Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,26 @@ saved_model/**/* 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
+docs/example_crop.gif filter=lfs diff=lfs merge=lfs -text
+docs/example_crop_still.gif filter=lfs diff=lfs merge=lfs -text
+docs/example_full.gif filter=lfs diff=lfs merge=lfs -text
+docs/example_full_enhanced.gif filter=lfs diff=lfs merge=lfs -text
+docs/free_view_result.gif filter=lfs diff=lfs merge=lfs -text
+docs/resize_good.gif filter=lfs diff=lfs merge=lfs -text
+docs/resize_no.gif filter=lfs diff=lfs merge=lfs -text
+docs/using_ref_video.gif filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/chinese_news.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/deyu.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/eluosi.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/fayu.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/imagine.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/japanese.wav filter=lfs diff=lfs merge=lfs -text
+examples/ref_video/WDA_AlexandriaOcasioCortez_000.mp4 filter=lfs diff=lfs merge=lfs -text
+examples/ref_video/WDA_KatieHill_000.mp4 filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_16.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_17.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_3.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_4.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_5.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_8.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_9.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,174 @@
+# 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/
+
+examples/results/*
+gfpgan/*
+checkpoints/*
+assets/*
+results/*
+Dockerfile
+start_docker.sh
+start.sh
+
+checkpoints
+
+# Mac
+.DS_Store

.ipynb_checkpoints/Untitled-checkpoint.ipynb

@@ -0,0 +1,6 @@
+{
+ "cells": [],
+ "metadata": {},
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2023 Tencent AI Lab
+
+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

@@ -1,13 +1,268 @@
----
-title: SadTalker
-emoji: 🏃
-colorFrom: gray
-colorTo: blue
-sdk: gradio
-sdk_version: 3.37.0
-app_file: app.py
-pinned: false
-license: mit
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+<div align="center">
+
+<img src='https://user-images.githubusercontent.com/4397546/229094115-862c747e-7397-4b54-ba4a-bd368bfe2e0f.png' width='500px'/>
+
+
+<!--<h2> 😭 SadTalker： <span style="font-size:12px">Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation </span> </h2> -->
+
+  <a href='https://arxiv.org/abs/2211.12194'><img src='https://img.shields.io/badge/ArXiv-PDF-red'></a> &nbsp; <a href='https://sadtalker.github.io'><img src='https://img.shields.io/badge/Project-Page-Green'></a> &nbsp; [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Winfredy/SadTalker/blob/main/quick_demo.ipynb) &nbsp; [![Hugging Face Spaces](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces/vinthony/SadTalker) &nbsp; [![sd webui-colab](https://img.shields.io/badge/Automatic1111-Colab-green)](https://colab.research.google.com/github/camenduru/stable-diffusion-webui-colab/blob/main/video/stable/stable_diffusion_1_5_video_webui_colab.ipynb) &nbsp; [![Replicate](https://replicate.com/cjwbw/sadtalker/badge)](https://replicate.com/cjwbw/sadtalker) 
+
+<div>
+    <a target='_blank'>Wenxuan Zhang <sup>*,1,2</sup> </a>&emsp;
+    <a href='https://vinthony.github.io/' target='_blank'>Xiaodong Cun <sup>*,2</a>&emsp;
+    <a href='https://xuanwangvc.github.io/' target='_blank'>Xuan Wang <sup>3</sup></a>&emsp;
+    <a href='https://yzhang2016.github.io/' target='_blank'>Yong Zhang <sup>2</sup></a>&emsp;
+    <a href='https://xishen0220.github.io/' target='_blank'>Xi Shen <sup>2</sup></a>&emsp; </br>
+    <a href='https://yuguo-xjtu.github.io/' target='_blank'>Yu Guo<sup>1</sup> </a>&emsp;
+    <a href='https://scholar.google.com/citations?hl=zh-CN&user=4oXBp9UAAAAJ' target='_blank'>Ying Shan <sup>2</sup> </a>&emsp;
+    <a target='_blank'>Fei Wang <sup>1</sup> </a>&emsp;
+</div>
+<br>
+<div>
+    <sup>1</sup> Xi'an Jiaotong University &emsp; <sup>2</sup> Tencent AI Lab &emsp; <sup>3</sup> Ant Group &emsp; 
+</div>
+<br>
+<i><strong><a href='https://arxiv.org/abs/2211.12194' target='_blank'>CVPR 2023</a></strong></i>
+<br>
+<br>
+
+
+![sadtalker](https://user-images.githubusercontent.com/4397546/222490039-b1f6156b-bf00-405b-9fda-0c9a9156f991.gif)
+
+<b>TL;DR: &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; single portrait image 🙎‍♂️  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;+  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; audio 🎤  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; talking head video 🎞.</b>
+
+<br>
+
+</div>
+
+
+
+## 🔥 Highlight
+
+- 🔥 The extension of the [stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui) is online. Checkout more details [here](docs/webui_extension.md).
+
+https://user-images.githubusercontent.com/4397546/231495639-5d4bb925-ea64-4a36-a519-6389917dac29.mp4
+
+- 🔥 `full image mode` is online! checkout [here](https://github.com/Winfredy/SadTalker#full-bodyimage-generation) for more details.
+
+| still+enhancer in v0.0.1                 | still + enhancer   in v0.0.2       |   [input image @bagbag1815](https://twitter.com/bagbag1815/status/1642754319094108161) |
+|:--------------------: |:--------------------: | :----: |
+| <video  src="https://user-images.githubusercontent.com/48216707/229484996-5d7be64f-2553-4c9e-a452-c5cf0b8ebafe.mp4" type="video/mp4"> </video> | <video  src="https://user-images.githubusercontent.com/4397546/230717873-355b7bf3-d3de-49f9-a439-9220e623fce7.mp4" type="video/mp4"> </video>  | <img src='./examples/source_image/full_body_2.png' width='380'> 
+
+- 🔥 Several new mode, eg, `still mode`, `reference mode`, `resize mode` are online for better and custom applications.
+
+- 🔥 Happy to  see more community demos at [bilibili](https://search.bilibili.com/all?keyword=sadtalker&from_source=webtop_search&spm_id_from=333.1007&search_source=3
+), [Youtube](https://www.youtube.com/results?search_query=sadtalker&sp=CAM%253D) and [twitter #sadtalker](https://twitter.com/search?q=%23sadtalker&src=typed_query).
+
+## 📋 Changelog (Previous changelog can be founded [here](docs/changlelog.md))
+
+- __[2023.06.12]__: add more new features in WEBUI extension, see the discussion [here](https://github.com/OpenTalker/SadTalker/discussions/386).
+
+- __[2023.06.05]__: release a new 512 beta face model. Fixed some bugs and improve the performance.
+
+- __[2023.04.15]__: Adding automatic1111 colab by @camenduru, thanks for this awesome colab: [![sd webui-colab](https://img.shields.io/badge/Automatic1111-Colab-green)](https://colab.research.google.com/github/camenduru/stable-diffusion-webui-colab/blob/main/video/stable/stable_diffusion_1_5_video_webui_colab.ipynb).
+
+- __[2023.04.12]__: adding a more detailed sd-webui installation document, fixed reinstallation problem.
+
+- __[2023.04.12]__: Fixed the sd-webui safe issues becasue of the 3rd packages, optimize the output path in `sd-webui-extension`.
+
+- __[2023.04.08]__: ❗️❗️❗️ In v0.0.2, we add a logo watermark to the generated video to prevent abusing since it is very realistic.
+
+- __[2023.04.08]__: v0.0.2, full image animation, adding baidu driver for download checkpoints. Optimizing the logic about enhancer.
+
+
+## 🚧 TODO: See the Discussion https://github.com/OpenTalker/SadTalker/issues/280
+
+## If you have any problem, please view our [FAQ](docs/FAQ.md) before opening an issue.
+
+
+
+## ⚙️ 1. Installation.
+
+Tutorials from communities: [中文windows教程](https://www.bilibili.com/video/BV1Dc411W7V6/) | [日本語コース](https://br-d.fanbox.cc/posts/5685086?utm_campaign=manage_post_page&utm_medium=share&utm_source=twitter) 
+
+### Linux:
+
+1. Installing [anaconda](https://www.anaconda.com/), python and git.
+
+2. Creating the env and install the requirements.
+  ```bash
+  git clone https://github.com/Winfredy/SadTalker.git
+
+  cd SadTalker 
+
+  conda create -n sadtalker python=3.8
+
+  conda activate sadtalker
+
+  pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113
+
+  conda install ffmpeg
+
+  pip install -r requirements.txt
+
+  ### tts is optional for gradio demo. 
+  ### pip install TTS
+
+  ```  
+### Windows ([中文windows教程](https://www.bilibili.com/video/BV1Dc411W7V6/)):
+
+1. Install [Python 3.10.6](https://www.python.org/downloads/windows/), checking "Add Python to PATH".
+2. Install [git](https://git-scm.com/download/win) manually (OR `scoop install git` via [scoop](https://scoop.sh/)).
+3. Install `ffmpeg`, following [this instruction](https://www.wikihow.com/Install-FFmpeg-on-Windows) (OR using `scoop install ffmpeg` via [scoop](https://scoop.sh/)).
+4. Download our SadTalker repository, for example by running `git clone https://github.com/Winfredy/SadTalker.git`.
+5. Download the `checkpoint` and `gfpgan` [below↓](https://github.com/Winfredy/SadTalker#-2-download-trained-models).
+5. Run `start.bat` from Windows Explorer as normal, non-administrator, user, a gradio WebUI demo will be started.
+
+### Macbook:
+
+More tips about installnation on Macbook and the Docker file can be founded [here](docs/install.md)
+
+## 📥 2. Download Trained Models.
+
+You can run the following script to put all the models in the right place.
+
+```bash
+bash scripts/download_models.sh
+```
+
+Other alternatives:
+> we also provide an offline patch (`gfpgan/`), thus, no model will be downloaded when generating.
+
+**Google Driver**: download our pre-trained model from [ this link (main checkpoints)](https://drive.google.com/file/d/1gwWh45pF7aelNP_P78uDJL8Sycep-K7j/view?usp=sharing) and [ gfpgan (offline patch)](https://drive.google.com/file/d/19AIBsmfcHW6BRJmeqSFlG5fL445Xmsyi?usp=sharing)
+
+**Github Release Page**: download all the files from the [lastest github release page](https://github.com/Winfredy/SadTalker/releases), and then, put it in ./checkpoints.
+
+**百度云盘**: we provided the downloaded model in [checkpoints,  提取码: sadt.](https://pan.baidu.com/s/1P4fRgk9gaSutZnn8YW034Q?pwd=sadt) And [gfpgan,  提取码: sadt.](https://pan.baidu.com/s/1kb1BCPaLOWX1JJb9Czbn6w?pwd=sadt)
+
+
+
+<details><summary>Model Details</summary>
+
+
+Model explains:
+
+##### New version 
+| Model | Description
+| :--- | :----------
+|checkpoints/mapping_00229-model.pth.tar | Pre-trained MappingNet in Sadtalker.
+|checkpoints/mapping_00109-model.pth.tar | Pre-trained MappingNet in Sadtalker.
+|checkpoints/SadTalker_V0.0.2_256.safetensors | packaged sadtalker checkpoints of old version, 256 face render).
+|checkpoints/SadTalker_V0.0.2_512.safetensors | packaged sadtalker checkpoints of old version, 512 face render).
+|gfpgan/weights | Face detection and enhanced models used in `facexlib` and `gfpgan`.
+  
+  
+##### Old version
+| Model | Description
+| :--- | :----------
+|checkpoints/auido2exp_00300-model.pth | Pre-trained ExpNet in Sadtalker.
+|checkpoints/auido2pose_00140-model.pth | Pre-trained PoseVAE in Sadtalker.
+|checkpoints/mapping_00229-model.pth.tar | Pre-trained MappingNet in Sadtalker.
+|checkpoints/mapping_00109-model.pth.tar | Pre-trained MappingNet in Sadtalker.
+|checkpoints/facevid2vid_00189-model.pth.tar | Pre-trained face-vid2vid model from [the reappearance of face-vid2vid](https://github.com/zhanglonghao1992/One-Shot_Free-View_Neural_Talking_Head_Synthesis).
+|checkpoints/epoch_20.pth | Pre-trained 3DMM extractor in [Deep3DFaceReconstruction](https://github.com/microsoft/Deep3DFaceReconstruction).
+|checkpoints/wav2lip.pth | Highly accurate lip-sync model in [Wav2lip](https://github.com/Rudrabha/Wav2Lip).
+|checkpoints/shape_predictor_68_face_landmarks.dat | Face landmark model used in [dilb](http://dlib.net/). 
+|checkpoints/BFM | 3DMM library file.  
+|checkpoints/hub | Face detection models used in [face alignment](https://github.com/1adrianb/face-alignment).
+|gfpgan/weights | Face detection and enhanced models used in `facexlib` and `gfpgan`.
+
+The final folder will be shown as:
+
+<img width="331" alt="image" src="https://user-images.githubusercontent.com/4397546/232511411-4ca75cbf-a434-48c5-9ae0-9009e8316484.png">
+
+
+</details>
+
+## 🔮 3. Quick Start ([Best Practice](docs/best_practice.md)).
+
+### WebUI Demos:
+
+**Online**: [Huggingface](https://huggingface.co/spaces/vinthony/SadTalker) | [SDWebUI-Colab](https://colab.research.google.com/github/camenduru/stable-diffusion-webui-colab/blob/main/video/stable/stable_diffusion_1_5_video_webui_colab.ipynb) | [Colab](https://colab.research.google.com/github/Winfredy/SadTalker/blob/main/quick_demo.ipynb)
+
+**Local Autiomatic1111 stable-diffusion webui extension**: please refer to [Autiomatic1111 stable-diffusion webui docs](docs/webui_extension.md).
+
+**Local gradio demo(highly recommanded!)**: Similar to our [hugging-face demo](https://huggingface.co/spaces/vinthony/SadTalker) can be run by:
+
+```bash
+## you need manually install TTS(https://github.com/coqui-ai/TTS) via `pip install tts` in advanced.
+python app.py
+```
+
+**Local gradio demo(highly recommanded!)**: 
+
+- windows: just double click `webui.bat`, the requirements will be installed automatically.
+- Linux/Mac OS: run `bash webui.sh` to start the webui.
+
+
+### Manually usages:
+
+##### Animating a portrait image from default config:
+```bash
+python inference.py --driven_audio <audio.wav> \
+                    --source_image <video.mp4 or picture.png> \
+                    --enhancer gfpgan 
+```
+The results will be saved in `results/$SOME_TIMESTAMP/*.mp4`.
+
+##### Full body/image Generation:
+
+Using `--still` to generate a natural full body video. You can add `enhancer` to improve the quality of the generated video. 
+
+```bash
+python inference.py --driven_audio <audio.wav> \
+                    --source_image <video.mp4 or picture.png> \
+                    --result_dir <a file to store results> \
+                    --still \
+                    --preprocess full \
+                    --enhancer gfpgan 
+```
+
+More examples and configuration and tips can be founded in the [ >>> best practice documents <<<](docs/best_practice.md).
+
+## 🛎 Citation
+
+If you find our work useful in your research, please consider citing:
+
+```bibtex
+@article{zhang2022sadtalker,
+  title={SadTalker: Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation},
+  author={Zhang, Wenxuan and Cun, Xiaodong and Wang, Xuan and Zhang, Yong and Shen, Xi and Guo, Yu and Shan, Ying and Wang, Fei},
+  journal={arXiv preprint arXiv:2211.12194},
+  year={2022}
+}
+```
+
+
+
+## 💗 Acknowledgements
+
+Facerender code borrows heavily from [zhanglonghao's reproduction of face-vid2vid](https://github.com/zhanglonghao1992/One-Shot_Free-View_Neural_Talking_Head_Synthesis) and [PIRender](https://github.com/RenYurui/PIRender). We thank the authors for sharing their wonderful code. In training process, We also use the model from [Deep3DFaceReconstruction](https://github.com/microsoft/Deep3DFaceReconstruction) and [Wav2lip](https://github.com/Rudrabha/Wav2Lip). We thank for their wonderful work.
+
+See also these wonderful 3rd libraries we use:
+
+- **Face Utils**: https://github.com/xinntao/facexlib
+- **Face Enhancement**: https://github.com/TencentARC/GFPGAN
+- **Image/Video Enhancement**:https://github.com/xinntao/Real-ESRGAN
+
+## 🥂 Extensions:
+
+- [SadTalker-Video-Lip-Sync](https://github.com/Zz-ww/SadTalker-Video-Lip-Sync) from [@Zz-ww](https://github.com/Zz-ww): SadTalker for Video Lip Editing
+
+## 🥂 Related Works
+- [StyleHEAT: One-Shot High-Resolution Editable Talking Face Generation via Pre-trained StyleGAN (ECCV 2022)](https://github.com/FeiiYin/StyleHEAT)
+- [CodeTalker: Speech-Driven 3D Facial Animation with Discrete Motion Prior (CVPR 2023)](https://github.com/Doubiiu/CodeTalker)
+- [VideoReTalking: Audio-based Lip Synchronization for Talking Head Video Editing In the Wild (SIGGRAPH Asia 2022)](https://github.com/vinthony/video-retalking)
+- [DPE: Disentanglement of Pose and Expression for General Video Portrait Editing (CVPR 2023)](https://github.com/Carlyx/DPE)
+- [3D GAN Inversion with Facial Symmetry Prior (CVPR 2023)](https://github.com/FeiiYin/SPI/)
+- [T2M-GPT: Generating Human Motion from Textual Descriptions with Discrete Representations (CVPR 2023)](https://github.com/Mael-zys/T2M-GPT)
+
+## 📢 Disclaimer
+
+This is not an official product of Tencent. This repository can only be used for personal/research/non-commercial purposes.
+
+LOGO: color and font suggestion: [ChatGPT](ai.com), logo font：[Montserrat Alternates
+](https://fonts.google.com/specimen/Montserrat+Alternates?preview.text=SadTalker&preview.text_type=custom&query=mont).
+
+All the copyright of the demo images and audio are from communities users or the geneartion from stable diffusion. Free free to contact us if you feel uncomfortable.
+

Untitled.ipynb

@@ -0,0 +1,533 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "c69a901b-8c3f-4188-97bd-5594c4496ec5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from huggingface_hub import HfApi\n",
+    "api = HfApi()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "8b249adc-ccd0-4145-86ce-64509ad276cf",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "faf7b41b81e54705bae8921f2a86e9fd",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "VBox(children=(HTML(value='<center> <img\\nsrc=https://huggingface.co/front/assets/huggingface_logo-noborder.sv…"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "from huggingface_hub import login\n",
+    "login()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "41815b66-b12b-45da-9a5e-92c527fe2b4a",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "6bb4c00dda664acd9f70e029e15dbd57",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "SadTalker_V0.0.2_512.safetensors:   0%|          | 0.00/725M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "b9a78d85c20642ac86da6b6d0ab97031",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "mapping_00109-model.pth.tar:   0%|          | 0.00/156M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "d396b1372ea04ae79dd82677c1c33352",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Upload 31 LFS files:   0%|          | 0/31 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "7b924c6f560b40b78152a87f5c388d6b",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "mapping_00229-model.pth.tar:   0%|          | 0.00/156M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "4245c014a4d04d6daacd47d7d81c6249",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "SadTalker_V0.0.2_256.safetensors:   0%|          | 0.00/725M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "d7f2b0a6b3ed4d5e895da21237f42577",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "example_crop.gif:   0%|          | 0.00/1.55M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "24ea098c821c4fc79ffa38b2fe5a1575",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "example_crop_still.gif:   0%|          | 0.00/1.25M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "b828fea9017846748cbe423f1b5fada7",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "example_full.gif:   0%|          | 0.00/1.46M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "a0b664fbf3f8465d9b286290e5f14d8f",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "example_full_enhanced.gif:   0%|          | 0.00/5.78M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "e7483527580845429ce33f4f987e7ddb",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "free_view_result.gif:   0%|          | 0.00/5.61M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "7662c74dc37e4e8a9e756da854c7267a",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "resize_good.gif:   0%|          | 0.00/1.73M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0e6e3f1dbb26419c895a7fed7a7f31ac",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "resize_no.gif:   0%|          | 0.00/2.14M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "bdf7d2a9e5784350af4be3d57cc72fa4",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "using_ref_video.gif:   0%|          | 0.00/8.11M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "3348de1636bc4d85b2674eeed42d03d9",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "chinese_news.wav:   0%|          | 0.00/1.54M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "40d1dc6faf634304a13e18feab5816af",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "deyu.wav:   0%|          | 0.00/2.69M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "2c91050e17e34841ae74d4a9b414eaf7",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "eluosi.wav:   0%|          | 0.00/1.79M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "35ed59910dd04972b65bd0684dd497c1",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "fayu.wav:   0%|          | 0.00/1.94M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "6cc44214a3224b9fa952d8fdd69b262d",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "imagine.wav:   0%|          | 0.00/1.62M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "21b4c8e2efbe4205b961d849c52c0c66",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "japanese.wav:   0%|          | 0.00/2.62M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "8df2d5959a394789b9defb2d8dea9680",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "WDA_AlexandriaOcasioCortez_000.mp4:   0%|          | 0.00/2.26M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "bee85421a93345f88e1ba92bdc627a92",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "WDA_KatieHill_000.mp4:   0%|          | 0.00/3.55M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "1dfceecdadf2443bb935b58c06116ac3",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_16.png:   0%|          | 0.00/1.48M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "7b054b4faa49444786ac49f10bc2ca27",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_17.png:   0%|          | 0.00/2.09M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "78db6e2bdf3f4e8289cda18189d2f9bd",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_3.png:   0%|          | 0.00/1.35M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ef5a284a239548c4964ea3a924975cbc",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_4.png:   0%|          | 0.00/3.63M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "fbdf097df06247c98da1a3f7b4322c92",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_5.png:   0%|          | 0.00/1.23M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "1099a9330eae49178d53d35827f9fa05",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_8.png:   0%|          | 0.00/3.12M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "f53f1a45946f4a8db6c7a48f525b8f7c",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "art_9.png:   0%|          | 0.00/1.26M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "de038d6f5aa84c63b2335af75c82e337",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "GFPGANv1.4.pth:   0%|          | 0.00/349M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "69daa05b13a745fa8a27e6447282fdff",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "alignment_WFLW_4HG.pth:   0%|          | 0.00/194M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ee0b32bbe61647359a6e8324a092b95e",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "detection_Resnet50_Final.pth:   0%|          | 0.00/109M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "709b3e709e704234826dd1050f260688",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "parsing_parsenet.pth:   0%|          | 0.00/85.3M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "api.upload_folder(\n",
+    "    folder_path=\"/workspace/SadTalker\",\n",
+    "    repo_id=\"kevinwang676/SadTalker\",\n",
+    "    repo_type=\"space\",\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "c15c0fad-c4ac-4fd5-b886-89d2f97bc285",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.11"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

app_sadtalker.py

@@ -0,0 +1,111 @@
+import os, sys
+import gradio as gr
+from src.gradio_demo import SadTalker  
+
+
+try:
+    import webui  # in webui
+    in_webui = True
+except:
+    in_webui = False
+
+
+def toggle_audio_file(choice):
+    if choice == False:
+        return gr.update(visible=True), gr.update(visible=False)
+    else:
+        return gr.update(visible=False), gr.update(visible=True)
+    
+def ref_video_fn(path_of_ref_video):
+    if path_of_ref_video is not None:
+        return gr.update(value=True)
+    else:
+        return gr.update(value=False)
+
+def sadtalker_demo(checkpoint_path='checkpoints', config_path='src/config', warpfn=None):
+
+    sad_talker = SadTalker(checkpoint_path, config_path, lazy_load=True)
+
+    with gr.Blocks(analytics_enabled=False) as sadtalker_interface:
+        gr.Markdown("<div align='center'> <h2> 😭 SadTalker: Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation (CVPR 2023) </span> </h2> \
+                    <a style='font-size:18px;color: #efefef' href='https://arxiv.org/abs/2211.12194'>Arxiv</a> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; \
+                    <a style='font-size:18px;color: #efefef' href='https://sadtalker.github.io'>Homepage</a>  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp; \
+                     <a style='font-size:18px;color: #efefef' href='https://github.com/Winfredy/SadTalker'> Github </div>")
+        
+        with gr.Row().style(equal_height=False):
+            with gr.Column(variant='panel'):
+                with gr.Tabs(elem_id="sadtalker_source_image"):
+                    with gr.TabItem('Upload image'):
+                        with gr.Row():
+                            source_image = gr.Image(label="Source image", source="upload", type="filepath", elem_id="img2img_image").style(width=512)
+
+                with gr.Tabs(elem_id="sadtalker_driven_audio"):
+                    with gr.TabItem('Upload OR TTS'):
+                        with gr.Column(variant='panel'):
+                            driven_audio = gr.Audio(label="Input audio", source="upload", type="filepath")
+
+                        if sys.platform != 'win32' and not in_webui: 
+                            from src.utils.text2speech import TTSTalker
+                            tts_talker = TTSTalker()
+                            with gr.Column(variant='panel'):
+                                input_text = gr.Textbox(label="Generating audio from text", lines=5, placeholder="please enter some text here, we genreate the audio from text using @Coqui.ai TTS.")
+                                tts = gr.Button('Generate audio',elem_id="sadtalker_audio_generate", variant='primary')
+                                tts.click(fn=tts_talker.test, inputs=[input_text], outputs=[driven_audio])
+                            
+            with gr.Column(variant='panel'): 
+                with gr.Tabs(elem_id="sadtalker_checkbox"):
+                    with gr.TabItem('Settings'):
+                        gr.Markdown("need help? please visit our [best practice page](https://github.com/OpenTalker/SadTalker/blob/main/docs/best_practice.md) for more detials")
+                        with gr.Column(variant='panel'):
+                            # width = gr.Slider(minimum=64, elem_id="img2img_width", maximum=2048, step=8, label="Manually Crop Width", value=512) # img2img_width
+                            # height = gr.Slider(minimum=64, elem_id="img2img_height", maximum=2048, step=8, label="Manually Crop Height", value=512) # img2img_width
+                            pose_style = gr.Slider(minimum=0, maximum=46, step=1, label="Pose style", value=0) # 
+                            size_of_image = gr.Radio([256, 512], value=256, label='face model resolution', info="use 256/512 model?") # 
+                            preprocess_type = gr.Radio(['crop', 'resize','full', 'extcrop', 'extfull'], value='crop', label='preprocess', info="How to handle input image?")
+                            is_still_mode = gr.Checkbox(label="Still Mode (fewer hand motion, works with preprocess `full`)")
+                            batch_size = gr.Slider(label="batch size in generation", step=1, maximum=10, value=2)
+                            enhancer = gr.Checkbox(label="GFPGAN as Face enhancer")
+                            submit = gr.Button('Generate', elem_id="sadtalker_generate", variant='primary')
+                            
+                with gr.Tabs(elem_id="sadtalker_genearted"):
+                        gen_video = gr.Video(label="Generated video", format="mp4").style(width=256)
+
+        if warpfn:
+            submit.click(
+                        fn=warpfn(sad_talker.test), 
+                        inputs=[source_image,
+                                driven_audio,
+                                preprocess_type,
+                                is_still_mode,
+                                enhancer,
+                                batch_size,                            
+                                size_of_image,
+                                pose_style
+                                ], 
+                        outputs=[gen_video]
+                        )
+        else:
+            submit.click(
+                        fn=sad_talker.test, 
+                        inputs=[source_image,
+                                driven_audio,
+                                preprocess_type,
+                                is_still_mode,
+                                enhancer,
+                                batch_size,                            
+                                size_of_image,
+                                pose_style
+                                ], 
+                        outputs=[gen_video]
+                        )
+
+    return sadtalker_interface
+ 
+
+if __name__ == "__main__":
+
+    demo = sadtalker_demo()
+    demo.queue()
+    demo.launch()
+
+

checkpoints/SadTalker_V0.0.2_256.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c211f5d6de003516bf1bbda9f47049a4c9c99133b1ab565c6961e5af16477bff
+size 725066984

checkpoints/SadTalker_V0.0.2_512.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e063f7ff5258240bdb0f7690783a7b1374e6a4a81ce8fa33456f4cd49694340
+size 725066984

checkpoints/mapping_00109-model.pth.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84a8642468a3fcfdd9ab6be955267043116c2bec2284686a5262f1eaf017f64c
+size 155779231

checkpoints/mapping_00229-model.pth.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62a1e06006cc963220f6477438518ed86e9788226c62ae382ddc42fbcefb83f1
+size 155521183

cog.yaml

@@ -0,0 +1,35 @@
+build:
+  gpu: true
+  cuda: "11.3"
+  python_version: "3.8"
+  system_packages:
+    - "ffmpeg"
+    - "libgl1-mesa-glx"
+    - "libglib2.0-0"
+  python_packages:
+    - "torch==1.12.1"
+    - "torchvision==0.13.1"
+    - "torchaudio==0.12.1"
+    - "joblib==1.1.0"
+    - "scikit-image==0.19.3"
+    - "basicsr==1.4.2"
+    - "facexlib==0.3.0"
+    - "resampy==0.3.1"
+    - "pydub==0.25.1"
+    - "scipy==1.10.1"
+    - "kornia==0.6.8"
+    - "face_alignment==1.3.5"
+    - "imageio==2.19.3"
+    - "imageio-ffmpeg==0.4.7"
+    - "librosa==0.9.2" #
+    - "tqdm==4.65.0"
+    - "yacs==0.1.8"
+    - "gfpgan==1.3.8"
+    - "dlib-bin==19.24.1"
+    - "av==10.0.0"
+    - "trimesh==3.9.20"
+  run:
+    - mkdir -p /root/.cache/torch/hub/checkpoints/ && wget --output-document "/root/.cache/torch/hub/checkpoints/s3fd-619a316812.pth" "https://www.adrianbulat.com/downloads/python-fan/s3fd-619a316812.pth"
+    - mkdir -p /root/.cache/torch/hub/checkpoints/ && wget --output-document "/root/.cache/torch/hub/checkpoints/2DFAN4-cd938726ad.zip" "https://www.adrianbulat.com/downloads/python-fan/2DFAN4-cd938726ad.zip"
+
+predict: "predict.py:Predictor"

docs/FAQ.md

@@ -0,0 +1,46 @@
+
+## Frequency Asked Question
+
+**Q: `ffmpeg` is not recognized as an internal or external command**
+
+In Linux, you can install the ffmpeg via `conda install ffmpeg`. Or on Mac OS X, try to install ffmpeg via `brew install ffmpeg`. On windows, make sure you have `ffmpeg` in the `%PATH%` as suggested in [#54](https://github.com/Winfredy/SadTalker/issues/54), then, following [this](https://www.geeksforgeeks.org/how-to-install-ffmpeg-on-windows/) installation to install `ffmpeg`.
+
+**Q: Running Requirments.**
+
+Please refer to the discussion here: https://github.com/Winfredy/SadTalker/issues/124#issuecomment-1508113989
+
+
+**Q: ModuleNotFoundError: No module named 'ai'**
+
+please check the checkpoint's size of the `epoch_20.pth`. (https://github.com/Winfredy/SadTalker/issues/167, https://github.com/Winfredy/SadTalker/issues/113)
+
+**Q: Illegal Hardware Error: Mac M1**
+
+please reinstall the `dlib` by `pip install dlib` individually. (https://github.com/Winfredy/SadTalker/issues/129, https://github.com/Winfredy/SadTalker/issues/109)
+
+
+**Q: FileNotFoundError: [Errno 2] No such file or directory: checkpoints\BFM_Fitting\similarity_Lm3D_all.mat**
+
+Make sure you have downloaded the checkpoints and gfpgan as [here](https://github.com/Winfredy/SadTalker#-2-download-trained-models) and placed them in the right place. 
+
+**Q: RuntimeError: unexpected EOF, expected 237192 more bytes. The file might be corrupted.**
+
+The files are not automatically downloaded. Please update the code and download the gfpgan folders as [here](https://github.com/Winfredy/SadTalker#-2-download-trained-models).
+
+**Q: CUDA out of memory error**
+
+please refer to https://stackoverflow.com/questions/73747731/runtimeerror-cuda-out-of-memory-how-setting-max-split-size-mb
+
+``` 
+# windows
+set PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:128 
+python inference.py ...
+
+# linux
+export PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:128 
+python inference.py ...
+```
+
+**Q: Error while decoding stream #0:0: Invalid data found when processing input [mp3float @ 0000015037628c00] Header missing**
+
+Our method only support wav or mp3 files as input, please make sure the feeded audios are in these formats.

docs/best_practice.md

@@ -0,0 +1,94 @@
+# Best Practice and Tips for configuration
+
+> Our model only works on REAL person's photo or the portrait image similar to REAL person. The anime talking head genreation method will be released in future.
+
+Advanced confiurations for `inference.py`:
+
+| Name        | Configuration | default |   Explaination  | 
+|:------------- |:------------- |:----- | :------------- |
+| Enhance Mode | `--enhancer` | None | Using `gfpgan` or `RestoreFormer` to enhance the generated face via face restoration network 
+| Background Enhancer | `--background_enhancer` | None | Using `realesrgan` to enhance the full video. 
+| Still Mode   | ` --still` | False |  Using the same pose parameters as the original image, fewer head motion.
+| Expressive Mode | `--expression_scale` | 1.0 | a larger value will make the expression motion stronger.
+| save path | `--result_dir` |`./results` | The file will be save in the newer location.
+| preprocess | `--preprocess` | `crop` | Run and produce the results in the croped input image. Other choices: `resize`, where the images will be resized to the specific resolution. `full` Run the full image animation, use with `--still` to get better results.
+| ref Mode (eye) | `--ref_eyeblink` | None | A video path, where we borrow the eyeblink from this reference video to provide more natural eyebrow movement.
+| ref Mode (pose) | `--ref_pose` | None | A video path, where we borrow the pose from the head reference video. 
+| 3D Mode | `--face3dvis` | False | Need additional installation. More details to generate the 3d face can be founded [here](docs/face3d.md). 
+| free-view Mode | `--input_yaw`,<br> `--input_pitch`,<br> `--input_roll` | None | Genearting novel view or free-view 4D talking head from a single image. More details can be founded [here](https://github.com/Winfredy/SadTalker#generating-4d-free-view-talking-examples-from-audio-and-a-single-image).
+
+
+### About `--preprocess`
+
+Our method automatically handle the input images via `crop`, `resize` and `full`.
+
+ In `crop` mode, we only generate the croped image via the facial keypoints and generated the facial anime avator. The animation of both expression and head pose are realistic.
+
+> still mode will stop the eyeblink and head pose movement.
+
+|  [input image @bagbag1815](https://twitter.com/bagbag1815/status/1642754319094108161) | crop | crop w/still |
+|:--------------------: |:--------------------: | :----: |
+| <img src='../examples/source_image/full_body_2.png' width='380'> | ![full_body_2](example_crop.gif) | ![full_body_2](example_crop_still.gif) |
+
+
+ In `resize` mode, we resize the whole images to generate the fully talking head video. Thus, an image similar to the ID photo can be produced. ⚠️ It will produce bad results for full person images.
+
+
+ 
+
+| <img src='../examples/source_image/full_body_2.png' width='380'> |  <img src='../examples/source_image/full4.jpeg' width='380'> |
+|:--------------------: |:--------------------: |
+| ❌ not suitable for resize mode | ✅ good for resize mode |
+| <img src='resize_no.gif'> |  <img src='resize_good.gif' width='380'> |
+
+In `full` mode, our model will automatically process the croped region and paste back to the original image. Remember to use `--still` to keep the original head pose.
+
+| input | `--still` | `--still` & `enhancer` |
+|:--------------------: |:--------------------: | :--:|
+| <img src='../examples/source_image/full_body_2.png' width='380'> |  <img src='./example_full.gif' width='380'> |  <img src='./example_full_enhanced.gif' width='380'> 
+
+
+### About `--enhancer`
+
+For better facial quality, we intergate [gfpgan](https://github.com/TencentARC/GFPGAN) and [real-esrgan](https://github.com/xinntao/Real-ESRGAN) for different purpose. Just adding `--enhancer <gfpgan or RestoreFormer>` or `--background_enhancer <realesrgan>` for the enhancement of the face and the full image.
+
+```bash
+# make sure above packages are available:
+pip install gfpgan
+pip install realesrgan
+```
+
+### About `--face3dvis`
+
+This flag indicate that we can generated the 3d-rendered face and it's 3d facial landmarks. More details can be founded [here](face3d.md).
+
+| Input        | Animated 3d face | 
+|:-------------: | :-------------: |
+|  <img src='../examples/source_image/art_0.png' width='200px'> | <video src="https://user-images.githubusercontent.com/4397546/226856847-5a6a0a4d-a5ec-49e2-9b05-3206db65e8e3.mp4"></video>  | 
+
+> Kindly ensure to activate the audio as the default audio playing is incompatible with GitHub.
+
+
+
+#### reference eye-link mode.
+
+| Input, w/ reference video   ,  reference video    | 
+|:-------------: | 
+|  ![free_view](using_ref_video.gif)| 
+| If the reference video is shorter than the input audio, we will loop the reference video . 
+
+
+
+#### Generating 4D free-view talking examples from audio and a single image
+
+We use `input_yaw`, `input_pitch`, `input_roll` to control head pose. For example, `--input_yaw -20 30 10` means the input head yaw degree changes from -20 to 30 and then changes from 30 to 10.
+```bash
+python inference.py --driven_audio <audio.wav> \
+                    --source_image <video.mp4 or picture.png> \
+                    --result_dir <a file to store results> \
+                    --input_yaw -20 30 10
+```
+
+| Results, Free-view results,  Novel view results  | 
+|:-------------: | 
+|  ![free_view](free_view_result.gif)| 

docs/changlelog.md

@@ -0,0 +1,29 @@
+## changelogs
+
+
+- __[2023.04.06]__: stable-diffiusion webui extension is release.
+
+- __[2023.04.03]__: Enable TTS in huggingface and gradio local demo.
+
+- __[2023.03.30]__: Launch beta version of the full body mode.
+
+- __[2023.03.30]__: Launch new feature: through using reference videos, our algorithm can generate videos with more natural eye blinking and some eyebrow movement.
+
+- __[2023.03.29]__: `resize mode` is online by `python infererence.py --preprocess resize`! Where we can produce a larger crop of the image as discussed in https://github.com/Winfredy/SadTalker/issues/35.
+
+- __[2023.03.29]__: local gradio demo is online! `python app.py` to start the demo. New `requirments.txt` is used to avoid the bugs in `librosa`.
+
+- __[2023.03.28]__: Online demo is launched in [![Hugging Face Spaces](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces/vinthony/SadTalker), thanks AK!
+ 
+- __[2023.03.22]__: Launch new feature: generating the 3d face animation from a single image. New applications about it will be updated.
+
+- __[2023.03.22]__: Launch new feature: `still mode`, where only a small head pose will be produced via `python inference.py --still`. 
+
+- __[2023.03.18]__: Support `expression intensity`, now you can change the intensity of the generated motion: `python inference.py --expression_scale 1.3 (some value > 1)`.
+
+- __[2023.03.18]__: Reconfig the data folders, now you can download the checkpoint automatically using `bash scripts/download_models.sh`.
+- __[2023.03.18]__: We have offically integrate the [GFPGAN](https://github.com/TencentARC/GFPGAN) for face enhancement, using `python inference.py --enhancer gfpgan` for  better visualization performance.
+- __[2023.03.14]__: Specify the version of package `joblib` to remove the errors in using `librosa`, [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Winfredy/SadTalker/blob/main/quick_demo.ipynb) is online!
+- __[2023.03.06]__: Solve some bugs in code and errors in installation 
+- __[2023.03.03]__: Release the test code for audio-driven single image animation!
+- __[2023.02.28]__: SadTalker has been accepted by CVPR 2023!

docs/face3d.md

@@ -0,0 +1,48 @@
+## 3D Face visualization
+
+We use pytorch3d to visualize the produced 3d face from a single image.
+
+Since it is not easy to install, we produce a new install guidence here:
+
+```bash
+git clone https://github.com/Winfredy/SadTalker.git
+cd SadTalker 
+conda create -n sadtalker3d python=3.8
+source activate sadtalker3d
+
+conda install ffmpeg
+conda install -c fvcore -c iopath -c conda-forge fvcore iopath
+conda install libgcc gmp
+
+pip install torch==1.11.0+cu113 torchvision==0.12.0+cu113 torchaudio==0.11.0 --extra-index-url https://download.pytorch.org/whl/cu113
+
+# insintall pytorch3d
+pip install --no-index --no-cache-dir pytorch3d -f https://dl.fbaipublicfiles.com/pytorch3d/packaging/wheels/py38_cu113_pyt1110/download.html
+
+pip install -r requirements3d.txt
+
+### install gpfgan for enhancer
+pip install git+https://github.com/TencentARC/GFPGAN
+
+
+### when occurs gcc version problem `from pytorch import _C` from pytorch3d, add the anaconda path to LD_LIBRARY_PATH
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/$YOUR_ANACONDA_PATH/lib/
+
+```  
+
+
+Then, generating the result via:
+
+```bash
+
+
+python inference.py --driven_audio <audio.wav> \
+                    --source_image <video.mp4 or picture.png> \
+                    --result_dir <a file to store results> \
+                    --face3dvis
+
+```
+
+Then, the results will be given in the folders with the file name of `face3d.mp4`.
+
+More applications about 3d face will be released.

docs/install.md

@@ -0,0 +1,47 @@
+
+### Mac (Tested on M1 Mac OS 13.3)
+
+```
+git clone https://github.com/Winfredy/SadTalker.git
+
+cd SadTalker 
+
+conda create -n sadtalker python=3.8
+
+conda activate sadtalker
+
+# install pytorch 2.0
+pip install torch torchvision torchaudio
+
+conda install ffmpeg
+
+pip install -r requirements.txt
+
+pip install dlib # mac need to install the original dlib.
+
+```
+
+
+
+### Windows Native
+
+- Make sure you have `ffmpeg` in the `%PATH%` as suggested in [#54](https://github.com/Winfredy/SadTalker/issues/54), following [this](https://www.geeksforgeeks.org/how-to-install-ffmpeg-on-windows/) installation to install `ffmpeg`. 
+
+
+### Windows WSL
+- Make sure the environment: `export LD_LIBRARY_PATH=/usr/lib/wsl/lib:$LD_LIBRARY_PATH`
+
+
+### Docker installnation
+
+A dockerfile are also provided by [@thegenerativegeneration](https://github.com/thegenerativegeneration) in [docker hub](https://hub.docker.com/repository/docker/wawa9000/sadtalker), which can be used directly as:
+
+```bash
+docker run --gpus "all" --rm -v $(pwd):/host_dir wawa9000/sadtalker \
+    --driven_audio /host_dir/deyu.wav \
+    --source_image /host_dir/image.jpg \
+    --expression_scale 1.0 \
+    --still \
+    --result_dir /host_dir
+```
+

docs/webui_extension.md

@@ -0,0 +1,50 @@
+
+## Run SadTalker as a Stable Diffusion WebUI Extension.
+
+1. Installing the lastest version of [stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui) and install the sadtalker via `extension`.
+<img width="726" alt="image" src="https://user-images.githubusercontent.com/4397546/230698519-267d1d1f-6e99-4dd4-81e1-7b889259efbd.png">
+
+2. Download the checkpoints manually, for Linux and Mac:
+
+    ```bash
+
+    cd SOMEWHERE_YOU_LIKE
+
+    bash <(wget -qO- https://raw.githubusercontent.com/Winfredy/SadTalker/main/scripts/download_models.sh)
+    ```
+
+    For windows, you can download all the checkpoints from [google drive](https://drive.google.com/drive/folders/1Wd88VDoLhVzYsQ30_qDVluQr_Xm46yHT?usp=sharing) or [百度云盘](https://pan.baidu.com/s/1nXuVNd0exUl37ISwWqbFGA?pwd=sadt) 提取码: sadt.
+
+3.1. options 1: put the checkpoint in `stable-diffusion-webui/models/SadTalker` or `stable-diffusion-webui/extensions/SadTalker/checkpoints/`, the checkpoints will be detected automatically.
+
+3.2. Options 2: Set the path of `SADTALKTER_CHECKPOINTS` in `webui_user.sh`(linux) or `webui_user.bat`(windows) by:
+
+    > only works if you are directly starting webui from `webui_user.sh` or `webui_user.bat`.
+
+    ```bash
+    # windows (webui_user.bat)
+    set SADTALKER_CHECKPOINTS=D:\SadTalker\checkpoints
+
+    # linux (webui_user.sh)
+    export SADTALKER_CHECKPOINTS=/path/to/SadTalker/checkpoints
+    ```
+
+4. Then, starting the webui via `webui.sh or webui_user.sh(linux)` or `webui_user.bat(windows)` or any other methods, the SadTalker can be used in stable-diffusion-webui directly. 
+    
+    <img width="726" alt="image" src="https://user-images.githubusercontent.com/4397546/230698614-58015182-2916-4240-b324-e69022ef75b3.png">
+    
+## Questsions
+
+1. if you are running on CPU, you need to specific `--disable-safe-unpickle` in `webui_user.sh` or `webui_user.bat`.
+
+    ```bash
+    # windows (webui_user.bat)
+    set COMMANDLINE_ARGS="--disable-safe-unpickle"
+
+    # linux (webui_user.sh)
+    export COMMANDLINE_ARGS="--disable-safe-unpickle"
+    ```
+
+
+
+(Some [important discussion](https://github.com/Winfredy/SadTalker/issues/78) if you are unable to use `full` mode).

examples/driven_audio/chinese_news.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b0f4d313a1ca671bc4831d60bcf0c12225efbffe6c0e93e54fbfe9bcd4021cb
+size 1536078

examples/driven_audio/deyu.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba1839c57770a2ab0b593ce814344bfd4d750da02acc9be9e8cf5b9113a0f88a
+size 2694784

examples/driven_audio/eluosi.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4a3593815dc7b68c256672baa61934c9479efa770af2065fb0886f02713606e
+size 1786672

examples/driven_audio/fayu.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16ebd13626ae4171030b4ea05cceef06078483c352e4b68d469fc2a52bfffceb
+size 1940428

examples/driven_audio/imagine.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2db410217e074d91ae6011e1c5dc0b94f02d05d381c50af8e54253eeacad17d2
+size 1618510

examples/driven_audio/japanese.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3db5426d0b158799e2be4f609b11f75bfbd4affffe18e9a1c8e6f241fcdedcfc
+size 2622712

examples/ref_video/WDA_AlexandriaOcasioCortez_000.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a85242c3fc4d50e2202cea393b9e7ee59019759b68e78e26a254d528c22615a7
+size 2257667

examples/ref_video/WDA_KatieHill_000.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1fbb4cfd64eedc49b170c441714a9c4fd5e2c2f8a11592070ad89fbd257f2817
+size 3548230