Initial deployment of SadTalker with models

.gitattributes

@@ -33,3 +33,4 @@ 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
+*.mp4 filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,13 +1,53 @@
 ---
-title: Sadtalker
-emoji: 🌍
-colorFrom: green
-colorTo: pink
+title: SadTalker
+emoji: 😭
+colorFrom: blue
+colorTo: purple
 sdk: gradio
-sdk_version: 6.3.0
+sdk_version: 3.50.2
 app_file: app.py
 pinned: false
 license: mit
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# SadTalker: Audio-Driven Talking Face Animation
+
+This Space demonstrates **SadTalker**, a system for generating realistic talking head videos from a single image and audio input.
+
+## Features
+
+- 🎭 **Single Image Animation**: Animate any portrait image with audio
+- 🎵 **Audio-Driven**: Synchronize facial movements with speech
+- ⚙️ **Customizable Settings**: Adjust pose style, preprocessing, and enhancement options
+- 🎨 **Face Enhancement**: Optional GFPGAN integration for improved quality
+
+## Usage
+
+1. **Upload an Image**: Choose a portrait image (preferably with a clear, front-facing face)
+2. **Upload Audio**: Provide an audio file with speech
+3. **Adjust Settings** (optional):
+   - **Pose Style**: Choose from 47 different pose styles (0-46)
+   - **Face Model Resolution**: 256 or 512 (higher = better quality but slower)
+   - **Preprocess Type**: How to handle the input image
+   - **Still Mode**: Reduces head motion for a more static result
+   - **GFPGAN Enhancer**: Improves face quality in the output
+4. **Generate**: Click the "Generate Video" button and wait for processing
+
+## Tips for Best Results
+
+- Use clear, well-lit portrait images
+- Front-facing photos work best
+- Audio should be clear with minimal background noise
+- For more details, see the [best practices guide](https://github.com/OpenTalker/SadTalker/blob/main/docs/best_practice.md)
+
+## Credits
+
+**Paper**: [SadTalker: Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation](https://arxiv.org/abs/2211.12194) (CVPR 2023)
+
+**Authors**: Wenxuan Zhang, Xiaodong Cun, Xuan Wang, Yong Zhang, Xi Shen, Yu Guo, Ying Shan, Fei Wang
+
+**GitHub**: [OpenTalker/SadTalker](https://github.com/OpenTalker/SadTalker)
+
+## License
+
+This project is licensed under the MIT License.

SadTalker

@@ -0,0 +1 @@
+Subproject commit cd4c0465ae0b54a6f85af57f5c65fec9fe23e7f8

app.py

@@ -0,0 +1,109 @@
+import os
+import sys
+
+# Add SadTalker to path
+sys.path.append('SadTalker')
+
+import gradio as gr
+from SadTalker.src.gradio_demo import SadTalker
+
+def sadtalker_demo():
+    """Initialize and return the SadTalker Gradio interface"""
+    
+    checkpoint_path = 'SadTalker/checkpoints'
+    config_path = 'SadTalker/src/config'
+    
+    sad_talker = SadTalker(checkpoint_path, config_path, lazy_load=True)
+
+    with gr.Blocks(analytics_enabled=False, title="SadTalker - Talking Head Generator") as sadtalker_interface:
+        gr.Markdown("""
+        <div align='center'> 
+            <h2>😭 SadTalker: Audio-Driven Talking Face Animation</h2>
+            <p>Upload an image and audio to generate a talking head video</p>
+            <a href='https://arxiv.org/abs/2211.12194'>Paper</a> &nbsp;&nbsp;
+            <a href='https://sadtalker.github.io'>Homepage</a> &nbsp;&nbsp;
+            <a href='https://github.com/OpenTalker/SadTalker'>GitHub</a>
+        </div>
+        """)
+        
+        with gr.Row():
+            with gr.Column(variant='panel'):
+                with gr.Tabs():
+                    with gr.TabItem('Upload Image'):
+                        source_image = gr.Image(
+                            label="Source Image", 
+                            source="upload", 
+                            type="filepath"
+                        )
+
+                with gr.Tabs():
+                    with gr.TabItem('Upload Audio'):
+                        driven_audio = gr.Audio(
+                            label="Input Audio", 
+                            source="upload", 
+                            type="filepath"
+                        )
+                            
+            with gr.Column(variant='panel'): 
+                with gr.Tabs():
+                    with gr.TabItem('Settings'):
+                        gr.Markdown("Need help? Visit our [best practice page](https://github.com/OpenTalker/SadTalker/blob/main/docs/best_practice.md)")
+                        
+                        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'
+                        )
+                        preprocess_type = gr.Radio(
+                            ['crop', 'resize', 'full', 'extcrop', 'extfull'], 
+                            value='crop', 
+                            label='Preprocess Type'
+                        )
+                        is_still_mode = gr.Checkbox(
+                            label="Still Mode (less head motion)"
+                        )
+                        batch_size = gr.Slider(
+                            label="Batch Size", 
+                            step=1, 
+                            minimum=1,
+                            maximum=10, 
+                            value=2
+                        )
+                        enhancer = gr.Checkbox(
+                            label="GFPGAN Face Enhancer"
+                        )
+                        
+                        submit = gr.Button('Generate Video', variant='primary')
+                        
+                with gr.Tabs():
+                    with gr.TabItem('Generated Video'):
+                        gen_video = gr.Video(label="Generated Video", format="mp4")
+
+        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()

requirements.txt

@@ -0,0 +1,23 @@
+torch>=1.12.0
+torchvision>=0.13.0
+numpy==1.23.4
+face_alignment==1.3.5
+imageio==2.19.3
+imageio-ffmpeg==0.4.7
+librosa==0.9.2
+numba
+resampy==0.3.1
+pydub==0.25.1
+scipy==1.10.1
+kornia==0.6.8
+tqdm
+yacs==0.1.8
+pyyaml
+joblib==1.1.0
+scikit-image==0.19.3
+basicsr==1.4.2
+facexlib==0.3.0
+gradio
+gfpgan
+av
+safetensors