Duplicate from Tune-A-Video-library/Tune-A-Video-inference

Co-authored-by: hysts <hysts@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,35 @@
+*.whl filter=lfs diff=lfs merge=lfs -text
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,164 @@
+checkpoints/
+experiments/
+
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "Tune-A-Video"]
+	path = Tune-A-Video
+	url = https://github.com/showlab/Tune-A-Video

.pre-commit-config.yaml

@@ -0,0 +1,37 @@
+exclude: patch
+repos:
+- repo: https://github.com/pre-commit/pre-commit-hooks
+  rev: v4.2.0
+  hooks:
+  - id: check-executables-have-shebangs
+  - id: check-json
+  - id: check-merge-conflict
+  - id: check-shebang-scripts-are-executable
+  - id: check-toml
+  - id: check-yaml
+  - id: double-quote-string-fixer
+  - id: end-of-file-fixer
+  - id: mixed-line-ending
+    args: ['--fix=lf']
+  - id: requirements-txt-fixer
+  - id: trailing-whitespace
+- repo: https://github.com/myint/docformatter
+  rev: v1.4
+  hooks:
+  - id: docformatter
+    args: ['--in-place']
+- repo: https://github.com/pycqa/isort
+  rev: 5.12.0
+  hooks:
+    - id: isort
+- repo: https://github.com/pre-commit/mirrors-mypy
+  rev: v0.991
+  hooks:
+    - id: mypy
+      args: ['--ignore-missing-imports']
+      additional_dependencies: ['types-python-slugify']
+- repo: https://github.com/google/yapf
+  rev: v0.32.0
+  hooks:
+  - id: yapf
+    args: ['--parallel', '--in-place']

.style.yapf

@@ -0,0 +1,5 @@
+[style]
+based_on_style = pep8
+blank_line_before_nested_class_or_def = false
+spaces_before_comment = 2
+split_before_logical_operator = true

Dockerfile

@@ -0,0 +1,57 @@
+FROM nvidia/cuda:11.7.1-cudnn8-devel-ubuntu22.04
+ENV DEBIAN_FRONTEND=noninteractive
+RUN apt-get update && \
+    apt-get upgrade -y && \
+    apt-get install -y --no-install-recommends \
+    git \
+    git-lfs \
+    wget \
+    curl \
+    # ffmpeg \
+    ffmpeg \
+    x264 \
+    # python build dependencies \
+    build-essential \
+    libssl-dev \
+    zlib1g-dev \
+    libbz2-dev \
+    libreadline-dev \
+    libsqlite3-dev \
+    libncursesw5-dev \
+    xz-utils \
+    tk-dev \
+    libxml2-dev \
+    libxmlsec1-dev \
+    libffi-dev \
+    liblzma-dev && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:${PATH}
+WORKDIR ${HOME}/app
+
+RUN curl https://pyenv.run | bash
+ENV PATH=${HOME}/.pyenv/shims:${HOME}/.pyenv/bin:${PATH}
+ENV PYTHON_VERSION=3.10.9
+RUN pyenv install ${PYTHON_VERSION} && \
+    pyenv global ${PYTHON_VERSION} && \
+    pyenv rehash && \
+    pip install --no-cache-dir -U pip setuptools wheel
+
+RUN pip install --no-cache-dir -U torch==1.13.1 torchvision==0.14.1
+COPY --chown=1000 requirements.txt /tmp/requirements.txt
+RUN pip install --no-cache-dir -U -r /tmp/requirements.txt
+
+COPY --chown=1000 . ${HOME}/app
+RUN cd Tune-A-Video && patch -p1 < ../patch
+ENV PYTHONPATH=${HOME}/app \
+    PYTHONUNBUFFERED=1 \
+    GRADIO_ALLOW_FLAGGING=never \
+    GRADIO_NUM_PORTS=1 \
+    GRADIO_SERVER_NAME=0.0.0.0 \
+    GRADIO_THEME=huggingface \
+    SYSTEM=spaces
+CMD ["python", "app.py"]

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2022 hysts
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -0,0 +1,12 @@
+---
+title: Tune-A-Video Inference
+emoji: 🐠
+colorFrom: red
+colorTo: purple
+sdk: docker
+pinned: false
+license: mit
+duplicated_from: Tune-A-Video-library/Tune-A-Video-inference
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Tune-A-Video

@@ -0,0 +1 @@
+Subproject commit b2c8c3eeac0df5c5d9eccc4dd2153e17b83c638c

app.py

@@ -0,0 +1,217 @@
+#!/usr/bin/env python
+
+from __future__ import annotations
+
+import os
+
+import gradio as gr
+
+from inference import InferencePipeline
+
+
+class InferenceUtil:
+    def __init__(self, hf_token: str | None):
+        self.hf_token = hf_token
+
+    def load_model_info(self, model_id: str) -> tuple[str, str]:
+        try:
+            card = InferencePipeline.get_model_card(model_id, self.hf_token)
+        except Exception:
+            return '', ''
+        base_model = getattr(card.data, 'base_model', '')
+        training_prompt = getattr(card.data, 'training_prompt', '')
+        return base_model, training_prompt
+
+
+TITLE = '# [Tune-A-Video](https://tuneavideo.github.io/)'
+HF_TOKEN = os.getenv('HF_TOKEN')
+pipe = InferencePipeline(HF_TOKEN)
+app = InferenceUtil(HF_TOKEN)
+
+with gr.Blocks(css='style.css') as demo:
+    gr.Markdown(TITLE)
+
+    with gr.Row():
+        with gr.Column():
+            with gr.Box():
+                model_id = gr.Dropdown(
+                    label='Model ID',
+                    choices=[
+                        'Tune-A-Video-library/a-man-is-surfing',
+                        'Tune-A-Video-library/mo-di-bear-guitar',
+                        'Tune-A-Video-library/redshift-man-skiing',
+                    ],
+                    value='Tune-A-Video-library/a-man-is-surfing')
+                with gr.Accordion(
+                        label=
+                        'Model info (Base model and prompt used for training)',
+                        open=False):
+                    with gr.Row():
+                        base_model_used_for_training = gr.Text(
+                            label='Base model', interactive=False)
+                        prompt_used_for_training = gr.Text(
+                            label='Training prompt', interactive=False)
+            prompt = gr.Textbox(label='Prompt',
+                                max_lines=1,
+                                placeholder='Example: "A panda is surfing"')
+            video_length = gr.Slider(label='Video length',
+                                     minimum=4,
+                                     maximum=12,
+                                     step=1,
+                                     value=8)
+            fps = gr.Slider(label='FPS',
+                            minimum=1,
+                            maximum=12,
+                            step=1,
+                            value=1)
+            seed = gr.Slider(label='Seed',
+                             minimum=0,
+                             maximum=100000,
+                             step=1,
+                             value=0)
+            with gr.Accordion('Other Parameters', open=False):
+                num_steps = gr.Slider(label='Number of Steps',
+                                      minimum=0,
+                                      maximum=100,
+                                      step=1,
+                                      value=50)
+                guidance_scale = gr.Slider(label='CFG Scale',
+                                           minimum=0,
+                                           maximum=50,
+                                           step=0.1,
+                                           value=7.5)
+
+            run_button = gr.Button('Generate')
+
+            gr.Markdown('''
+            - It takes a few minutes to download model first.
+            - Expected time to generate an 8-frame video: 70 seconds with T4, 24 seconds with A10G, (10 seconds with A100)
+            ''')
+        with gr.Column():
+            result = gr.Video(label='Result')
+    with gr.Row():
+        examples = [
+            [
+                'Tune-A-Video-library/a-man-is-surfing',
+                'A panda is surfing.',
+                8,
+                1,
+                3,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/a-man-is-surfing',
+                'A racoon is surfing, cartoon style.',
+                8,
+                1,
+                3,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/mo-di-bear-guitar',
+                'a handsome prince is playing guitar, modern disney style.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/mo-di-bear-guitar',
+                'a magical princess is playing guitar, modern disney style.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/mo-di-bear-guitar',
+                'a rabbit is playing guitar, modern disney style.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/mo-di-bear-guitar',
+                'a baby is playing guitar, modern disney style.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/redshift-man-skiing',
+                '(redshift style) spider man is skiing.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/redshift-man-skiing',
+                '(redshift style) black widow is skiing.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/redshift-man-skiing',
+                '(redshift style) batman is skiing.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+            [
+                'Tune-A-Video-library/redshift-man-skiing',
+                '(redshift style) hulk is skiing.',
+                8,
+                1,
+                123,
+                50,
+                7.5,
+            ],
+        ]
+        gr.Examples(examples=examples,
+                    inputs=[
+                        model_id,
+                        prompt,
+                        video_length,
+                        fps,
+                        seed,
+                        num_steps,
+                        guidance_scale,
+                    ],
+                    outputs=result,
+                    fn=pipe.run,
+                    cache_examples=os.getenv('SYSTEM') == 'spaces')
+
+    model_id.change(fn=app.load_model_info,
+                    inputs=model_id,
+                    outputs=[
+                        base_model_used_for_training,
+                        prompt_used_for_training,
+                    ])
+    inputs = [
+        model_id,
+        prompt,
+        video_length,
+        fps,
+        seed,
+        num_steps,
+        guidance_scale,
+    ]
+    prompt.submit(fn=pipe.run, inputs=inputs, outputs=result)
+    run_button.click(fn=pipe.run, inputs=inputs, outputs=result)
+
+demo.queue().launch()

inference.py

@@ -0,0 +1,109 @@
+from __future__ import annotations
+
+import gc
+import pathlib
+import sys
+import tempfile
+
+import gradio as gr
+import imageio
+import PIL.Image
+import torch
+from diffusers.utils.import_utils import is_xformers_available
+from einops import rearrange
+from huggingface_hub import ModelCard
+
+sys.path.append('Tune-A-Video')
+
+from tuneavideo.models.unet import UNet3DConditionModel
+from tuneavideo.pipelines.pipeline_tuneavideo import TuneAVideoPipeline
+
+
+class InferencePipeline:
+    def __init__(self, hf_token: str | None = None):
+        self.hf_token = hf_token
+        self.pipe = None
+        self.device = torch.device(
+            'cuda:0' if torch.cuda.is_available() else 'cpu')
+        self.model_id = None
+
+    def clear(self) -> None:
+        self.model_id = None
+        del self.pipe
+        self.pipe = None
+        torch.cuda.empty_cache()
+        gc.collect()
+
+    @staticmethod
+    def check_if_model_is_local(model_id: str) -> bool:
+        return pathlib.Path(model_id).exists()
+
+    @staticmethod
+    def get_model_card(model_id: str,
+                       hf_token: str | None = None) -> ModelCard:
+        if InferencePipeline.check_if_model_is_local(model_id):
+            card_path = (pathlib.Path(model_id) / 'README.md').as_posix()
+        else:
+            card_path = model_id
+        return ModelCard.load(card_path, token=hf_token)
+
+    @staticmethod
+    def get_base_model_info(model_id: str, hf_token: str | None = None) -> str:
+        card = InferencePipeline.get_model_card(model_id, hf_token)
+        return card.data.base_model
+
+    def load_pipe(self, model_id: str) -> None:
+        if model_id == self.model_id:
+            return
+        base_model_id = self.get_base_model_info(model_id, self.hf_token)
+        unet = UNet3DConditionModel.from_pretrained(
+            model_id,
+            subfolder='unet',
+            torch_dtype=torch.float16,
+            use_auth_token=self.hf_token)
+        pipe = TuneAVideoPipeline.from_pretrained(base_model_id,
+                                                  unet=unet,
+                                                  torch_dtype=torch.float16,
+                                                  use_auth_token=self.hf_token)
+        pipe = pipe.to(self.device)
+        if is_xformers_available():
+            pipe.unet.enable_xformers_memory_efficient_attention()
+        self.pipe = pipe
+        self.model_id = model_id  # type: ignore
+
+    def run(
+        self,
+        model_id: str,
+        prompt: str,
+        video_length: int,
+        fps: int,
+        seed: int,
+        n_steps: int,
+        guidance_scale: float,
+    ) -> PIL.Image.Image:
+        if not torch.cuda.is_available():
+            raise gr.Error('CUDA is not available.')
+
+        self.load_pipe(model_id)
+
+        generator = torch.Generator(device=self.device).manual_seed(seed)
+        out = self.pipe(
+            prompt,
+            video_length=video_length,
+            width=512,
+            height=512,
+            num_inference_steps=n_steps,
+            guidance_scale=guidance_scale,
+            generator=generator,
+        )  # type: ignore
+
+        frames = rearrange(out.videos[0], 'c t h w -> t h w c')
+        frames = (frames * 255).to(torch.uint8).numpy()
+
+        out_file = tempfile.NamedTemporaryFile(suffix='.mp4', delete=False)
+        writer = imageio.get_writer(out_file.name, fps=fps)
+        for frame in frames:
+            writer.append_data(frame)
+        writer.close()
+
+        return out_file.name

packages.txt

@@ -0,0 +1 @@
+ffmpeg

patch

@@ -0,0 +1,15 @@
+diff --git a/train_tuneavideo.py b/train_tuneavideo.py
+index 66d51b2..86b2a5d 100644
+--- a/train_tuneavideo.py
++++ b/train_tuneavideo.py
+@@ -94,8 +94,8 @@ def main(
+ 
+     # Handle the output folder creation
+     if accelerator.is_main_process:
+-        now = datetime.datetime.now().strftime("%Y-%m-%dT%H-%M-%S")
+-        output_dir = os.path.join(output_dir, now)
++        #now = datetime.datetime.now().strftime("%Y-%m-%dT%H-%M-%S")
++        #output_dir = os.path.join(output_dir, now)
+         os.makedirs(output_dir, exist_ok=True)
+         OmegaConf.save(config, os.path.join(output_dir, 'config.yaml'))
+ 

requirements.txt

@@ -0,0 +1,19 @@
+accelerate==0.15.0
+bitsandbytes==0.35.4
+decord==0.6.0
+diffusers[torch]==0.11.1
+einops==0.6.0
+ftfy==6.1.1
+gradio==3.18.0
+huggingface-hub==0.12.0
+imageio==2.25.0
+imageio-ffmpeg==0.4.8
+omegaconf==2.3.0
+Pillow==9.4.0
+python-slugify==7.0.0
+tensorboard==2.11.2
+torch==1.13.1
+torchvision==0.14.1
+transformers==4.26.0
+triton==2.0.0.dev20221202
+xformers==0.0.16

style.css

@@ -0,0 +1,3 @@
+h1 {
+  text-align: center;
+}