• 🚀 Stable Video 4D (SV4D): A video-to-4D diffusion model for novel-view video synthesis.

  • Quickstart: Run python scripts/sampling/simple_video_sample_4d.py --input_path assets/sv4d_videos/test_video1.mp4 --output_folder outputs/sv4d.
  • Technical Details:
    • Trained to generate 40 frames at 576x576 resolution.
    • Run demo locally with python -m scripts.demo.gradio_app_sv4d.
  • Resources: Project page, Technical report, Video summary.

  

• 🎥 SV3D: An image-to-video model for multi-view synthesis.

  • Run a demo with streamlit run scripts/demo/video_sampling.py.
  • Models:
    • SV3D_u: Generates orbital videos from single images.
    • SV3D_p: Allows for camera path specification for 3D video.
  • Documentation: Project page, Tech report, Video summary.

  

• 🌟 Released SD-Turbo.

• ⚡ SDXL-Turbo: A fast, high-quality text-to-image model.
  • Installation: pip install streamlit-keyup.
  • Demo: Run streamlit run scripts/demo/turbo.py.
  • Technical Report.
  • 

• 🎬 Stable Video Diffusion (SVD) for image-to-video generation.
  • Model Details:
    • SVD: Generates 14 frames at 576x1024 resolution.
    • SVD-XT: Finetuned for 25 frames.
  • Run Locally: python -m scripts.demo.gradio_app.
  • Demos: scripts/demo/video_sampling.py and scripts/sampling/simple_video_sample.py.
  • Technical Report.
  • 

• 🌐 Released SDXL-base-1.0 and SDXL-refiner-1.0 models under the CreativeML Open RAIL++-M license.

• 📄 Technical report for SDXL available here.

• 🌱 Released two new diffusion models:
  • SDXL-base-0.9: Trained with various aspect ratios and resolution 1024².
  • SDXL-refiner-0.9: Denoises data with small noise levels for image-to-image use only.

git clone https://github.com/Stability-AI/generative-models.git
cd generative-models

Note: Tested with python3.10.

# Create virtual environment
python3 -m venv .pt2
source .pt2/bin/activate
pip3 install -r requirements/pt2.txt

pip3 install .

pip3 install -e git+https://github.com/Stability-AI/datapipelines.git@main#egg=sdata

• Build Wheel:

pip install hatch
hatch build -t wheel

• Install Wheel:

pip install dist/*.whl

Streamlit Demo: scripts/demo/sampling.py.

Supported Models:

• SDXL-base-1.0
• SDXL-refiner-1.0

Weights available under the CreativeML Open RAIL++-M license.

Run detection with:

python -m venv .detect
source .detect/bin/activate

pip install "numpy>=1.17" "PyWavelets>=1.1.1" "opencv-python>=4.1.0.25"
pip install --no-deps invisible-watermark

Test Commands:

python scripts/demo/detect.py <your filename>
python scripts/demo/detect.py <filename 1> <filename 2> ... <filename n>
python scripts/demo/detect.py <your folder name>/*

Training configurations are available in configs/example_training. Start training with:

python main.py...

• Icons for visual engagement.
• Bold headers for improved section visibility.
• Consistent layout for easier navigation.
• Shortened key instructions with emphasis on main points.

Let me know if you need further tweaks!