                                                                                                                                                                                                                                                                           |
+| [prs-eth/marigold-normals-v0-1](https://huggingface.co/prs-eth/marigold-normals-v0-1)         | Normals  | A preview checkpoint for the Marigold Normals pipeline. Designed to be used with the `DDIMScheduler` at inference, it requires at least 10 steps to get reliable predictions. The surface normals predictions are unit-length 3D vectors with values in the range from -1 to 1. *This checkpoint will be phased out after the release of `v1-0` version.*                                                                                                                                                                                                                                              |
+| [prs-eth/marigold-normals-lcm-v0-1](https://huggingface.co/prs-eth/marigold-normals-lcm-v0-1) | Normals  | The fast Marigold Normals checkpoint, fine-tuned from `prs-eth/marigold-normals-v0-1`. Designed to be used with the `LCMScheduler` at inference, it requires as little as 1 step to get reliable predictions. The prediction reliability saturates at 4 steps and declines after that. *This checkpoint will be phased out after the release of `v1-0` version.*                                                                                                                                                                                                                                       |
+The examples below are mostly given for depth prediction, but they can be universally applied with other supported modalities. 
+We showcase the predictions using the same input image of Albert Einstein generated by Midjourney. 
+This makes it easier to compare visualizations of the predictions across various modalities and checkpoints.
+
+<div class="flex gap-4" style="justify-content: center; width: 100%;">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://marigoldmonodepth.github.io/images/einstein.jpg"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Example input image for all Marigold pipelines
+	</figcaption>
+  </div>
+</div>
+
+### Depth Prediction Quick Start
+
+To get the first depth prediction, load `prs-eth/marigold-depth-lcm-v1-0` checkpoint into `MarigoldDepthPipeline` pipeline, put the image through the pipeline, and save the predictions: 
+
+```python
+import diffusers
+import torch
+
+pipe = diffusers.MarigoldDepthPipeline.from_pretrained(
+    "prs-eth/marigold-depth-lcm-v1-0", variant="fp16", torch_dtype=torch.float16
+).to("cuda")
+
+image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+depth = pipe(image)
+
+vis = pipe.image_processor.visualize_depth(depth.prediction)
+vis[0].save("einstein_depth.png")
+
+depth_16bit = pipe.image_processor.export_depth_to_16bit_png(depth.prediction)
+depth_16bit[0].save("einstein_depth_16bit.png")
+```
+
+The visualization function for depth [`~pipelines.marigold.marigold_image_processing.MarigoldImageProcessor.visualize_depth`] applies one of [matplotlib's colormaps](https://matplotlib.org/stable/users/explain/colors/colormaps.html) (`Spectral` by default) to map the predicted pixel values from a single-channel `[0, 1]` depth range into an RGB image. 
+With the `Spectral` colormap, pixels with near depth are painted red, and far pixels are assigned blue color.
+The 16-bit PNG file stores the single channel values mapped linearly from the `[0, 1]` range into `[0, 65535]`.
+Below are the raw and the visualized predictions; as can be seen, dark areas (mustache) are easier to distinguish in the visualization:
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_lcm_depth_16bit.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Predicted depth (16-bit PNG)
+	</figcaption>
+  </div>
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_lcm_depth.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Predicted depth visualization (Spectral)
+	</figcaption>
+  </div>
+</div>
+
+### Surface Normals Prediction Quick Start
+
+Load `prs-eth/marigold-normals-lcm-v0-1` checkpoint into `MarigoldNormalsPipeline` pipeline, put the image through the pipeline, and save the predictions: 
+
+```python
+import diffusers
+import torch
+
+pipe = diffusers.MarigoldNormalsPipeline.from_pretrained(
+    "prs-eth/marigold-normals-lcm-v0-1", variant="fp16", torch_dtype=torch.float16
+).to("cuda")
+
+image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+normals = pipe(image)
+
+vis = pipe.image_processor.visualize_normals(normals.prediction)
+vis[0].save("einstein_normals.png")
+```
+
+The visualization function for normals [`~pipelines.marigold.marigold_image_processing.MarigoldImageProcessor.visualize_normals`] maps the three-dimensional prediction with pixel values in the range `[-1, 1]` into an RGB image. 
+The visualization function supports flipping surface normals axes to make the visualization compatible with other choices of the frame of reference. 
+Conceptually, each pixel is painted according to the surface normal vector in the frame of reference, where `X` axis points right, `Y` axis points up, and `Z` axis points at the viewer.
+Below is the visualized prediction:
+
+<div class="flex gap-4" style="justify-content: center; width: 100%;">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_lcm_normals.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Predicted surface normals visualization
+	</figcaption>
+  </div>
+</div>
+
+In this example, the nose tip almost certainly has a point on the surface, in which the surface normal vector points straight at the viewer, meaning that its coordinates are `[0, 0, 1]`.
+This vector maps to the RGB `[128, 128, 255]`, which corresponds to the violet-blue color.
+Similarly, a surface normal on the cheek in the right part of the image has a large `X` component, which increases the red hue. 
+Points on the shoulders pointing up with a large `Y` promote green color. 
+
+### Speeding up inference
+
+The above quick start snippets are already optimized for speed: they load the LCM checkpoint, use the `fp16` variant of weights and computation, and perform just one denoising diffusion step. 
+The `pipe(image)` call completes in 280ms on RTX 3090 GPU.
+Internally, the input image is encoded with the Stable Diffusion VAE encoder, then the U-Net performs one denoising step, and finally, the prediction latent is decoded with the VAE decoder into pixel space.
+In this case, two out of three module calls are dedicated to converting between pixel and latent space of LDM.
+Because Marigold's latent space is compatible with the base Stable Diffusion, it is possible to speed up the pipeline call by more than 3x (85ms on RTX 3090) by using a [lightweight replacement of the SD VAE](../api/models/autoencoder_tiny):
+
+```diff
+  import diffusers
+  import torch
+  
+  pipe = diffusers.MarigoldDepthPipeline.from_pretrained(
+      "prs-eth/marigold-depth-lcm-v1-0", variant="fp16", torch_dtype=torch.float16
+  ).to("cuda")
+  
++ pipe.vae = diffusers.AutoencoderTiny.from_pretrained(
++ 	"madebyollin/taesd", torch_dtype=torch.float16
++ ).cuda()
+  
+  image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+  depth = pipe(image)
+```
+
+As suggested in [Optimizations](../optimization/torch2.0#torch.compile), adding `torch.compile` may squeeze extra performance depending on the target hardware:
+
+```diff
+  import diffusers
+  import torch
+  
+  pipe = diffusers.MarigoldDepthPipeline.from_pretrained(
+      "prs-eth/marigold-depth-lcm-v1-0", variant="fp16", torch_dtype=torch.float16
+  ).to("cuda")
+  
++ pipe.unet = torch.compile(pipe.unet, mode="reduce-overhead", fullgraph=True)
+  
+  image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+  depth = pipe(image)
+```
+
+## Qualitative Comparison with Depth Anything
+
+With the above speed optimizations, Marigold delivers predictions with more details and faster than [Depth Anything](https://huggingface.co/docs/transformers/main/en/model_doc/depth_anything) with the largest checkpoint [LiheYoung/depth-anything-large-hf](https://huggingface.co/LiheYoung/depth-anything-large-hf):
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_lcm_depth.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Marigold LCM fp16 with Tiny AutoEncoder
+	</figcaption>
+  </div>
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/einstein_depthanything_large.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Depth Anything Large
+	</figcaption>
+  </div>
+</div>
+
+## Maximizing Precision and Ensembling
+
+Marigold pipelines have a built-in ensembling mechanism combining multiple predictions from different random latents. 
+This is a brute-force way of improving the precision of predictions, capitalizing on the generative nature of diffusion.
+The ensembling path is activated automatically when the `ensemble_size` argument is set greater than `1`. 
+When aiming for maximum precision, it makes sense to adjust `num_inference_steps` simultaneously with `ensemble_size`. 
+The recommended values vary across checkpoints but primarily depend on the scheduler type.
+The effect of ensembling is particularly well-seen with surface normals:
+
+```python
+import diffusers
+
+model_path = "prs-eth/marigold-normals-v1-0"
+
+model_paper_kwargs = {
+	diffusers.schedulers.DDIMScheduler: {
+		"num_inference_steps": 10,
+		"ensemble_size": 10,
+	},
+	diffusers.schedulers.LCMScheduler: {
+		"num_inference_steps": 4,
+		"ensemble_size": 5,
+	},	
+}
+
+image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+
+pipe = diffusers.MarigoldNormalsPipeline.from_pretrained(model_path).to("cuda")
+pipe_kwargs = model_paper_kwargs[type(pipe.scheduler)]
+
+depth = pipe(image, **pipe_kwargs)
+
+vis = pipe.image_processor.visualize_normals(depth.prediction)
+vis[0].save("einstein_normals.png")
+```
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_lcm_normals.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Surface normals, no ensembling
+	</figcaption>
+  </div>
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_normals.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Surface normals, with ensembling
+	</figcaption>
+  </div>
+</div>
+
+As can be seen, all areas with fine-grained structurers, such as hair, got more conservative and on average more correct predictions.
+Such a result is more suitable for precision-sensitive downstream tasks, such as 3D reconstruction.
+
+## Quantitative Evaluation
+
+To evaluate Marigold quantitatively in standard leaderboards and benchmarks (such as NYU, KITTI, and other datasets), follow the evaluation protocol outlined in the paper: load the full precision fp32 model and use appropriate values for `num_inference_steps` and `ensemble_size`. 
+Optionally seed randomness to ensure reproducibility. Maximizing `batch_size` will deliver maximum device utilization.
+
+```python
+import diffusers
+import torch
+
+device = "cuda"
+seed = 2024
+model_path = "prs-eth/marigold-v1-0"
+
+model_paper_kwargs = {
+	diffusers.schedulers.DDIMScheduler: {
+		"num_inference_steps": 50,
+		"ensemble_size": 10,
+	},
+	diffusers.schedulers.LCMScheduler: {
+		"num_inference_steps": 4,
+		"ensemble_size": 10,
+	},	
+}
+
+image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+
+generator = torch.Generator(device=device).manual_seed(seed)
+pipe = diffusers.MarigoldDepthPipeline.from_pretrained(model_path).to(device)
+pipe_kwargs = model_paper_kwargs[type(pipe.scheduler)]
+
+depth = pipe(image, generator=generator, **pipe_kwargs)
+
+# evaluate metrics
+```
+
+## Using Predictive Uncertainty
+
+The ensembling mechanism built into Marigold pipelines combines multiple predictions obtained from different random latents. 
+As a side effect, it can be used to quantify epistemic (model) uncertainty; simply specify `ensemble_size` greater than 1 and set `output_uncertainty=True`.
+The resulting uncertainty will be available in the `uncertainty` field of the output.
+It can be visualized as follows:
+
+```python
+import diffusers
+import torch
+
+pipe = diffusers.MarigoldDepthPipeline.from_pretrained(
+    "prs-eth/marigold-depth-lcm-v1-0", variant="fp16", torch_dtype=torch.float16
+).to("cuda")
+
+image = diffusers.utils.load_image("https://marigoldmonodepth.github.io/images/einstein.jpg")
+depth = pipe(
+	image,
+	ensemble_size=10,  # any number greater than 1; higher values yield higher precision
+	output_uncertainty=True,
+)
+
+uncertainty = pipe.image_processor.visualize_uncertainty(depth.uncertainty)
+uncertainty[0].save("einstein_depth_uncertainty.png")
+```
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_depth_uncertainty.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Depth uncertainty
+	</figcaption>
+  </div>
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_einstein_normals_uncertainty.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+		Surface normals uncertainty
+	</figcaption>
+  </div>
+</div>
+
+The interpretation of uncertainty is easy: higher values (white) correspond to pixels, where the model struggles to make consistent predictions.
+Evidently, the depth model is the least confident around edges with discontinuity, where the object depth changes drastically.
+The surface normals model is the least confident in fine-grained structures, such as hair, and dark areas, such as the collar.
+
+## Frame-by-frame Video Processing with Temporal Consistency
+
+Due to Marigold's generative nature, each prediction is unique and defined by the random noise sampled for the latent initialization. 
+This becomes an obvious drawback compared to traditional end-to-end dense regression networks, as exemplified in the following videos:
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_obama.gif"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">Input video</figcaption>
+  </div>
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_obama_depth_independent.gif"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">Marigold Depth applied to input video frames independently</figcaption>
+  </div>
+</div>
+
+To address this issue, it is possible to pass `latents` argument to the pipelines, which defines the starting point of diffusion.
+Empirically, we found that a convex combination of the very same starting point noise latent and the latent corresponding to the previous frame prediction give sufficiently smooth results, as implemented in the snippet below:
+
+```python
+import imageio
+from PIL import Image
+from tqdm import tqdm
+import diffusers
+import torch
+
+device = "cuda"
+path_in = "obama.mp4"
+path_out = "obama_depth.gif"
+
+pipe = diffusers.MarigoldDepthPipeline.from_pretrained(
+    "prs-eth/marigold-depth-lcm-v1-0", variant="fp16", torch_dtype=torch.float16
+).to(device)
+pipe.vae = diffusers.AutoencoderTiny.from_pretrained(
+    "madebyollin/taesd", torch_dtype=torch.float16
+).to(device)
+pipe.set_progress_bar_config(disable=True)
+
+with imageio.get_reader(path_in) as reader:
+    size = reader.get_meta_data()['size']
+    last_frame_latent = None
+    latent_common = torch.randn(
+        (1, 4, 768 * size[1] // (8 * max(size)), 768 * size[0] // (8 * max(size)))
+    ).to(device=device, dtype=torch.float16)
+
+    out = []
+    for frame_id, frame in tqdm(enumerate(reader), desc="Processing Video"):
+        frame = Image.fromarray(frame)
+        latents = latent_common
+        if last_frame_latent is not None:
+            latents = 0.9 * latents + 0.1 * last_frame_latent
+
+        depth = pipe(
+			frame, match_input_resolution=False, latents=latents, output_latent=True,
+        )
+        last_frame_latent = depth.latent
+        out.append(pipe.image_processor.visualize_depth(depth.prediction)[0])
+
+    diffusers.utils.export_to_gif(out, path_out, fps=reader.get_meta_data()['fps'])
+```
+
+Here, the diffusion process starts from the given computed latent. 
+The pipeline sets `output_latent=True` to access `out.latent` and computes its contribution to the next frame's latent initialization.
+The result is much more stable now:
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_obama_depth_independent.gif"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">Marigold Depth applied to input video frames independently</figcaption>
+  </div>
+  <div style="flex: 1 1 50%; max-width: 50%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/marigold_obama_depth_consistent.gif"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">Marigold Depth with forced latents initialization</figcaption>
+  </div>
+</div>
+
+## Marigold for ControlNet
+
+A very common application for depth prediction with diffusion models comes in conjunction with ControlNet.
+Depth crispness plays a crucial role in obtaining high-quality results from ControlNet.
+As seen in comparisons with other methods above, Marigold excels at that task.
+The snippet below demonstrates how to load an image, compute depth, and pass it into ControlNet in a compatible format:
+
+```python
+import torch
+import diffusers
+
+device = "cuda"
+generator = torch.Generator(device=device).manual_seed(2024)
+image = diffusers.utils.load_image(
+    "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/controlnet_depth_source.png"
+)
+
+pipe = diffusers.MarigoldDepthPipeline.from_pretrained(
+    "prs-eth/marigold-depth-lcm-v1-0", torch_dtype=torch.float16, variant="fp16"
+).to("cuda")
+
+depth_image = pipe(image, generator=generator).prediction
+depth_image = pipe.image_processor.visualize_depth(depth_image, color_map="binary")
+depth_image[0].save("motorcycle_controlnet_depth.png")
+
+controlnet = diffusers.ControlNetModel.from_pretrained(
+    "diffusers/controlnet-depth-sdxl-1.0", torch_dtype=torch.float16, variant="fp16"
+).to("cuda")
+pipe = diffusers.StableDiffusionXLControlNetPipeline.from_pretrained(
+    "SG161222/RealVisXL_V4.0", torch_dtype=torch.float16, variant="fp16", controlnet=controlnet
+).to("cuda")
+pipe.scheduler = diffusers.DPMSolverMultistepScheduler.from_config(pipe.scheduler.config, use_karras_sigmas=True)
+
+controlnet_out = pipe(
+    prompt="high quality photo of a sports bike, city",
+    negative_prompt="",
+    guidance_scale=6.5,
+    num_inference_steps=25,
+    image=depth_image,
+    controlnet_conditioning_scale=0.7,
+    control_guidance_end=0.7,
+    generator=generator,
+).images
+controlnet_out[0].save("motorcycle_controlnet_out.png")
+```
+
+<div class="flex gap-4">
+  <div style="flex: 1 1 33%; max-width: 33%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/controlnet_depth_source.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+      Input image
+    </figcaption>
+  </div>
+  <div style="flex: 1 1 33%; max-width: 33%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/motorcycle_controlnet_depth.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+      Depth in the format compatible with ControlNet
+    </figcaption>
+  </div>
+  <div style="flex: 1 1 33%; max-width: 33%;">
+    <img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/marigold/motorcycle_controlnet_out.png"/>
+    <figcaption class="mt-1 text-center text-sm text-gray-500">
+      ControlNet generation, conditioned on depth and prompt: "high quality photo of a sports bike, city"
+    </figcaption>
+  </div>
+</div>
+
+Hopefully, you will find Marigold useful for solving your downstream tasks, be it a part of a more broad generative workflow, or a perception task, such as 3D reconstruction.

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

@@ -212,6 +212,62 @@ images = pipeline(prompt_ids, params, prng_seed, num_inference_steps, jit=True).
 images = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:])))
 ```
 
+## Custom Timestep Schedules
+
+With all our schedulers, you can choose one of the popular timestep schedules using configurations such as `timestep_spacing`, `interpolation_type`, and `use_karras_sigmas`. Some schedulers also provide the flexibility to use a custom timestep schedule. You can use any list of arbitrary timesteps, we will use the AYS timestep schedule here as example. It is a set of 10-step optimized timestep schedules released by researchers from Nvidia that can achieve significantly better quality compared to the preset timestep schedules. You can read more about their research [here](https://research.nvidia.com/labs/toronto-ai/AlignYourSteps/). 
+
+```python
+from diffusers.schedulers import AysSchedules
+sampling_schedule = AysSchedules["StableDiffusionXLTimesteps"]
+print(sampling_schedule)
+```
+```
+[999, 845, 730, 587, 443, 310, 193, 116, 53, 13]
+```
+
+You can then create a pipeline and pass this custom timestep schedule to it as `timesteps`.
+
+```python
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    "SG161222/RealVisXL_V4.0",
+    torch_dtype=torch.float16,
+    variant="fp16",
+).to("cuda")
+
+pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config, algorithm_type="sde-dpmsolver++")
+
+prompt = "A cinematic shot of a cute little rabbit wearing a jacket and doing a thumbs up"
+
+generator = torch.Generator(device="cpu").manual_seed(2487854446)
+
+image = pipe(
+    prompt=prompt,
+    negative_prompt="",
+    generator=generator,
+    timesteps=sampling_schedule,
+).images[0]
+```
+The generated image has better quality than the default linear timestep schedule for the same number of steps, and it is similar to the default timestep scheduler when running for 25 steps.
+
+<div class="flex gap-4">
+  <div>
+    <img class="rounded-xl" src="https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/ays.png"/>
+    <figcaption class="mt-2 text-center text-sm text-gray-500">AYS timestep schedule 10 steps</figcaption>
+  </div>
+  <div>
+    <img class="rounded-xl" src="https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/10.png"/>
+    <figcaption class="mt-2 text-center text-sm text-gray-500">Linearly-spaced timestep schedule 10 steps</figcaption>
+  </div>
+  <div>
+    <img class="rounded-xl" src="https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/25.png"/>
+    <figcaption class="mt-2 text-center text-sm text-gray-500">Linearly-spaced timestep schedule 25 steps</figcaption>
+  </div>
+</div>
+
+> [!TIP]
+> 🤗 Diffusers currently only supports `timesteps` and `sigmas` for a selected list of schedulers and pipelines, but feel free to open a [feature request](https://github.com/huggingface/diffusers/issues/new/choose) if you want to extend feature to a scheduler and pipeline that does not currently support it!
+
+
 ## Models
 
 Models are loaded from the [`ModelMixin.from_pretrained`] method, which downloads and caches the latest version of the model weights and configurations. If the latest files are available in the local cache, [`~ModelMixin.from_pretrained`] reuses files in the cache instead of re-downloading them.

docs/source/ja/installation.md

@@ -106,7 +106,7 @@ pip install -e ".[flax]"
 
 これらのコマンドは、リポジトリをクローンしたフォルダと Python のライブラリパスをリンクします。
 Python は通常のライブラリパスに加えて、クローンしたフォルダの中を探すようになります。
-例えば、Python パッケージが通常 `~/anaconda3/envs/main/lib/python3.8/site-packages/` にインストールされている場合、Python はクローンした `~/diffusers/` フォルダも同様に参照します。
+例えば、Python パッケージが通常 `~/anaconda3/envs/main/lib/python3.10/site-packages/` にインストールされている場合、Python はクローンした `~/diffusers/` フォルダも同様に参照します。
 
 <Tip warning={true}>
 

docs/source/ko/installation.md

@@ -105,7 +105,7 @@ pip install -e ".[flax]"
 
 이러한 명령어들은 저장소를 복제한 폴더와 Python 라이브러리 경로를 연결합니다.
 Python은 이제 일반 라이브러리 경로에 더하여 복제한 폴더 내부를 살펴봅니다.
-예를들어 Python 패키지가 `~/anaconda3/envs/main/lib/python3.8/site-packages/`에 설치되어 있는 경우 Python은 복제한 폴더인 `~/diffusers/`도 검색합니다.
+예를들어 Python 패키지가 `~/anaconda3/envs/main/lib/python3.10/site-packages/`에 설치되어 있는 경우 Python은 복제한 폴더인 `~/diffusers/`도 검색합니다.
 
 <Tip warning={true}>
 

docs/source/ko/optimization/fp16.md

@@ -339,7 +339,7 @@ from dataclasses import dataclass
 
 @dataclass
 class UNet2DConditionOutput:
-    sample: torch.FloatTensor
+    sample: torch.Tensor
 
 
 pipe = StableDiffusionPipeline.from_pretrained(

docs/source/ko/training/lora.md

@@ -49,15 +49,15 @@ huggingface-cli login
 
 ### 학습[[dreambooth-training]]
 
-[Pokémon BLIP 캡션](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) 데이터셋으로 [`stable-diffusion-v1-5`](https://huggingface.co/runwayml/stable-diffusion-v1-5)를 파인튜닝해 나만의 포켓몬을 생성해 보겠습니다.
+[Naruto BLIP 캡션](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) 데이터셋으로 [`stable-diffusion-v1-5`](https://huggingface.co/runwayml/stable-diffusion-v1-5)를 파인튜닝해 나만의 포켓몬을 생성해 보겠습니다.
 
 시작하려면 `MODEL_NAME` 및 `DATASET_NAME` 환경 변수가 설정되어 있는지 확인하십시오. `OUTPUT_DIR` 및 `HUB_MODEL_ID` 변수는 선택 사항이며 허브에서 모델을 저장할 위치를 지정합니다.
 
 ```bash
 export MODEL_NAME="runwayml/stable-diffusion-v1-5"
-export OUTPUT_DIR="/sddata/finetune/lora/pokemon"
-export HUB_MODEL_ID="pokemon-lora"
-export DATASET_NAME="lambdalabs/pokemon-blip-captions"
+export OUTPUT_DIR="/sddata/finetune/lora/naruto"
+export HUB_MODEL_ID="naruto-lora"
+export DATASET_NAME="lambdalabs/naruto-blip-captions"
 ```
 
 학습을 시작하기 전에 알아야 할 몇 가지 플래그가 있습니다.

docs/source/ko/training/text2image.md

@@ -73,12 +73,12 @@ xFormers는 Flax에 사용할 수 없습니다.
 
 <frameworkcontent>
 <pt>
-다음과 같이 [Pokémon BLIP 캡션](https://huggingface.co/datasets/lambdalabs/pokemon-blip-captions) 데이터셋에서 파인튜닝 실행을 위해 [PyTorch 학습 스크립트](https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image.py)를 실행합니다:
+다음과 같이 [Naruto BLIP 캡션](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) 데이터셋에서 파인튜닝 실행을 위해 [PyTorch 학습 스크립트](https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image.py)를 실행합니다:
 
 
 ```bash
 export MODEL_NAME="CompVis/stable-diffusion-v1-4"
-export dataset_name="lambdalabs/pokemon-blip-captions"
+export dataset_name="lambdalabs/naruto-blip-captions"
 
 accelerate launch train_text_to_image.py \
   --pretrained_model_name_or_path=$MODEL_NAME \
@@ -93,7 +93,7 @@ accelerate launch train_text_to_image.py \
   --learning_rate=1e-05 \
   --max_grad_norm=1 \
   --lr_scheduler="constant" --lr_warmup_steps=0 \
-  --output_dir="sd-pokemon-model" 
+  --output_dir="sd-naruto-model" 
 ```
 
 자체 데이터셋으로 파인튜닝하려면 🤗 [Datasets](https://huggingface.co/docs/datasets/index)에서 요구하는 형식에 따라 데이터셋을 준비하세요. [데이터셋을 허브에 업로드](https://huggingface.co/docs/datasets/image_dataset#upload-dataset-to-the-hub)하거나 [파일들이 있는 로컬 폴더를 준비](https ://huggingface.co/docs/datasets/image_dataset#imagefolder)할 수 있습니다.
@@ -136,7 +136,7 @@ pip install -U -r requirements_flax.txt
 
 ```bash
 export MODEL_NAME="runwayml/stable-diffusion-v1-5"
-export dataset_name="lambdalabs/pokemon-blip-captions"
+export dataset_name="lambdalabs/naruto-blip-captions"
 
 python train_text_to_image_flax.py \
   --pretrained_model_name_or_path=$MODEL_NAME \
@@ -146,7 +146,7 @@ python train_text_to_image_flax.py \
   --max_train_steps=15000 \
   --learning_rate=1e-05 \
   --max_grad_norm=1 \
-  --output_dir="sd-pokemon-model" 
+  --output_dir="sd-naruto-model" 
 ```
 
 자체 데이터셋으로 파인튜닝하려면 🤗 [Datasets](https://huggingface.co/docs/datasets/index)에서 요구하는 형식에 따라 데이터셋을 준비하세요. [데이터셋을 허브에 업로드](https://huggingface.co/docs/datasets/image_dataset#upload-dataset-to-the-hub)하거나 [파일들이 있는 로컬 폴더를 준비](https ://huggingface.co/docs/datasets/image_dataset#imagefolder)할 수 있습니다.
@@ -166,7 +166,7 @@ python train_text_to_image_flax.py \
   --max_train_steps=15000 \
   --learning_rate=1e-05 \
   --max_grad_norm=1 \
-  --output_dir="sd-pokemon-model"
+  --output_dir="sd-naruto-model"
 ```
 </jax>
 </frameworkcontent>
@@ -189,7 +189,7 @@ pipe = StableDiffusionPipeline.from_pretrained(model_path, torch_dtype=torch.flo
 pipe.to("cuda")
 
 image = pipe(prompt="yoda").images[0]
-image.save("yoda-pokemon.png")
+image.save("yoda-naruto.png")
 ```
 </pt>
 <jax>
@@ -203,7 +203,7 @@ from diffusers import FlaxStableDiffusionPipeline
 model_path = "path_to_saved_model"
 pipe, params = FlaxStableDiffusionPipeline.from_pretrained(model_path, dtype=jax.numpy.bfloat16)
 
-prompt = "yoda pokemon"
+prompt = "yoda naruto"
 prng_seed = jax.random.PRNGKey(0)
 num_inference_steps = 50
 
@@ -218,7 +218,7 @@ prompt_ids = shard(prompt_ids)
 
 images = pipeline(prompt_ids, params, prng_seed, num_inference_steps, jit=True).images
 images = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:])))
-image.save("yoda-pokemon.png")
+image.save("yoda-naruto.png")
 ```
 </jax>
 </frameworkcontent>

docs/source/ko/training/unconditional_training.md

@@ -103,13 +103,13 @@ accelerate launch train_unconditional.py \
 <div class="flex justify-center">
     <img src="https://user-images.githubusercontent.com/26864830/180248660-a0b143d0-b89a-42c5-8656-2ebf6ece7e52.png"/>
 </div>
-[Pokemon](https://huggingface.co/datasets/huggan/pokemon) 데이터셋을 사용할 경우:
+[Naruto](https://huggingface.co/datasets/lambdalabs/naruto-blip-captions) 데이터셋을 사용할 경우:
 
 ```bash
 accelerate launch train_unconditional.py \
-  --dataset_name="huggan/pokemon" \
+  --dataset_name="lambdalabs/naruto-blip-captions" \
   --resolution=64 \
-  --output_dir="ddpm-ema-pokemon-64" \
+  --output_dir="ddpm-ema-naruto-64" \
   --train_batch_size=16 \
   --num_epochs=100 \
   --gradient_accumulation_steps=1 \
@@ -129,9 +129,9 @@ accelerate launch train_unconditional.py \
 
 ```bash
 accelerate launch --mixed_precision="fp16" --multi_gpu train_unconditional.py \
-  --dataset_name="huggan/pokemon" \
+  --dataset_name="lambdalabs/naruto-blip-captions" \
   --resolution=64 --center_crop --random_flip \
-  --output_dir="ddpm-ema-pokemon-64" \
+  --output_dir="ddpm-ema-naruto-64" \
   --train_batch_size=16 \
   --num_epochs=100 \
   --gradient_accumulation_steps=1 \

docs/source/pt/installation.md

@@ -102,7 +102,7 @@ pip install -e ".[flax]"
 
 Esses comandos irá linkar a pasta que você clonou o repositório e os caminhos das suas bibliotecas Python.
 Python então irá procurar dentro da pasta que você clonou além dos caminhos normais das bibliotecas.
-Por exemplo, se o pacote python for tipicamente instalado no `~/anaconda3/envs/main/lib/python3.8/site-packages/`, o Python também irá procurar na pasta `~/diffusers/` que você clonou.
+Por exemplo, se o pacote python for tipicamente instalado no `~/anaconda3/envs/main/lib/python3.10/site-packages/`, o Python também irá procurar na pasta `~/diffusers/` que você clonou.
 
 <Tip warning={true}>
 

docs/source/zh/installation.md

@@ -107,7 +107,7 @@ pip install -e ".[flax]"
 
 这些命令将连接到你克隆的版本库和你的 Python 库路径。
 现在，不只是在通常的库路径，Python 还会在你克隆的文件夹内寻找包。
-例如，如果你的 Python 包通常安装在 `~/anaconda3/envs/main/lib/python3.8/Site-packages/`，Python 也会搜索你克隆到的文件夹。`~/diffusers/`。
+例如，如果你的 Python 包通常安装在 `~/anaconda3/envs/main/lib/python3.10/Site-packages/`，Python 也会搜索你克隆到的文件夹。`~/diffusers/`。
 
 <Tip warning={true}>
 

examples/advanced_diffusion_training/train_dreambooth_lora_sd15_advanced.py

@@ -71,7 +71,7 @@ from diffusers.utils.import_utils import is_xformers_available
 
 
 # Will error if the minimal version of diffusers is not installed. Remove at your own risks.
-check_min_version("0.28.0.dev0")
+check_min_version("0.28.0")
 
 logger = get_logger(__name__)
 
@@ -981,7 +981,7 @@ def collate_fn(examples, with_prior_preservation=False):
 
 
 class PromptDataset(Dataset):
-    "A simple dataset to prepare the prompts to generate class images on multiple GPUs."
+    """A simple dataset to prepare the prompts to generate class images on multiple GPUs."""
 
     def __init__(self, prompt, num_samples):
         self.prompt = prompt

examples/advanced_diffusion_training/train_dreambooth_lora_sdxl_advanced.py

@@ -78,7 +78,7 @@ from diffusers.utils.torch_utils import is_compiled_module
 
 
 # Will error if the minimal version of diffusers is not installed. Remove at your own risks.
-check_min_version("0.28.0.dev0")
+check_min_version("0.28.0")
 
 logger = get_logger(__name__)
 
@@ -1136,7 +1136,7 @@ def collate_fn(examples, with_prior_preservation=False):
 
 
 class PromptDataset(Dataset):
-    "A simple dataset to prepare the prompts to generate class images on multiple GPUs."
+    """A simple dataset to prepare the prompts to generate class images on multiple GPUs."""
 
     def __init__(self, prompt, num_samples):
         self.prompt = prompt

examples/community/README.md

@@ -68,6 +68,8 @@ Please also check out our [Community Scripts](https://github.com/huggingface/dif
 |   InstantID Pipeline                                                                                               | Stable Diffusion XL Pipeline that supports InstantID                                                                                                                                                                                                                                                                                                                                                 |  [InstantID Pipeline](#instantid-pipeline) | [![Hugging Face Space](https://img.shields.io/badge/🤗%20Hugging%20Face-Space-yellow)](https://huggingface.co/spaces/InstantX/InstantID) | [Haofan Wang](https://github.com/haofanwang) |
 |   UFOGen Scheduler                                                                                               | Scheduler for UFOGen Model (compatible with Stable Diffusion pipelines)                                                                                                                                                                                                                                                                                                                                                 |  [UFOGen Scheduler](#ufogen-scheduler) | - | [dg845](https://github.com/dg845) |
 | Stable Diffusion XL IPEX Pipeline | Accelerate Stable Diffusion XL inference pipeline with BF16/FP32 precision on Intel Xeon CPUs with [IPEX](https://github.com/intel/intel-extension-for-pytorch) | [Stable Diffusion XL on IPEX](#stable-diffusion-xl-on-ipex) | - | [Dan Li](https://github.com/ustcuna/) |
+| Stable Diffusion BoxDiff Pipeline | Training-free controlled generation with bounding boxes using [BoxDiff](https://github.com/showlab/BoxDiff) | [Stable Diffusion BoxDiff Pipeline](#stable-diffusion-boxdiff) | - | [Jingyang Zhang](https://github.com/zjysteven/) |
+|   FRESCO V2V Pipeline                                                                                                    | Implementation of [[CVPR 2024] FRESCO: Spatial-Temporal Correspondence for Zero-Shot Video Translation](https://arxiv.org/abs/2403.12962)                                                                                                                                                                                                                                                                                                                                                                                                                                      | [FRESCO V2V Pipeline](#fresco)      | - |              [Yifan Zhou](https://github.com/SingleZombie) |
 
 To load a custom pipeline you just need to pass the `custom_pipeline` argument to `DiffusionPipeline`, as one of the files in `diffusers/examples/community`. Feel free to send a PR with your own pipelines, we will merge them quickly.
 
@@ -238,12 +240,12 @@ pipeline_output = pipe(
     # denoising_steps=10,     # (optional) Number of denoising steps of each inference pass. Default: 10.
     # ensemble_size=10,       # (optional) Number of inference passes in the ensemble. Default: 10.
     # ------------------------------------------------
-    
+
     # ----- recommended setting for LCM version ------
     # denoising_steps=4,
     # ensemble_size=5,
     # -------------------------------------------------
-    
+
     # processing_res=768,     # (optional) Maximum resolution of processing. If set to 0: will not resize at all. Defaults to 768.
     # match_input_res=True,   # (optional) Resize depth prediction to match input resolution.
     # batch_size=0,           # (optional) Inference batch size, no bigger than `num_ensemble`. If set to 0, the script will automatically decide the proper batch size. Defaults to 0.
@@ -1030,7 +1032,7 @@ image = pipe().images[0]
 
 Make sure you have @crowsonkb's <https://github.com/crowsonkb/k-diffusion> installed:
 
-```
+```sh
 pip install k-diffusion
 ```
 
@@ -1676,6 +1678,68 @@ image = pipe(prompt, image=input_image, strength=0.75,).images[0]
 image.save('tensorrt_img2img_new_zealand_hills.png')
 ```
 
+### Stable Diffusion BoxDiff
+BoxDiff is a training-free method for controlled generation with bounding box coordinates. It shoud work with any Stable Diffusion model. Below shows an example with `stable-diffusion-2-1-base`.
+```py
+import torch
+from PIL import Image, ImageDraw
+from copy import deepcopy
+
+from examples.community.pipeline_stable_diffusion_boxdiff import StableDiffusionBoxDiffPipeline
+
+def draw_box_with_text(img, boxes, names):
+    colors = ["red", "olive", "blue", "green", "orange", "brown", "cyan", "purple"]
+    img_new = deepcopy(img)
+    draw = ImageDraw.Draw(img_new)
+
+    W, H = img.size
+    for bid, box in enumerate(boxes):
+        draw.rectangle([box[0] * W, box[1] * H, box[2] * W, box[3] * H], outline=colors[bid % len(colors)], width=4)
+        draw.text((box[0] * W, box[1] * H), names[bid], fill=colors[bid % len(colors)])
+    return img_new
+
+pipe = StableDiffusionBoxDiffPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-2-1-base",
+    torch_dtype=torch.float16,
+)
+pipe.to("cuda")
+
+# example 1
+prompt = "as the aurora lights up the sky, a herd of reindeer leisurely wanders on the grassy meadow, admiring the breathtaking view, a serene lake quietly reflects the magnificent display, and in the distance, a snow-capped mountain stands majestically, fantasy, 8k, highly detailed"
+phrases = [
+    "aurora",
+    "reindeer",
+    "meadow",
+    "lake",
+    "mountain"
+]
+boxes = [[1,3,512,202], [75,344,421,495], [1,327,508,507], [2,217,507,341], [1,135,509,242]]
+
+# example 2
+# prompt = "A rabbit wearing sunglasses looks very proud"
+# phrases = ["rabbit", "sunglasses"]
+# boxes = [[67,87,366,512], [66,130,364,262]]
+
+boxes = [[x / 512 for x in box] for box in boxes]
+
+images = pipe(
+    prompt,
+    boxdiff_phrases=phrases,
+    boxdiff_boxes=boxes,
+    boxdiff_kwargs={
+        "attention_res": 16,
+        "normalize_eot": True
+    },
+    num_inference_steps=50,
+    guidance_scale=7.5,
+    generator=torch.manual_seed(42),
+    safety_checker=None
+).images
+
+draw_box_with_text(images[0], boxes, phrases).save("output.png")
+```
+
+
 ### Stable Diffusion Reference
 
 This pipeline uses the Reference Control. Refer to the [sd-webui-controlnet discussion: Reference-only Control](https://github.com/Mikubill/sd-webui-controlnet/discussions/1236)[sd-webui-controlnet discussion: Reference-adain Control](https://github.com/Mikubill/sd-webui-controlnet/discussions/1280).
@@ -1790,13 +1854,13 @@ To use this pipeline, you need to:
 
 You can simply use pip to install IPEX with the latest version.
 
-```python
+```sh
 python -m pip install intel_extension_for_pytorch
 ```
 
 **Note:** To install a specific version, run with the following command:
 
-```
+```sh
 python -m pip install intel_extension_for_pytorch==<version_name> -f https://developer.intel.com/ipex-whl-stable-cpu
 ```
 
@@ -1894,13 +1958,13 @@ To use this pipeline, you need to:
 
 You can simply use pip to install IPEX with the latest version.
 
-```python
+```sh
 python -m pip install intel_extension_for_pytorch
 ```
 
 **Note:** To install a specific version, run with the following command:
 
-```
+```sh
 python -m pip install intel_extension_for_pytorch==<version_name> -f https://developer.intel.com/ipex-whl-stable-cpu
 ```
 
@@ -2946,8 +3010,8 @@ This code implements a pipeline for the Stable Diffusion model, enabling the div
 
 ### Sample Code
 
-```
-from from examples.community.regional_prompting_stable_diffusion import RegionalPromptingStableDiffusionPipeline
+```py
+from examples.community.regional_prompting_stable_diffusion import RegionalPromptingStableDiffusionPipeline
 pipe = RegionalPromptingStableDiffusionPipeline.from_single_file(model_path, vae=vae)
 
 rp_args = {
@@ -3972,6 +4036,93 @@ onestep_image = pipe(prompt, num_inference_steps=1).images[0]
 multistep_image = pipe(prompt, num_inference_steps=4).images[0]
 ```
 
+### FRESCO
+
+This is the Diffusers implementation of zero-shot video-to-video translation pipeline [FRESCO](https://github.com/williamyang1991/FRESCO) (without Ebsynth postprocessing and background smooth). To run the code, please install gmflow. Then modify the path in `gmflow_dir`. After that, you can run the pipeline with:
+
+```py
+from PIL import Image
+import cv2
+import torch
+import numpy as np
+
+from diffusers import ControlNetModel,DDIMScheduler, DiffusionPipeline
+import sys
+gmflow_dir = "/path/to/gmflow"
+sys.path.insert(0, gmflow_dir)
+
+def video_to_frame(video_path: str, interval: int):
+    vidcap = cv2.VideoCapture(video_path)
+    success = True
+
+    count = 0
+    res = []
+    while success:
+        count += 1
+        success, image = vidcap.read()
+        if count % interval != 1:
+            continue
+        if image is not None:
+            image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
+            res.append(image)
+            if len(res) >= 8:
+                break
+
+    vidcap.release()
+    return res
+
+
+input_video_path = 'https://github.com/williamyang1991/FRESCO/raw/main/data/car-turn.mp4'
+output_video_path = 'car.gif'
+
+# You can use any fintuned SD here
+model_path = 'SG161222/Realistic_Vision_V2.0'
+
+prompt = 'a red car turns in the winter'
+a_prompt = ', RAW photo, subject, (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3, '
+n_prompt = '(deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime, mutated hands and fingers:1.4), (deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, disconnected limbs, mutation, mutated, ugly, disgusting, amputation'
+
+input_interval = 5
+frames = video_to_frame(
+    input_video_path, input_interval)
+
+control_frames = []
+# get canny image
+for frame in frames:
+    image = cv2.Canny(frame, 50, 100)
+    np_image = np.array(image)
+    np_image = np_image[:, :, None]
+    np_image = np.concatenate([np_image, np_image, np_image], axis=2)
+    canny_image = Image.fromarray(np_image)
+    control_frames.append(canny_image)
+
+# You can use any ControlNet here
+controlnet = ControlNetModel.from_pretrained(
+    "lllyasviel/sd-controlnet-canny").to('cuda')
+
+pipe = DiffusionPipeline.from_pretrained(
+    model_path, controlnet=controlnet, custom_pipeline='fresco_v2v').to('cuda')
+pipe.scheduler = DDIMScheduler.from_config(pipe.scheduler.config)
+
+generator = torch.manual_seed(0)
+frames = [Image.fromarray(frame) for frame in frames]
+
+output_frames = pipe(
+    prompt + a_prompt,
+    frames,
+    control_frames,
+    num_inference_steps=20,
+    strength=0.75,
+    controlnet_conditioning_scale=0.7,
+    generator=generator,
+    negative_prompt=n_prompt
+).images
+
+output_frames[0].save(output_video_path, save_all=True,
+                 append_images=output_frames[1:], duration=100, loop=0)
+
+```
+
 # Perturbed-Attention Guidance
 
 [Project](https://ku-cvlab.github.io/Perturbed-Attention-Guidance/) / [arXiv](https://arxiv.org/abs/2403.17377) / [GitHub](https://github.com/KU-CVLAB/Perturbed-Attention-Guidance)
@@ -3980,7 +4131,7 @@ This implementation is based on [Diffusers](https://huggingface.co/docs/diffuser
 
 ## Example Usage
 
-```
+```py
 import os
 import torch
 

examples/community/bit_diffusion.py

@@ -44,9 +44,9 @@ def bits_to_decimal(x, bits=BITS):
 # modified scheduler step functions for clamping the predicted x_0 between -bit_scale and +bit_scale
 def ddim_bit_scheduler_step(
     self,
-    model_output: torch.FloatTensor,
+    model_output: torch.Tensor,
     timestep: int,
-    sample: torch.FloatTensor,
+    sample: torch.Tensor,
     eta: float = 0.0,
     use_clipped_model_output: bool = True,
     generator=None,
@@ -56,9 +56,9 @@ def ddim_bit_scheduler_step(
     Predict the sample at the previous timestep by reversing the SDE. Core function to propagate the diffusion
     process from the learned model outputs (most often the predicted noise).
     Args:
-        model_output (`torch.FloatTensor`): direct output from learned diffusion model.
+        model_output (`torch.Tensor`): direct output from learned diffusion model.
         timestep (`int`): current discrete timestep in the diffusion chain.
-        sample (`torch.FloatTensor`):
+        sample (`torch.Tensor`):
             current instance of sample being created by diffusion process.
         eta (`float`): weight of noise for added noise in diffusion step.
         use_clipped_model_output (`bool`): TODO
@@ -134,9 +134,9 @@ def ddim_bit_scheduler_step(
 
 def ddpm_bit_scheduler_step(
     self,
-    model_output: torch.FloatTensor,
+    model_output: torch.Tensor,
     timestep: int,
-    sample: torch.FloatTensor,
+    sample: torch.Tensor,
     prediction_type="epsilon",
     generator=None,
     return_dict: bool = True,
@@ -145,9 +145,9 @@ def ddpm_bit_scheduler_step(
     Predict the sample at the previous timestep by reversing the SDE. Core function to propagate the diffusion
     process from the learned model outputs (most often the predicted noise).
     Args:
-        model_output (`torch.FloatTensor`): direct output from learned diffusion model.
+        model_output (`torch.Tensor`): direct output from learned diffusion model.
         timestep (`int`): current discrete timestep in the diffusion chain.
-        sample (`torch.FloatTensor`):
+        sample (`torch.Tensor`):
             current instance of sample being created by diffusion process.
         prediction_type (`str`, default `epsilon`):
             indicates whether the model predicts the noise (epsilon), or the samples (`sample`).

examples/community/checkpoint_merger.py

@@ -138,7 +138,6 @@ class CheckpointMergerPipeline(DiffusionPipeline):
             comparison_result &= self._compare_model_configs(config_dicts[idx - 1], config_dicts[idx])
             if not force and comparison_result is False:
                 raise ValueError("Incompatible checkpoints. Please check model_index.json for the models.")
-                print(config_dicts[0], config_dicts[1])
         print("Compatible model_index.json files found")
         # Step 2: Basic Validation has succeeded. Let's download the models and save them into our local files.
         cached_folders = []

examples/community/clip_guided_images_mixing_stable_diffusion.py

@@ -233,8 +233,8 @@ class CLIPGuidedImagesMixingStableDiffusion(DiffusionPipeline, StableDiffusionMi
     @torch.no_grad()
     def __call__(
         self,
-        style_image: Union[torch.FloatTensor, PIL.Image.Image],
-        content_image: Union[torch.FloatTensor, PIL.Image.Image],
+        style_image: Union[torch.Tensor, PIL.Image.Image],
+        content_image: Union[torch.Tensor, PIL.Image.Image],
         style_prompt: Optional[str] = None,
         content_prompt: Optional[str] = None,
         height: Optional[int] = 512,

examples/community/clip_guided_stable_diffusion.py

@@ -180,7 +180,7 @@ class CLIPGuidedStableDiffusion(DiffusionPipeline, StableDiffusionMixin):
         num_cutouts: Optional[int] = 4,
         use_cutouts: Optional[bool] = True,
         generator: Optional[torch.Generator] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
     ):

examples/community/clip_guided_stable_diffusion_img2img.py

@@ -306,7 +306,7 @@ class CLIPGuidedStableDiffusion(DiffusionPipeline, StableDiffusionMixin):
         prompt: Union[str, List[str]],
         height: Optional[int] = 512,
         width: Optional[int] = 512,
-        image: Union[torch.FloatTensor, PIL.Image.Image] = None,
+        image: Union[torch.Tensor, PIL.Image.Image] = None,
         strength: float = 0.8,
         num_inference_steps: Optional[int] = 50,
         guidance_scale: Optional[float] = 7.5,
@@ -317,7 +317,7 @@ class CLIPGuidedStableDiffusion(DiffusionPipeline, StableDiffusionMixin):
         num_cutouts: Optional[int] = 4,
         use_cutouts: Optional[bool] = True,
         generator: Optional[torch.Generator] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
     ):

examples/community/composable_stable_diffusion.py

@@ -354,10 +354,10 @@ class ComposableStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin)
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[torch.Generator] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         weights: Optional[str] = "",
     ):
@@ -391,7 +391,7 @@ class ComposableStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin)
             generator (`torch.Generator`, *optional*):
                 A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
                 deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
@@ -403,7 +403,7 @@ class ComposableStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin)
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/ddim_noise_comparative_analysis.py

@@ -103,7 +103,7 @@ class DDIMNoiseComparativeAnalysisPipeline(DiffusionPipeline):
     @torch.no_grad()
     def __call__(
         self,
-        image: Union[torch.FloatTensor, PIL.Image.Image] = None,
+        image: Union[torch.Tensor, PIL.Image.Image] = None,
         strength: float = 0.8,
         batch_size: int = 1,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
@@ -115,7 +115,7 @@ class DDIMNoiseComparativeAnalysisPipeline(DiffusionPipeline):
     ) -> Union[ImagePipelineOutput, Tuple]:
         r"""
         Args:
-            image (`torch.FloatTensor` or `PIL.Image.Image`):
+            image (`torch.Tensor` or `PIL.Image.Image`):
                 `Image`, or tensor representing an image batch, that will be used as the starting point for the
                 process.
             strength (`float`, *optional*, defaults to 0.8):

examples/community/gluegen.py

@@ -205,7 +205,7 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
         safety_checker: StableDiffusionSafetyChecker,
         feature_extractor: CLIPImageProcessor,
         language_adapter: TranslatorNoLN = None,
-        tensor_norm: torch.FloatTensor = None,
+        tensor_norm: torch.Tensor = None,
         requires_safety_checker: bool = True,
     ):
         super().__init__()
@@ -231,7 +231,7 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
         num_token: int,
         dim: int,
         dim_out: int,
-        tensor_norm: torch.FloatTensor,
+        tensor_norm: torch.Tensor,
         mult: int = 2,
         depth: int = 5,
     ):
@@ -242,7 +242,7 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
         )
         self.language_adapter.load_state_dict(torch.load(model_path))
 
-    def _adapt_language(self, prompt_embeds: torch.FloatTensor):
+    def _adapt_language(self, prompt_embeds: torch.Tensor):
         prompt_embeds = prompt_embeds / 3
         prompt_embeds = self.language_adapter(prompt_embeds) * (self.tensor_norm / 2)
         return prompt_embeds
@@ -254,8 +254,8 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -275,10 +275,10 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -535,7 +535,7 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
                 data type of the generated embeddings
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -594,9 +594,9 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -635,14 +635,14 @@ class GlueGenStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin, Lo
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.

examples/community/hd_painter.py

@@ -28,10 +28,10 @@ class RASGAttnProcessor:
     def __call__(
         self,
         attn: Attention,
-        hidden_states: torch.FloatTensor,
-        encoder_hidden_states: Optional[torch.FloatTensor] = None,
-        attention_mask: Optional[torch.FloatTensor] = None,
-        temb: Optional[torch.FloatTensor] = None,
+        hidden_states: torch.Tensor,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        temb: Optional[torch.Tensor] = None,
         scale: float = 1.0,
     ) -> torch.Tensor:
         # Same as the default AttnProcessor up untill the part where similarity matrix gets saved
@@ -111,10 +111,10 @@ class PAIntAAttnProcessor:
     def __call__(
         self,
         attn: Attention,
-        hidden_states: torch.FloatTensor,
-        encoder_hidden_states: Optional[torch.FloatTensor] = None,
-        attention_mask: Optional[torch.FloatTensor] = None,
-        temb: Optional[torch.FloatTensor] = None,
+        hidden_states: torch.Tensor,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        temb: Optional[torch.Tensor] = None,
         scale: float = 1.0,
     ) -> torch.Tensor:
         # Automatically recognize the resolution of the current attention layer and resize the masks accordingly
@@ -454,7 +454,7 @@ class StableDiffusionHDPainterPipeline(StableDiffusionInpaintPipeline):
         prompt: Union[str, List[str]] = None,
         image: PipelineImageInput = None,
         mask_image: PipelineImageInput = None,
-        masked_image_latents: torch.FloatTensor = None,
+        masked_image_latents: torch.Tensor = None,
         height: Optional[int] = None,
         width: Optional[int] = None,
         padding_mask_crop: Optional[int] = None,
@@ -467,9 +467,9 @@ class StableDiffusionHDPainterPipeline(StableDiffusionInpaintPipeline):
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.01,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,

examples/community/iadb.py

@@ -17,21 +17,21 @@ class IADBScheduler(SchedulerMixin, ConfigMixin):
 
     def step(
         self,
-        model_output: torch.FloatTensor,
+        model_output: torch.Tensor,
         timestep: int,
-        x_alpha: torch.FloatTensor,
-    ) -> torch.FloatTensor:
+        x_alpha: torch.Tensor,
+    ) -> torch.Tensor:
         """
         Predict the sample at the previous timestep by reversing the ODE. Core function to propagate the diffusion
         process from the learned model outputs (most often the predicted noise).
 
         Args:
-            model_output (`torch.FloatTensor`): direct output from learned diffusion model. It is the direction from x0 to x1.
+            model_output (`torch.Tensor`): direct output from learned diffusion model. It is the direction from x0 to x1.
             timestep (`float`): current timestep in the diffusion chain.
-            x_alpha (`torch.FloatTensor`): x_alpha sample for the current timestep
+            x_alpha (`torch.Tensor`): x_alpha sample for the current timestep
 
         Returns:
-            `torch.FloatTensor`: the sample at the previous timestep
+            `torch.Tensor`: the sample at the previous timestep
 
         """
         if self.num_inference_steps is None:
@@ -53,10 +53,10 @@ class IADBScheduler(SchedulerMixin, ConfigMixin):
 
     def add_noise(
         self,
-        original_samples: torch.FloatTensor,
-        noise: torch.FloatTensor,
-        alpha: torch.FloatTensor,
-    ) -> torch.FloatTensor:
+        original_samples: torch.Tensor,
+        noise: torch.Tensor,
+        alpha: torch.Tensor,
+    ) -> torch.Tensor:
         return original_samples * alpha + noise * (1 - alpha)
 
     def __len__(self):

examples/community/imagic_stable_diffusion.py

@@ -110,7 +110,7 @@ class ImagicStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin):
     def train(
         self,
         prompt: Union[str, List[str]],
-        image: Union[torch.FloatTensor, PIL.Image.Image],
+        image: Union[torch.Tensor, PIL.Image.Image],
         height: Optional[int] = 512,
         width: Optional[int] = 512,
         generator: Optional[torch.Generator] = None,
@@ -144,7 +144,7 @@ class ImagicStableDiffusionPipeline(DiffusionPipeline, StableDiffusionMixin):
             generator (`torch.Generator`, *optional*):
                 A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
                 deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.

examples/community/img2img_inpainting.py

@@ -133,9 +133,9 @@ class ImageToImageInpaintingPipeline(DiffusionPipeline):
     def __call__(
         self,
         prompt: Union[str, List[str]],
-        image: Union[torch.FloatTensor, PIL.Image.Image],
-        inner_image: Union[torch.FloatTensor, PIL.Image.Image],
-        mask_image: Union[torch.FloatTensor, PIL.Image.Image],
+        image: Union[torch.Tensor, PIL.Image.Image],
+        inner_image: Union[torch.Tensor, PIL.Image.Image],
+        mask_image: Union[torch.Tensor, PIL.Image.Image],
         height: int = 512,
         width: int = 512,
         num_inference_steps: int = 50,
@@ -144,10 +144,10 @@ class ImageToImageInpaintingPipeline(DiffusionPipeline):
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[torch.Generator] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         **kwargs,
     ):
@@ -194,7 +194,7 @@ class ImageToImageInpaintingPipeline(DiffusionPipeline):
             generator (`torch.Generator`, *optional*):
                 A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
                 deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
@@ -206,7 +206,7 @@ class ImageToImageInpaintingPipeline(DiffusionPipeline):
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/instaflow_one_step.py

@@ -189,8 +189,8 @@ class InstaFlowPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
     ):
         deprecation_message = "`_encode_prompt()` is deprecated and it will be removed in a future version. Use `encode_prompt()` instead. Also, be aware that the output format changed from a concatenated tensor to a tuple."
@@ -219,8 +219,8 @@ class InstaFlowPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
     ):
         r"""
@@ -239,10 +239,10 @@ class InstaFlowPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -501,12 +501,12 @@ class InstaFlowPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         guidance_rescale: float = 0.0,
@@ -538,14 +538,14 @@ class InstaFlowPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):
@@ -555,7 +555,7 @@ class InstaFlowPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that calls every `callback_steps` steps during inference. The function is called with the
-                following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function is called. If not specified, the callback is called at
                 every step.

examples/community/interpolate_stable_diffusion.py

@@ -132,12 +132,12 @@ class StableDiffusionWalkPipeline(DiffusionPipeline, StableDiffusionMixin):
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[torch.Generator] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
-        text_embeddings: Optional[torch.FloatTensor] = None,
+        text_embeddings: Optional[torch.Tensor] = None,
         **kwargs,
     ):
         r"""
@@ -170,7 +170,7 @@ class StableDiffusionWalkPipeline(DiffusionPipeline, StableDiffusionMixin):
             generator (`torch.Generator`, *optional*):
                 A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
                 deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
@@ -182,11 +182,11 @@ class StableDiffusionWalkPipeline(DiffusionPipeline, StableDiffusionMixin):
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.
-            text_embeddings (`torch.FloatTensor`, *optional*, defaults to `None`):
+            text_embeddings (`torch.Tensor`, *optional*, defaults to `None`):
                 Pre-generated text embeddings to be used as inputs for image generation. Can be used in place of
                 `prompt` to avoid re-computing the embeddings. If not provided, the embeddings will be generated from
                 the supplied `prompt`.

examples/community/ip_adapter_face_id.py

@@ -62,7 +62,7 @@ class IPAdapterFullImageProjection(nn.Module):
         self.ff = FeedForward(image_embed_dim, cross_attention_dim * num_tokens, mult=mult, activation_fn="gelu")
         self.norm = nn.LayerNorm(cross_attention_dim)
 
-    def forward(self, image_embeds: torch.FloatTensor):
+    def forward(self, image_embeds: torch.Tensor):
         x = self.ff(image_embeds)
         x = x.reshape(-1, self.num_tokens, self.cross_attention_dim)
         return self.norm(x)
@@ -452,8 +452,8 @@ class IPAdapterFaceIDStableDiffusionPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         **kwargs,
     ):
@@ -484,8 +484,8 @@ class IPAdapterFaceIDStableDiffusionPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -505,10 +505,10 @@ class IPAdapterFaceIDStableDiffusionPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -788,7 +788,7 @@ class IPAdapterFaceIDStableDiffusionPipeline(
                 data type of the generated embeddings
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -847,10 +847,10 @@ class IPAdapterFaceIDStableDiffusionPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        image_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        image_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -891,17 +891,17 @@ class IPAdapterFaceIDStableDiffusionPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
-            image_embeds (`torch.FloatTensor`, *optional*):
+            image_embeds (`torch.Tensor`, *optional*):
                 Pre-generated image embeddings.
             output_type (`str`, *optional*, defaults to `"pil"`):
                 The output format of the generated image. Choose between `PIL.Image` or `np.array`.

examples/community/latent_consistency_img2img.py

@@ -88,7 +88,7 @@ class LatentConsistencyModelImg2ImgPipeline(DiffusionPipeline):
                 torch device
             num_images_per_prompt (`int`):
                 number of images that should be generated per prompt
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
         """
@@ -240,14 +240,6 @@ class LatentConsistencyModelImg2ImgPipeline(DiffusionPipeline):
 
         return latents
 
-        if latents is None:
-            latents = torch.randn(shape, dtype=dtype).to(device)
-        else:
-            latents = latents.to(device)
-        # scale the initial noise by the standard deviation required by the scheduler
-        latents = latents * self.scheduler.init_noise_sigma
-        return latents
-
     def get_w_embedding(self, w, embedding_dim=512, dtype=torch.float32):
         """
         see https://github.com/google-research/vdm/blob/dc27b98a554f65cdc654b800da5aa1846545d41b/model_vdm.py#L298
@@ -290,10 +282,10 @@ class LatentConsistencyModelImg2ImgPipeline(DiffusionPipeline):
         width: Optional[int] = 768,
         guidance_scale: float = 7.5,
         num_images_per_prompt: Optional[int] = 1,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         num_inference_steps: int = 4,
         lcm_origin_steps: int = 50,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -403,16 +395,16 @@ class LCMSchedulerOutput(BaseOutput):
     """
     Output class for the scheduler's `step` function output.
     Args:
-        prev_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
+        prev_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
             Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the
             denoising loop.
-        pred_original_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
+        pred_original_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
             The predicted denoised sample `(x_{0})` based on the model output from the current timestep.
             `pred_original_sample` can be used to preview progress or for guidance.
     """
 
-    prev_sample: torch.FloatTensor
-    denoised: Optional[torch.FloatTensor] = None
+    prev_sample: torch.Tensor
+    denoised: Optional[torch.Tensor] = None
 
 
 # Copied from diffusers.schedulers.scheduling_ddpm.betas_for_alpha_bar
@@ -460,10 +452,10 @@ def rescale_zero_terminal_snr(betas):
     """
     Rescales betas to have zero terminal SNR Based on https://arxiv.org/pdf/2305.08891.pdf (Algorithm 1)
     Args:
-        betas (`torch.FloatTensor`):
+        betas (`torch.Tensor`):
             the betas that the scheduler is being initialized with.
     Returns:
-        `torch.FloatTensor`: rescaled betas with zero terminal SNR
+        `torch.Tensor`: rescaled betas with zero terminal SNR
     """
     # Convert betas to alphas_bar_sqrt
     alphas = 1.0 - betas
@@ -573,7 +565,7 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
             # Glide cosine schedule
             self.betas = betas_for_alpha_bar(num_train_timesteps)
         else:
-            raise NotImplementedError(f"{beta_schedule} does is not implemented for {self.__class__}")
+            raise NotImplementedError(f"{beta_schedule} is not implemented for {self.__class__}")
 
         # Rescale for zero SNR
         if rescale_betas_zero_snr:
@@ -595,17 +587,17 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
         self.num_inference_steps = None
         self.timesteps = torch.from_numpy(np.arange(0, num_train_timesteps)[::-1].copy().astype(np.int64))
 
-    def scale_model_input(self, sample: torch.FloatTensor, timestep: Optional[int] = None) -> torch.FloatTensor:
+    def scale_model_input(self, sample: torch.Tensor, timestep: Optional[int] = None) -> torch.Tensor:
         """
         Ensures interchangeability with schedulers that need to scale the denoising model input depending on the
         current timestep.
         Args:
-            sample (`torch.FloatTensor`):
+            sample (`torch.Tensor`):
                 The input sample.
             timestep (`int`, *optional*):
                 The current timestep in the diffusion chain.
         Returns:
-            `torch.FloatTensor`:
+            `torch.Tensor`:
                 A scaled input sample.
         """
         return sample
@@ -621,7 +613,7 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
         return variance
 
     # Copied from diffusers.schedulers.scheduling_ddpm.DDPMScheduler._threshold_sample
-    def _threshold_sample(self, sample: torch.FloatTensor) -> torch.FloatTensor:
+    def _threshold_sample(self, sample: torch.Tensor) -> torch.Tensor:
         """
         "Dynamic thresholding: At each sampling step we set s to a certain percentile absolute pixel value in xt0 (the
         prediction of x_0 at timestep t), and if s > 1, then we threshold xt0 to the range [-s, s] and then divide by
@@ -693,25 +685,25 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
 
     def step(
         self,
-        model_output: torch.FloatTensor,
+        model_output: torch.Tensor,
         timeindex: int,
         timestep: int,
-        sample: torch.FloatTensor,
+        sample: torch.Tensor,
         eta: float = 0.0,
         use_clipped_model_output: bool = False,
         generator=None,
-        variance_noise: Optional[torch.FloatTensor] = None,
+        variance_noise: Optional[torch.Tensor] = None,
         return_dict: bool = True,
     ) -> Union[LCMSchedulerOutput, Tuple]:
         """
         Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
         process from the learned model outputs (most often the predicted noise).
         Args:
-            model_output (`torch.FloatTensor`):
+            model_output (`torch.Tensor`):
                 The direct output from learned diffusion model.
             timestep (`float`):
                 The current discrete timestep in the diffusion chain.
-            sample (`torch.FloatTensor`):
+            sample (`torch.Tensor`):
                 A current instance of a sample created by the diffusion process.
             eta (`float`):
                 The weight of noise for added noise in diffusion step.
@@ -722,7 +714,7 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
                 `use_clipped_model_output` has no effect.
             generator (`torch.Generator`, *optional*):
                 A random number generator.
-            variance_noise (`torch.FloatTensor`):
+            variance_noise (`torch.Tensor`):
                 Alternative to generating noise with `generator` by directly providing the noise for the variance
                 itself. Useful for methods such as [`CycleDiffusion`].
             return_dict (`bool`, *optional*, defaults to `True`):
@@ -785,10 +777,10 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
     # Copied from diffusers.schedulers.scheduling_ddpm.DDPMScheduler.add_noise
     def add_noise(
         self,
-        original_samples: torch.FloatTensor,
-        noise: torch.FloatTensor,
+        original_samples: torch.Tensor,
+        noise: torch.Tensor,
         timesteps: torch.IntTensor,
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         # Make sure alphas_cumprod and timestep have same device and dtype as original_samples
         alphas_cumprod = self.alphas_cumprod.to(device=original_samples.device, dtype=original_samples.dtype)
         timesteps = timesteps.to(original_samples.device)
@@ -807,9 +799,7 @@ class LCMSchedulerWithTimestamp(SchedulerMixin, ConfigMixin):
         return noisy_samples
 
     # Copied from diffusers.schedulers.scheduling_ddpm.DDPMScheduler.get_velocity
-    def get_velocity(
-        self, sample: torch.FloatTensor, noise: torch.FloatTensor, timesteps: torch.IntTensor
-    ) -> torch.FloatTensor:
+    def get_velocity(self, sample: torch.Tensor, noise: torch.Tensor, timesteps: torch.IntTensor) -> torch.Tensor:
         # Make sure alphas_cumprod and timestep have same device and dtype as sample
         alphas_cumprod = self.alphas_cumprod.to(device=sample.device, dtype=sample.dtype)
         timesteps = timesteps.to(sample.device)

examples/community/latent_consistency_interpolate.py

@@ -281,8 +281,8 @@ class LatentConsistencyModelWalkPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -302,10 +302,10 @@ class LatentConsistencyModelWalkPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -506,7 +506,7 @@ class LatentConsistencyModelWalkPipeline(
                 data type of the generated embeddings
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -546,7 +546,7 @@ class LatentConsistencyModelWalkPipeline(
         height: int,
         width: int,
         callback_steps: int,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
         callback_on_step_end_tensor_inputs=None,
     ):
         if height % 8 != 0 or width % 8 != 0:
@@ -580,11 +580,11 @@ class LatentConsistencyModelWalkPipeline(
     @torch.no_grad()
     def interpolate_embedding(
         self,
-        start_embedding: torch.FloatTensor,
-        end_embedding: torch.FloatTensor,
+        start_embedding: torch.Tensor,
+        end_embedding: torch.Tensor,
         num_interpolation_steps: Union[int, List[int]],
         interpolation_type: str,
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         if interpolation_type == "lerp":
             interpolation_fn = lerp
         elif interpolation_type == "slerp":
@@ -611,11 +611,11 @@ class LatentConsistencyModelWalkPipeline(
     @torch.no_grad()
     def interpolate_latent(
         self,
-        start_latent: torch.FloatTensor,
-        end_latent: torch.FloatTensor,
+        start_latent: torch.Tensor,
+        end_latent: torch.Tensor,
         num_interpolation_steps: Union[int, List[int]],
         interpolation_type: str,
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         if interpolation_type == "lerp":
             interpolation_fn = lerp
         elif interpolation_type == "slerp":
@@ -663,8 +663,8 @@ class LatentConsistencyModelWalkPipeline(
         guidance_scale: float = 8.5,
         num_images_per_prompt: Optional[int] = 1,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -705,11 +705,11 @@ class LatentConsistencyModelWalkPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):

examples/community/latent_consistency_txt2img.py

@@ -86,7 +86,7 @@ class LatentConsistencyModelPipeline(DiffusionPipeline):
                 torch device
             num_images_per_prompt (`int`):
                 number of images that should be generated per prompt
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
         """
@@ -208,10 +208,10 @@ class LatentConsistencyModelPipeline(DiffusionPipeline):
         width: Optional[int] = 768,
         guidance_scale: float = 7.5,
         num_images_per_prompt: Optional[int] = 1,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         num_inference_steps: int = 4,
         lcm_origin_steps: int = 50,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -310,16 +310,16 @@ class LCMSchedulerOutput(BaseOutput):
     """
     Output class for the scheduler's `step` function output.
     Args:
-        prev_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
+        prev_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
             Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the
             denoising loop.
-        pred_original_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
+        pred_original_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
             The predicted denoised sample `(x_{0})` based on the model output from the current timestep.
             `pred_original_sample` can be used to preview progress or for guidance.
     """
 
-    prev_sample: torch.FloatTensor
-    denoised: Optional[torch.FloatTensor] = None
+    prev_sample: torch.Tensor
+    denoised: Optional[torch.Tensor] = None
 
 
 # Copied from diffusers.schedulers.scheduling_ddpm.betas_for_alpha_bar
@@ -367,10 +367,10 @@ def rescale_zero_terminal_snr(betas):
     """
     Rescales betas to have zero terminal SNR Based on https://arxiv.org/pdf/2305.08891.pdf (Algorithm 1)
     Args:
-        betas (`torch.FloatTensor`):
+        betas (`torch.Tensor`):
             the betas that the scheduler is being initialized with.
     Returns:
-        `torch.FloatTensor`: rescaled betas with zero terminal SNR
+        `torch.Tensor`: rescaled betas with zero terminal SNR
     """
     # Convert betas to alphas_bar_sqrt
     alphas = 1.0 - betas
@@ -477,7 +477,7 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
             # Glide cosine schedule
             self.betas = betas_for_alpha_bar(num_train_timesteps)
         else:
-            raise NotImplementedError(f"{beta_schedule} does is not implemented for {self.__class__}")
+            raise NotImplementedError(f"{beta_schedule} is not implemented for {self.__class__}")
 
         # Rescale for zero SNR
         if rescale_betas_zero_snr:
@@ -499,17 +499,17 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
         self.num_inference_steps = None
         self.timesteps = torch.from_numpy(np.arange(0, num_train_timesteps)[::-1].copy().astype(np.int64))
 
-    def scale_model_input(self, sample: torch.FloatTensor, timestep: Optional[int] = None) -> torch.FloatTensor:
+    def scale_model_input(self, sample: torch.Tensor, timestep: Optional[int] = None) -> torch.Tensor:
         """
         Ensures interchangeability with schedulers that need to scale the denoising model input depending on the
         current timestep.
         Args:
-            sample (`torch.FloatTensor`):
+            sample (`torch.Tensor`):
                 The input sample.
             timestep (`int`, *optional*):
                 The current timestep in the diffusion chain.
         Returns:
-            `torch.FloatTensor`:
+            `torch.Tensor`:
                 A scaled input sample.
         """
         return sample
@@ -525,7 +525,7 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
         return variance
 
     # Copied from diffusers.schedulers.scheduling_ddpm.DDPMScheduler._threshold_sample
-    def _threshold_sample(self, sample: torch.FloatTensor) -> torch.FloatTensor:
+    def _threshold_sample(self, sample: torch.Tensor) -> torch.Tensor:
         """
         "Dynamic thresholding: At each sampling step we set s to a certain percentile absolute pixel value in xt0 (the
         prediction of x_0 at timestep t), and if s > 1, then we threshold xt0 to the range [-s, s] and then divide by
@@ -593,25 +593,25 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
 
     def step(
         self,
-        model_output: torch.FloatTensor,
+        model_output: torch.Tensor,
         timeindex: int,
         timestep: int,
-        sample: torch.FloatTensor,
+        sample: torch.Tensor,
         eta: float = 0.0,
         use_clipped_model_output: bool = False,
         generator=None,
-        variance_noise: Optional[torch.FloatTensor] = None,
+        variance_noise: Optional[torch.Tensor] = None,
         return_dict: bool = True,
     ) -> Union[LCMSchedulerOutput, Tuple]:
         """
         Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
         process from the learned model outputs (most often the predicted noise).
         Args:
-            model_output (`torch.FloatTensor`):
+            model_output (`torch.Tensor`):
                 The direct output from learned diffusion model.
             timestep (`float`):
                 The current discrete timestep in the diffusion chain.
-            sample (`torch.FloatTensor`):
+            sample (`torch.Tensor`):
                 A current instance of a sample created by the diffusion process.
             eta (`float`):
                 The weight of noise for added noise in diffusion step.
@@ -622,7 +622,7 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
                 `use_clipped_model_output` has no effect.
             generator (`torch.Generator`, *optional*):
                 A random number generator.
-            variance_noise (`torch.FloatTensor`):
+            variance_noise (`torch.Tensor`):
                 Alternative to generating noise with `generator` by directly providing the noise for the variance
                 itself. Useful for methods such as [`CycleDiffusion`].
             return_dict (`bool`, *optional*, defaults to `True`):
@@ -685,10 +685,10 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
     # Copied from diffusers.schedulers.scheduling_ddpm.DDPMScheduler.add_noise
     def add_noise(
         self,
-        original_samples: torch.FloatTensor,
-        noise: torch.FloatTensor,
+        original_samples: torch.Tensor,
+        noise: torch.Tensor,
         timesteps: torch.IntTensor,
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         # Make sure alphas_cumprod and timestep have same device and dtype as original_samples
         alphas_cumprod = self.alphas_cumprod.to(device=original_samples.device, dtype=original_samples.dtype)
         timesteps = timesteps.to(original_samples.device)
@@ -707,9 +707,7 @@ class LCMScheduler(SchedulerMixin, ConfigMixin):
         return noisy_samples
 
     # Copied from diffusers.schedulers.scheduling_ddpm.DDPMScheduler.get_velocity
-    def get_velocity(
-        self, sample: torch.FloatTensor, noise: torch.FloatTensor, timesteps: torch.IntTensor
-    ) -> torch.FloatTensor:
+    def get_velocity(self, sample: torch.Tensor, noise: torch.Tensor, timesteps: torch.IntTensor) -> torch.Tensor:
         # Make sure alphas_cumprod and timestep have same device and dtype as sample
         alphas_cumprod = self.alphas_cumprod.to(device=sample.device, dtype=sample.dtype)
         timesteps = timesteps.to(sample.device)

examples/community/llm_grounded_diffusion.py

@@ -756,13 +756,13 @@ class LLMGroundedDiffusionPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         clip_skip: Optional[int] = None,
@@ -807,14 +807,14 @@ class LLMGroundedDiffusionPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
@@ -825,7 +825,7 @@ class LLMGroundedDiffusionPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that calls every `callback_steps` steps during inference. The function is called with the
-                following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function is called. If not specified, the callback is called at
                 every step.
@@ -1194,8 +1194,8 @@ class LLMGroundedDiffusionPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         **kwargs,
     ):
@@ -1227,8 +1227,8 @@ class LLMGroundedDiffusionPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -1248,10 +1248,10 @@ class LLMGroundedDiffusionPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -1509,7 +1509,7 @@ class LLMGroundedDiffusionPipeline(
                 data type of the generated embeddings
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0

examples/community/lpw_stable_diffusion.py

@@ -378,7 +378,7 @@ def preprocess_image(image, batch_size):
 
 
 def preprocess_mask(mask, batch_size, scale_factor=8):
-    if not isinstance(mask, torch.FloatTensor):
+    if not isinstance(mask, torch.Tensor):
         mask = mask.convert("L")
         w, h = mask.size
         w, h = (x - x % 8 for x in (w, h))  # resize to integer multiple of 8
@@ -543,8 +543,8 @@ class StableDiffusionLongPromptWeightingPipeline(
         do_classifier_free_guidance,
         negative_prompt=None,
         max_embeddings_multiples=3,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
     ):
         r"""
         Encodes the prompt into text encoder hidden states.
@@ -767,8 +767,8 @@ class StableDiffusionLongPromptWeightingPipeline(
         self,
         prompt: Union[str, List[str]],
         negative_prompt: Optional[Union[str, List[str]]] = None,
-        image: Union[torch.FloatTensor, PIL.Image.Image] = None,
-        mask_image: Union[torch.FloatTensor, PIL.Image.Image] = None,
+        image: Union[torch.Tensor, PIL.Image.Image] = None,
+        mask_image: Union[torch.Tensor, PIL.Image.Image] = None,
         height: int = 512,
         width: int = 512,
         num_inference_steps: int = 50,
@@ -778,13 +778,13 @@ class StableDiffusionLongPromptWeightingPipeline(
         add_predicted_noise: Optional[bool] = False,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         max_embeddings_multiples: Optional[int] = 3,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         is_cancelled_callback: Optional[Callable[[], bool]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -798,10 +798,10 @@ class StableDiffusionLongPromptWeightingPipeline(
             negative_prompt (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
                 if `guidance_scale` is less than `1`).
-            image (`torch.FloatTensor` or `PIL.Image.Image`):
+            image (`torch.Tensor` or `PIL.Image.Image`):
                 `Image`, or tensor representing an image batch, that will be used as the starting point for the
                 process.
-            mask_image (`torch.FloatTensor` or `PIL.Image.Image`):
+            mask_image (`torch.Tensor` or `PIL.Image.Image`):
                 `Image`, or tensor representing an image batch, to mask `image`. White pixels in the mask will be
                 replaced by noise and therefore repainted, while black pixels will be preserved. If `mask_image` is a
                 PIL image, it will be converted to a single channel (luminance) before use. If it's a tensor, it should
@@ -836,14 +836,14 @@ class StableDiffusionLongPromptWeightingPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -857,7 +857,7 @@ class StableDiffusionLongPromptWeightingPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             is_cancelled_callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. If the function returns
                 `True`, the inference will be cancelled.
@@ -1032,13 +1032,13 @@ class StableDiffusionLongPromptWeightingPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         max_embeddings_multiples: Optional[int] = 3,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         is_cancelled_callback: Optional[Callable[[], bool]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -1072,14 +1072,14 @@ class StableDiffusionLongPromptWeightingPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -1093,7 +1093,7 @@ class StableDiffusionLongPromptWeightingPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             is_cancelled_callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. If the function returns
                 `True`, the inference will be cancelled.
@@ -1137,7 +1137,7 @@ class StableDiffusionLongPromptWeightingPipeline(
 
     def img2img(
         self,
-        image: Union[torch.FloatTensor, PIL.Image.Image],
+        image: Union[torch.Tensor, PIL.Image.Image],
         prompt: Union[str, List[str]],
         negative_prompt: Optional[Union[str, List[str]]] = None,
         strength: float = 0.8,
@@ -1146,12 +1146,12 @@ class StableDiffusionLongPromptWeightingPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: Optional[float] = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         max_embeddings_multiples: Optional[int] = 3,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         is_cancelled_callback: Optional[Callable[[], bool]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -1159,7 +1159,7 @@ class StableDiffusionLongPromptWeightingPipeline(
         r"""
         Function for image-to-image generation.
         Args:
-            image (`torch.FloatTensor` or `PIL.Image.Image`):
+            image (`torch.Tensor` or `PIL.Image.Image`):
                 `Image`, or tensor representing an image batch, that will be used as the starting point for the
                 process.
             prompt (`str` or `List[str]`):
@@ -1190,10 +1190,10 @@ class StableDiffusionLongPromptWeightingPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -1207,7 +1207,7 @@ class StableDiffusionLongPromptWeightingPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             is_cancelled_callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. If the function returns
                 `True`, the inference will be cancelled.
@@ -1249,8 +1249,8 @@ class StableDiffusionLongPromptWeightingPipeline(
 
     def inpaint(
         self,
-        image: Union[torch.FloatTensor, PIL.Image.Image],
-        mask_image: Union[torch.FloatTensor, PIL.Image.Image],
+        image: Union[torch.Tensor, PIL.Image.Image],
+        mask_image: Union[torch.Tensor, PIL.Image.Image],
         prompt: Union[str, List[str]],
         negative_prompt: Optional[Union[str, List[str]]] = None,
         strength: float = 0.8,
@@ -1260,12 +1260,12 @@ class StableDiffusionLongPromptWeightingPipeline(
         add_predicted_noise: Optional[bool] = False,
         eta: Optional[float] = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         max_embeddings_multiples: Optional[int] = 3,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         is_cancelled_callback: Optional[Callable[[], bool]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -1273,10 +1273,10 @@ class StableDiffusionLongPromptWeightingPipeline(
         r"""
         Function for inpaint.
         Args:
-            image (`torch.FloatTensor` or `PIL.Image.Image`):
+            image (`torch.Tensor` or `PIL.Image.Image`):
                 `Image`, or tensor representing an image batch, that will be used as the starting point for the
                 process. This is the image whose masked region will be inpainted.
-            mask_image (`torch.FloatTensor` or `PIL.Image.Image`):
+            mask_image (`torch.Tensor` or `PIL.Image.Image`):
                 `Image`, or tensor representing an image batch, to mask `image`. White pixels in the mask will be
                 replaced by noise and therefore repainted, while black pixels will be preserved. If `mask_image` is a
                 PIL image, it will be converted to a single channel (luminance) before use. If it's a tensor, it should
@@ -1311,10 +1311,10 @@ class StableDiffusionLongPromptWeightingPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -1328,7 +1328,7 @@ class StableDiffusionLongPromptWeightingPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             is_cancelled_callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. If the function returns
                 `True`, the inference will be cancelled.

examples/community/lpw_stable_diffusion_xl.py

@@ -694,10 +694,10 @@ class SDXLLongPromptWeightingPipeline(
         do_classifier_free_guidance: bool = True,
         negative_prompt: Optional[str] = None,
         negative_prompt_2: Optional[str] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
     ):
         r"""
@@ -722,17 +722,17 @@ class SDXLLongPromptWeightingPipeline(
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in both text-encoders
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -1320,7 +1320,7 @@ class SDXLLongPromptWeightingPipeline(
                 data type of the generated embeddings
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -1378,7 +1378,7 @@ class SDXLLongPromptWeightingPipeline(
         prompt_2: Optional[str] = None,
         image: Optional[PipelineImageInput] = None,
         mask_image: Optional[PipelineImageInput] = None,
-        masked_image_latents: Optional[torch.FloatTensor] = None,
+        masked_image_latents: Optional[torch.Tensor] = None,
         height: Optional[int] = None,
         width: Optional[int] = None,
         strength: float = 0.8,
@@ -1392,12 +1392,12 @@ class SDXLLongPromptWeightingPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -1481,23 +1481,23 @@ class SDXLLongPromptWeightingPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
             ip_adapter_image: (`PipelineImageInput`, *optional*):
                 Optional image input to work with IP Adapters.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -1926,12 +1926,12 @@ class SDXLLongPromptWeightingPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -2001,12 +2001,12 @@ class SDXLLongPromptWeightingPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -2066,7 +2066,7 @@ class SDXLLongPromptWeightingPipeline(
         prompt_2: Optional[str] = None,
         image: Optional[PipelineImageInput] = None,
         mask_image: Optional[PipelineImageInput] = None,
-        masked_image_latents: Optional[torch.FloatTensor] = None,
+        masked_image_latents: Optional[torch.Tensor] = None,
         height: Optional[int] = None,
         width: Optional[int] = None,
         strength: float = 0.8,
@@ -2080,12 +2080,12 @@ class SDXLLongPromptWeightingPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,

examples/community/marigold_depth_estimation.py

@@ -43,8 +43,7 @@ from diffusers.utils import BaseOutput, check_min_version
 
 
 # Will error if the minimal version of diffusers is not installed. Remove at your own risks.
-check_min_version("0.25.0")
-
+check_min_version("0.28.0")
 
 class MarigoldDepthOutput(BaseOutput):
     """

examples/community/masked_stable_diffusion_img2img.py

@@ -16,10 +16,10 @@ class MaskedStableDiffusionImg2ImgPipeline(StableDiffusionImg2ImgPipeline):
         self,
         prompt: Union[str, List[str]] = None,
         image: Union[
-            torch.FloatTensor,
+            torch.Tensor,
             PIL.Image.Image,
             np.ndarray,
-            List[torch.FloatTensor],
+            List[torch.Tensor],
             List[PIL.Image.Image],
             List[np.ndarray],
         ] = None,
@@ -30,18 +30,18 @@ class MaskedStableDiffusionImg2ImgPipeline(StableDiffusionImg2ImgPipeline):
         num_images_per_prompt: Optional[int] = 1,
         eta: Optional[float] = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         mask: Union[
-            torch.FloatTensor,
+            torch.Tensor,
             PIL.Image.Image,
             np.ndarray,
-            List[torch.FloatTensor],
+            List[torch.Tensor],
             List[PIL.Image.Image],
             List[np.ndarray],
         ] = None,
@@ -52,7 +52,7 @@ class MaskedStableDiffusionImg2ImgPipeline(StableDiffusionImg2ImgPipeline):
         Args:
             prompt (`str` or `List[str]`, *optional*):
                 The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
-            image (`torch.FloatTensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.FloatTensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`):
+            image (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`):
                 `Image` or tensor representing an image batch to be used as the starting point. Can also accept image
                 latents as `image`, but if passing latents directly it is not encoded again.
             strength (`float`, *optional*, defaults to 0.8):
@@ -78,10 +78,10 @@ class MaskedStableDiffusionImg2ImgPipeline(StableDiffusionImg2ImgPipeline):
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):
@@ -91,14 +91,14 @@ class MaskedStableDiffusionImg2ImgPipeline(StableDiffusionImg2ImgPipeline):
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that calls every `callback_steps` steps during inference. The function is called with the
-                following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function is called. If not specified, the callback is called at
                 every step.
             cross_attention_kwargs (`dict`, *optional*):
                 A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
                 [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
-            mask (`torch.FloatTensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.FloatTensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`, *optional*):
+            mask (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`, *optional*):
                 A mask with non-zero elements for the area to be inpainted. If not specified, no mask is applied.
         Examples:
 

examples/community/mixture_canvas.py

@@ -154,7 +154,7 @@ class Text2ImageRegion(DiffusionRegion):
 class Image2ImageRegion(DiffusionRegion):
     """Class defining a region where an image guided diffusion process is acting"""
 
-    reference_image: torch.FloatTensor = None
+    reference_image: torch.Tensor = None
     strength: float = 0.8  # Strength of the image
 
     def __post_init__(self):

examples/community/multilingual_stable_diffusion.py

@@ -147,10 +147,10 @@ class MultilingualStableDiffusion(DiffusionPipeline, StableDiffusionMixin):
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[torch.Generator] = None,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         **kwargs,
     ):
@@ -184,7 +184,7 @@ class MultilingualStableDiffusion(DiffusionPipeline, StableDiffusionMixin):
             generator (`torch.Generator`, *optional*):
                 A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
                 deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
@@ -196,7 +196,7 @@ class MultilingualStableDiffusion(DiffusionPipeline, StableDiffusionMixin):
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/pipeline_animatediff_controlnet.py

@@ -198,8 +198,8 @@ class AnimateDiffControlNetPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -219,10 +219,10 @@ class AnimateDiffControlNetPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -752,9 +752,9 @@ class AnimateDiffControlNetPipeline(
         num_videos_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
         ip_adapter_image_embeds: Optional[PipelineImageInput] = None,
         conditioning_frames: Optional[List[PipelineImageInput]] = None,
@@ -798,20 +798,20 @@ class AnimateDiffControlNetPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for video
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`. Latents should be of shape
                 `(batch_size, num_channel, num_frames, height, width)`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             ip_adapter_image (`PipelineImageInput`, *optional*):
                 Optional image input to work with IP Adapters.
-            ip_adapter_image_embeds (`List[torch.FloatTensor]`, *optional*):
+            ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*):
                 Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters.
                 Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. It should contain the negative image embedding
                 if `do_classifier_free_guidance` is set to `True`.
@@ -821,7 +821,7 @@ class AnimateDiffControlNetPipeline(
                 are specified, images must be passed as a list such that each element of the list can be correctly
                 batched for input to a single ControlNet.
             output_type (`str`, *optional*, defaults to `"pil"`):
-                The output format of the generated video. Choose between `torch.FloatTensor`, `PIL.Image` or
+                The output format of the generated video. Choose between `torch.Tensor`, `PIL.Image` or
                 `np.array`.
             return_dict (`bool`, *optional*, defaults to `True`):
                 Whether or not to return a [`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] instead

examples/community/pipeline_animatediff_img2video.py

@@ -315,8 +315,8 @@ class AnimateDiffImgToVideoPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -336,10 +336,10 @@ class AnimateDiffImgToVideoPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -746,14 +746,14 @@ class AnimateDiffImgToVideoPipeline(
         num_videos_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
         ip_adapter_image_embeds: Optional[PipelineImageInput] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: Optional[int] = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         clip_skip: Optional[int] = None,
@@ -791,33 +791,33 @@ class AnimateDiffImgToVideoPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for video
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`. Latents should be of shape
                 `(batch_size, num_channel, num_frames, height, width)`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             ip_adapter_image: (`PipelineImageInput`, *optional*):
                 Optional image input to work with IP Adapters.
-            ip_adapter_image_embeds (`List[torch.FloatTensor]`, *optional*):
+            ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*):
                 Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters.
                 Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. It should contain the negative image embedding
                 if `do_classifier_free_guidance` is set to `True`.
                 If not provided, embeddings are computed from the `ip_adapter_image` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):
-                The output format of the generated video. Choose between `torch.FloatTensor`, `PIL.Image` or
+                The output format of the generated video. Choose between `torch.Tensor`, `PIL.Image` or
                 `np.array`.
             return_dict (`bool`, *optional*, defaults to `True`):
                 Whether or not to return a [`AnimateDiffImgToVideoPipelineOutput`] instead
                 of a plain tuple.
             callback (`Callable`, *optional*):
                 A function that calls every `callback_steps` steps during inference. The function is called with the
-                following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function is called. If not specified, the callback is called at
                 every step.

examples/community/pipeline_demofusion_sdxl.py

@@ -187,10 +187,10 @@ class DemoFusionSDXLPipeline(
         do_classifier_free_guidance: bool = True,
         negative_prompt: Optional[str] = None,
         negative_prompt_2: Optional[str] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
     ):
         r"""
@@ -215,17 +215,17 @@ class DemoFusionSDXLPipeline(
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in both text-encoders
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -642,14 +642,14 @@ class DemoFusionSDXLPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = False,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         guidance_rescale: float = 0.0,
@@ -720,21 +720,21 @@ class DemoFusionSDXLPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -746,7 +746,7 @@ class DemoFusionSDXLPipeline(
                 of a plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/pipeline_fabric.py

@@ -190,8 +190,8 @@ class FabricPipeline(DiffusionPipeline):
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
     ):
         r"""
@@ -210,10 +210,10 @@ class FabricPipeline(DiffusionPipeline):
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -512,7 +512,7 @@ class FabricPipeline(DiffusionPipeline):
         neg_scale: float = 0.5,
         pos_bottleneck_scale: float = 1.0,
         neg_bottleneck_scale: float = 1.0,
-        latents: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
     ):
         r"""
         The call function to the pipeline for generation. Generate a trajectory of images with binary feedback. The

examples/community/pipeline_prompt2prompt.py

@@ -217,8 +217,8 @@ class Prompt2PromptPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         **kwargs,
     ):
@@ -250,8 +250,8 @@ class Prompt2PromptPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -271,10 +271,10 @@ class Prompt2PromptPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -564,12 +564,12 @@ class Prompt2PromptPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: Optional[int] = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         guidance_rescale: float = 0.0,
@@ -604,7 +604,7 @@ class Prompt2PromptPipeline(
             generator (`torch.Generator`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
@@ -616,7 +616,7 @@ class Prompt2PromptPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/pipeline_sdxl_style_aligned.py

@@ -514,10 +514,10 @@ class StyleAlignedSDXLPipeline(
         do_classifier_free_guidance: bool = True,
         negative_prompt: Optional[str] = None,
         negative_prompt_2: Optional[str] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -543,17 +543,17 @@ class StyleAlignedSDXLPipeline(
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in both text-encoders
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -1325,7 +1325,7 @@ class StyleAlignedSDXLPipeline(
                 data type of the generated embeddings
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -1387,7 +1387,7 @@ class StyleAlignedSDXLPipeline(
         prompt_2: Optional[Union[str, List[str]]] = None,
         image: Optional[PipelineImageInput] = None,
         mask_image: Optional[PipelineImageInput] = None,
-        masked_image_latents: Optional[torch.FloatTensor] = None,
+        masked_image_latents: Optional[torch.Tensor] = None,
         strength: float = 0.3,
         height: Optional[int] = None,
         width: Optional[int] = None,
@@ -1401,11 +1401,11 @@ class StyleAlignedSDXLPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
@@ -1474,21 +1474,21 @@ class StyleAlignedSDXLPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.

examples/community/pipeline_stable_diffusion_pag.py

@@ -1,4 +1,5 @@
-# Implementation of StableDiffusionPAGPipeline
+# Implementation of StableDiffusionPipeline with PAG
+# https://ku-cvlab.github.io/Perturbed-Attention-Guidance
 
 import inspect
 from typing import Any, Callable, Dict, List, Optional, Union
@@ -56,13 +57,13 @@ class PAGIdentitySelfAttnProcessor:
     def __call__(
         self,
         attn: Attention,
-        hidden_states: torch.FloatTensor,
-        encoder_hidden_states: Optional[torch.FloatTensor] = None,
-        attention_mask: Optional[torch.FloatTensor] = None,
-        temb: Optional[torch.FloatTensor] = None,
+        hidden_states: torch.Tensor,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        temb: Optional[torch.Tensor] = None,
         *args,
         **kwargs,
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         if len(args) > 0 or kwargs.get("scale", None) is not None:
             deprecation_message = "The `scale` argument is deprecated and will be ignored. Please remove it, as passing it will raise an error in the future. `scale` should directly be passed while calling the underlying pipeline component i.e., via `cross_attention_kwargs`."
             deprecate("scale", "1.0.0", deprecation_message)
@@ -134,8 +135,8 @@ class PAGIdentitySelfAttnProcessor:
 
         value = attn.to_v(hidden_states_ptb)
 
-        hidden_states_ptb = torch.zeros(value.shape).to(value.get_device())
-        # hidden_states_ptb = value
+        # hidden_states_ptb = torch.zeros(value.shape).to(value.get_device())
+        hidden_states_ptb = value
 
         hidden_states_ptb = hidden_states_ptb.to(query.dtype)
 
@@ -170,13 +171,13 @@ class PAGCFGIdentitySelfAttnProcessor:
     def __call__(
         self,
         attn: Attention,
-        hidden_states: torch.FloatTensor,
-        encoder_hidden_states: Optional[torch.FloatTensor] = None,
-        attention_mask: Optional[torch.FloatTensor] = None,
-        temb: Optional[torch.FloatTensor] = None,
+        hidden_states: torch.Tensor,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        temb: Optional[torch.Tensor] = None,
         *args,
         **kwargs,
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         if len(args) > 0 or kwargs.get("scale", None) is not None:
             deprecation_message = "The `scale` argument is deprecated and will be ignored. Please remove it, as passing it will raise an error in the future. `scale` should directly be passed while calling the underlying pipeline component i.e., via `cross_attention_kwargs`."
             deprecate("scale", "1.0.0", deprecation_message)
@@ -492,8 +493,8 @@ class StableDiffusionPAGPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         **kwargs,
     ):
@@ -524,8 +525,8 @@ class StableDiffusionPAGPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -544,10 +545,10 @@ class StableDiffusionPAGPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -965,7 +966,7 @@ class StableDiffusionPAGPipeline(
             dtype:
                 data type of the generated embeddings
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
+            `torch.Tensor`: Embedding vectors with shape `(len(timesteps), embedding_dim)`
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -1045,7 +1046,7 @@ class StableDiffusionPAGPipeline(
         return self._pag_scale
 
     @property
-    def do_adversarial_guidance(self):
+    def do_perturbed_attention_guidance(self):
         return self._pag_scale > 0
 
     @property
@@ -1056,14 +1057,6 @@ class StableDiffusionPAGPipeline(
     def do_pag_adaptive_scaling(self):
         return self._pag_adaptive_scaling > 0
 
-    @property
-    def pag_drop_rate(self):
-        return self._pag_drop_rate
-
-    @property
-    def pag_applied_layers(self):
-        return self._pag_applied_layers
-
     @property
     def pag_applied_layers_index(self):
         return self._pag_applied_layers_index
@@ -1080,18 +1073,16 @@ class StableDiffusionPAGPipeline(
         guidance_scale: float = 7.5,
         pag_scale: float = 0.0,
         pag_adaptive_scaling: float = 0.0,
-        pag_drop_rate: float = 0.5,
-        pag_applied_layers: List[str] = ["down"],  # ['down', 'mid', 'up']
         pag_applied_layers_index: List[str] = ["d4"],  # ['d4', 'd5', 'm0']
         negative_prompt: Optional[Union[str, List[str]]] = None,
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
-        ip_adapter_image_embeds: Optional[List[torch.FloatTensor]] = None,
+        ip_adapter_image_embeds: Optional[List[torch.Tensor]] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -1131,18 +1122,18 @@ class StableDiffusionPAGPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
-            ip_adapter_image_embeds (`List[torch.FloatTensor]`, *optional*):
+            ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*):
                 Pre-generated image embeddings for IP-Adapter. If not
                 provided, embeddings are computed from the `ip_adapter_image` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):
@@ -1221,8 +1212,6 @@ class StableDiffusionPAGPipeline(
 
         self._pag_scale = pag_scale
         self._pag_adaptive_scaling = pag_adaptive_scaling
-        self._pag_drop_rate = pag_drop_rate
-        self._pag_applied_layers = pag_applied_layers
         self._pag_applied_layers_index = pag_applied_layers_index
 
         # 2. Define call parameters
@@ -1257,13 +1246,13 @@ class StableDiffusionPAGPipeline(
         # to avoid doing two forward passes
 
         # cfg
-        if self.do_classifier_free_guidance and not self.do_adversarial_guidance:
+        if self.do_classifier_free_guidance and not self.do_perturbed_attention_guidance:
             prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
         # pag
-        elif not self.do_classifier_free_guidance and self.do_adversarial_guidance:
+        elif not self.do_classifier_free_guidance and self.do_perturbed_attention_guidance:
             prompt_embeds = torch.cat([prompt_embeds, prompt_embeds])
         # both
-        elif self.do_classifier_free_guidance and self.do_adversarial_guidance:
+        elif self.do_classifier_free_guidance and self.do_perturbed_attention_guidance:
             prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds, prompt_embeds])
 
         if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
@@ -1306,7 +1295,7 @@ class StableDiffusionPAGPipeline(
             ).to(device=device, dtype=latents.dtype)
 
         # 7. Denoising loop
-        if self.do_adversarial_guidance:
+        if self.do_perturbed_attention_guidance:
             down_layers = []
             mid_layers = []
             up_layers = []
@@ -1322,6 +1311,29 @@ class StableDiffusionPAGPipeline(
                     else:
                         raise ValueError(f"Invalid layer type: {layer_type}")
 
+        # change attention layer in UNet if use PAG
+        if self.do_perturbed_attention_guidance:
+            if self.do_classifier_free_guidance:
+                replace_processor = PAGCFGIdentitySelfAttnProcessor()
+            else:
+                replace_processor = PAGIdentitySelfAttnProcessor()
+
+            drop_layers = self.pag_applied_layers_index
+            for drop_layer in drop_layers:
+                try:
+                    if drop_layer[0] == "d":
+                        down_layers[int(drop_layer[1])].processor = replace_processor
+                    elif drop_layer[0] == "m":
+                        mid_layers[int(drop_layer[1])].processor = replace_processor
+                    elif drop_layer[0] == "u":
+                        up_layers[int(drop_layer[1])].processor = replace_processor
+                    else:
+                        raise ValueError(f"Invalid layer type: {drop_layer[0]}")
+                except IndexError:
+                    raise ValueError(
+                        f"Invalid layer index: {drop_layer}. Available layers: {len(down_layers)} down layers, {len(mid_layers)} mid layers, {len(up_layers)} up layers."
+                    )
+
         num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
         self._num_timesteps = len(timesteps)
         with self.progress_bar(total=num_inference_steps) as progress_bar:
@@ -1330,41 +1342,18 @@ class StableDiffusionPAGPipeline(
                     continue
 
                 # cfg
-                if self.do_classifier_free_guidance and not self.do_adversarial_guidance:
+                if self.do_classifier_free_guidance and not self.do_perturbed_attention_guidance:
                     latent_model_input = torch.cat([latents] * 2)
                 # pag
-                elif not self.do_classifier_free_guidance and self.do_adversarial_guidance:
+                elif not self.do_classifier_free_guidance and self.do_perturbed_attention_guidance:
                     latent_model_input = torch.cat([latents] * 2)
                 # both
-                elif self.do_classifier_free_guidance and self.do_adversarial_guidance:
+                elif self.do_classifier_free_guidance and self.do_perturbed_attention_guidance:
                     latent_model_input = torch.cat([latents] * 3)
                 # no
                 else:
                     latent_model_input = latents
 
-                # change attention layer in UNet if use PAG
-                if self.do_adversarial_guidance:
-                    if self.do_classifier_free_guidance:
-                        replace_processor = PAGCFGIdentitySelfAttnProcessor()
-                    else:
-                        replace_processor = PAGIdentitySelfAttnProcessor()
-
-                    drop_layers = self.pag_applied_layers_index
-                    for drop_layer in drop_layers:
-                        try:
-                            if drop_layer[0] == "d":
-                                down_layers[int(drop_layer[1])].processor = replace_processor
-                            elif drop_layer[0] == "m":
-                                mid_layers[int(drop_layer[1])].processor = replace_processor
-                            elif drop_layer[0] == "u":
-                                up_layers[int(drop_layer[1])].processor = replace_processor
-                            else:
-                                raise ValueError(f"Invalid layer type: {drop_layer[0]}")
-                        except IndexError:
-                            raise ValueError(
-                                f"Invalid layer index: {drop_layer}. Available layers: {len(down_layers)} down layers, {len(mid_layers)} mid layers, {len(up_layers)} up layers."
-                            )
-
                 latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
 
                 # predict the noise residual
@@ -1381,14 +1370,14 @@ class StableDiffusionPAGPipeline(
                 # perform guidance
 
                 # cfg
-                if self.do_classifier_free_guidance and not self.do_adversarial_guidance:
+                if self.do_classifier_free_guidance and not self.do_perturbed_attention_guidance:
                     noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
 
                     delta = noise_pred_text - noise_pred_uncond
                     noise_pred = noise_pred_uncond + self.guidance_scale * delta
 
                 # pag
-                elif not self.do_classifier_free_guidance and self.do_adversarial_guidance:
+                elif not self.do_classifier_free_guidance and self.do_perturbed_attention_guidance:
                     noise_pred_original, noise_pred_perturb = noise_pred.chunk(2)
 
                     signal_scale = self.pag_scale
@@ -1400,7 +1389,7 @@ class StableDiffusionPAGPipeline(
                     noise_pred = noise_pred_original + signal_scale * (noise_pred_original - noise_pred_perturb)
 
                 # both
-                elif self.do_classifier_free_guidance and self.do_adversarial_guidance:
+                elif self.do_classifier_free_guidance and self.do_perturbed_attention_guidance:
                     noise_pred_uncond, noise_pred_text, noise_pred_text_perturb = noise_pred.chunk(3)
 
                     signal_scale = self.pag_scale
@@ -1458,11 +1447,8 @@ class StableDiffusionPAGPipeline(
         # Offload all models
         self.maybe_free_model_hooks()
 
-        if not return_dict:
-            return (image, has_nsfw_concept)
-
         # change attention layer in UNet if use PAG
-        if self.do_adversarial_guidance:
+        if self.do_perturbed_attention_guidance:
             drop_layers = self.pag_applied_layers_index
             for drop_layer in drop_layers:
                 try:
@@ -1479,4 +1465,7 @@ class StableDiffusionPAGPipeline(
                         f"Invalid layer index: {drop_layer}. Available layers: {len(down_layers)} down layers, {len(mid_layers)} mid layers, {len(up_layers)} up layers."
                     )
 
+        if not return_dict:
+            return (image, has_nsfw_concept)
+
         return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

examples/community/pipeline_stable_diffusion_upscale_ldm3d.py

@@ -164,8 +164,8 @@ class StableDiffusionUpscaleLDM3DPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         **kwargs,
     ):
@@ -197,8 +197,8 @@ class StableDiffusionUpscaleLDM3DPipeline(
         num_images_per_prompt,
         do_classifier_free_guidance,
         negative_prompt=None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -218,10 +218,10 @@ class StableDiffusionUpscaleLDM3DPipeline(
                 The prompt or prompts not to guide the image generation. If not defined, one has to pass
                 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                 less than `1`).
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
@@ -460,7 +460,7 @@ class StableDiffusionUpscaleLDM3DPipeline(
             )
 
         # verify batch size of prompt and image are same if image is a list or tensor or numpy array
-        if isinstance(image, list) or isinstance(image, torch.Tensor) or isinstance(image, np.ndarray):
+        if isinstance(image, (list, np.ndarray, torch.Tensor)):
             if prompt is not None and isinstance(prompt, str):
                 batch_size = 1
             elif prompt is not None and isinstance(prompt, list):
@@ -535,12 +535,12 @@ class StableDiffusionUpscaleLDM3DPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         target_res: Optional[List[int]] = [1024, 1024],
@@ -551,7 +551,7 @@ class StableDiffusionUpscaleLDM3DPipeline(
         Args:
             prompt (`str` or `List[str]`, *optional*):
                 The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
-            image (`torch.FloatTensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.FloatTensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`):
+            image (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`):
                 `Image` or tensor representing an image batch to be upscaled.
             num_inference_steps (`int`, *optional*, defaults to 50):
                 The number of denoising steps. More denoising steps usually lead to a higher quality image at the
@@ -570,14 +570,14 @@ class StableDiffusionUpscaleLDM3DPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                 generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor is generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                 provided, text embeddings are generated from the `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):
@@ -587,7 +587,7 @@ class StableDiffusionUpscaleLDM3DPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that calls every `callback_steps` steps during inference. The function is called with the
-                following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function is called. If not specified, the callback is called at
                 every step.

examples/community/pipeline_stable_diffusion_xl_controlnet_adapter.py

@@ -248,10 +248,10 @@ class StableDiffusionXLControlNetAdapterPipeline(
         do_classifier_free_guidance: bool = True,
         negative_prompt: Optional[str] = None,
         negative_prompt_2: Optional[str] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -277,17 +277,17 @@ class StableDiffusionXLControlNetAdapterPipeline(
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in both text-encoders
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -887,14 +887,14 @@ class StableDiffusionXLControlNetAdapterPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         guidance_rescale: float = 0.0,
@@ -922,14 +922,14 @@ class StableDiffusionXLControlNetAdapterPipeline(
             prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts to be sent to the `tokenizer_2` and `text_encoder_2`. If not defined, `prompt` is
                 used in both text-encoders
-            adapter_image (`torch.FloatTensor`, `PIL.Image.Image`, `List[torch.FloatTensor]` or `List[PIL.Image.Image]` or `List[List[PIL.Image.Image]]`):
+            adapter_image (`torch.Tensor`, `PIL.Image.Image`, `List[torch.Tensor]` or `List[PIL.Image.Image]` or `List[List[PIL.Image.Image]]`):
                 The Adapter input condition. Adapter uses this input condition to generate guidance to Unet. If the
-                type is specified as `Torch.FloatTensor`, it is passed to Adapter as is. PIL.Image.Image` can also be
+                type is specified as `torch.Tensor`, it is passed to Adapter as is. PIL.Image.Image` can also be
                 accepted as an image. The control image is automatically resized to fit the output image.
-            control_image (`torch.FloatTensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.FloatTensor]`, `List[PIL.Image.Image]`, `List[np.ndarray]`,:
-                    `List[List[torch.FloatTensor]]`, `List[List[np.ndarray]]` or `List[List[PIL.Image.Image]]`):
+            control_image (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, `List[np.ndarray]`,:
+                    `List[List[torch.Tensor]]`, `List[List[np.ndarray]]` or `List[List[PIL.Image.Image]]`):
                 The ControlNet input condition to provide guidance to the `unet` for generation. If the type is
-                specified as `torch.FloatTensor`, it is passed to ControlNet as is. `PIL.Image.Image` can also be
+                specified as `torch.Tensor`, it is passed to ControlNet as is. `PIL.Image.Image` can also be
                 accepted as an image. The dimensions of the output image defaults to `image`'s dimensions. If height
                 and/or width are passed, `image` is resized accordingly. If multiple ControlNets are specified in
                 `init`, images must be passed as a list such that each element of the list can be correctly batched for
@@ -973,21 +973,21 @@ class StableDiffusionXLControlNetAdapterPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -999,7 +999,7 @@ class StableDiffusionXLControlNetAdapterPipeline(
                 instead of a plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/pipeline_stable_diffusion_xl_controlnet_adapter_inpaint.py

@@ -396,10 +396,10 @@ class StableDiffusionXLControlNetAdapterInpaintPipeline(
         do_classifier_free_guidance: bool = True,
         negative_prompt: Optional[str] = None,
         negative_prompt_2: Optional[str] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -425,17 +425,17 @@ class StableDiffusionXLControlNetAdapterInpaintPipeline(
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in both text-encoders
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -1229,14 +1229,14 @@ class StableDiffusionXLControlNetAdapterInpaintPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[Union[torch.FloatTensor]] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[Union[torch.Tensor]] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
-        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
+        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
         callback_steps: int = 1,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
         guidance_rescale: float = 0.0,
@@ -1270,14 +1270,14 @@ class StableDiffusionXLControlNetAdapterInpaintPipeline(
                 repainted, while black pixels will be preserved. If `mask_image` is a PIL image, it will be converted
                 to a single channel (luminance) before use. If it's a tensor, it should contain one color channel (L)
                 instead of 3, so the expected shape would be `(B, H, W, 1)`.
-            adapter_image (`torch.FloatTensor`, `PIL.Image.Image`, `List[torch.FloatTensor]` or `List[PIL.Image.Image]` or `List[List[PIL.Image.Image]]`):
+            adapter_image (`torch.Tensor`, `PIL.Image.Image`, `List[torch.Tensor]` or `List[PIL.Image.Image]` or `List[List[PIL.Image.Image]]`):
                 The Adapter input condition. Adapter uses this input condition to generate guidance to Unet. If the
-                type is specified as `Torch.FloatTensor`, it is passed to Adapter as is. PIL.Image.Image` can also be
+                type is specified as `torch.Tensor`, it is passed to Adapter as is. PIL.Image.Image` can also be
                 accepted as an image. The control image is automatically resized to fit the output image.
-            control_image (`torch.FloatTensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.FloatTensor]`, `List[PIL.Image.Image]`, `List[np.ndarray]`,:
-                    `List[List[torch.FloatTensor]]`, `List[List[np.ndarray]]` or `List[List[PIL.Image.Image]]`):
+            control_image (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, `List[np.ndarray]`,:
+                    `List[List[torch.Tensor]]`, `List[List[np.ndarray]]` or `List[List[PIL.Image.Image]]`):
                 The ControlNet input condition to provide guidance to the `unet` for generation. If the type is
-                specified as `torch.FloatTensor`, it is passed to ControlNet as is. `PIL.Image.Image` can also be
+                specified as `torch.Tensor`, it is passed to ControlNet as is. `PIL.Image.Image` can also be
                 accepted as an image. The dimensions of the output image defaults to `image`'s dimensions. If height
                 and/or width are passed, `image` is resized accordingly. If multiple ControlNets are specified in
                 `init`, images must be passed as a list such that each element of the list can be correctly batched for
@@ -1330,21 +1330,21 @@ class StableDiffusionXLControlNetAdapterInpaintPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -1356,7 +1356,7 @@ class StableDiffusionXLControlNetAdapterInpaintPipeline(
                 instead of a plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.

examples/community/pipeline_stable_diffusion_xl_differential_img2img.py

@@ -280,10 +280,10 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
         do_classifier_free_guidance: bool = True,
         negative_prompt: Optional[str] = None,
         negative_prompt_2: Optional[str] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         lora_scale: Optional[float] = None,
         clip_skip: Optional[int] = None,
     ):
@@ -309,17 +309,17 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in both text-encoders
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
@@ -868,7 +868,7 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
     # Copied from diffusers.pipelines.latent_consistency_models.pipeline_latent_consistency_text2img.LatentConsistencyModelPipeline.get_guidance_scale_embedding
     def get_guidance_scale_embedding(
         self, w: torch.Tensor, embedding_dim: int = 512, dtype: torch.dtype = torch.float32
-    ) -> torch.FloatTensor:
+    ) -> torch.Tensor:
         """
         See https://github.com/google-research/vdm/blob/dc27b98a554f65cdc654b800da5aa1846545d41b/model_vdm.py#L298
 
@@ -881,7 +881,7 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
                 Data type of the generated embeddings.
 
         Returns:
-            `torch.FloatTensor`: Embedding vectors with shape `(len(w), embedding_dim)`.
+            `torch.Tensor`: Embedding vectors with shape `(len(w), embedding_dim)`.
         """
         assert len(w.shape) == 1
         w = w * 1000.0
@@ -942,10 +942,10 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
         prompt: Union[str, List[str]] = None,
         prompt_2: Optional[Union[str, List[str]]] = None,
         image: Union[
-            torch.FloatTensor,
+            torch.Tensor,
             PIL.Image.Image,
             np.ndarray,
-            List[torch.FloatTensor],
+            List[torch.Tensor],
             List[PIL.Image.Image],
             List[np.ndarray],
         ] = None,
@@ -960,13 +960,13 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
         num_images_per_prompt: Optional[int] = 1,
         eta: float = 0.0,
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
-        latents: Optional[torch.FloatTensor] = None,
-        prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
-        pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
-        negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
+        latents: Optional[torch.Tensor] = None,
+        prompt_embeds: Optional[torch.Tensor] = None,
+        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
         ip_adapter_image: Optional[PipelineImageInput] = None,
-        ip_adapter_image_embeds: Optional[List[torch.FloatTensor]] = None,
+        ip_adapter_image_embeds: Optional[List[torch.Tensor]] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -982,12 +982,12 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
         clip_skip: Optional[int] = None,
         callback_on_step_end: Optional[Callable[[int, int, Dict], None]] = None,
         callback_on_step_end_tensor_inputs: List[str] = ["latents"],
-        map: torch.FloatTensor = None,
+        map: torch.Tensor = None,
         original_image: Union[
-            torch.FloatTensor,
+            torch.Tensor,
             PIL.Image.Image,
             np.ndarray,
-            List[torch.FloatTensor],
+            List[torch.Tensor],
             List[PIL.Image.Image],
             List[np.ndarray],
         ] = None,
@@ -1003,7 +1003,7 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
             prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts to be sent to the `tokenizer_2` and `text_encoder_2`. If not defined, `prompt` is
                 used in both text-encoders
-            image (`torch.FloatTensor` or `PIL.Image.Image` or `np.ndarray` or `List[torch.FloatTensor]` or `List[PIL.Image.Image]` or `List[np.ndarray]`):
+            image (`torch.Tensor` or `PIL.Image.Image` or `np.ndarray` or `List[torch.Tensor]` or `List[PIL.Image.Image]` or `List[np.ndarray]`):
                 The image(s) to modify with the pipeline.
             strength (`float`, *optional*, defaults to 0.3):
                 Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1. `image`
@@ -1051,26 +1051,26 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                 to make generation deterministic.
-            latents (`torch.FloatTensor`, *optional*):
+            latents (`torch.Tensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                 tensor will ge generated by sampling using the supplied random `generator`.
-            prompt_embeds (`torch.FloatTensor`, *optional*):
+            prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
-            negative_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                 argument.
-            pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting.
                 If not provided, pooled text embeddings will be generated from `prompt` input argument.
-            negative_pooled_prompt_embeds (`torch.FloatTensor`, *optional*):
+            negative_pooled_prompt_embeds (`torch.Tensor`, *optional*):
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
             ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
-            ip_adapter_image_embeds (`List[torch.FloatTensor]`, *optional*):
+            ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*):
                 Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters.
                 Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. It should contain the negative image embedding
                 if `do_classifier_free_guidance` is set to `True`.
@@ -1083,7 +1083,7 @@ class StableDiffusionXLDifferentialImg2ImgPipeline(
                 plain tuple.
             callback (`Callable`, *optional*):
                 A function that will be called every `callback_steps` steps during inference. The function will be
-                called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
+                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
             callback_steps (`int`, *optional*, defaults to 1):
                 The frequency at which the `callback` function will be called. If not specified, the callback will be
                 called at every step.