🚀 Ultralytics Replicate

Deploy Ultralytics YOLO models to Replicate with ready-to-use Cog configurations and automated CI/CD workflows.

🗂️ Repository Structure

This repository provides optimized Replicate deployments for YOLO11n, YOLOv8s WorldV2, YOLOE-11S, and the full YOLO26 task family (detect, segment, semantic, classify, pose, OBB) with an automated CI/CD workflow.

ultralytics/replicate/
│
├── yolo11n/                  # YOLO11n model deployment
│   ├── cog.yaml              # Cog configuration
│   ├── predict.py            # Prediction interface
│   ├── download.py           # Weight download script (run before cog build)
│   └── README.md             # Model documentation
├── yolov8s-worldv2/          # YOLOv8s WorldV2 model deployment (same four-file layout)
├── yoloe11s/                 # YOLOE-11S model deployment (same four-file layout)
├── yolo26/                   # YOLO26 family — 6 tasks, each with selectable size n/s/m/l/x
│   ├── detect/               # Object detection       → r8.im/ultralytics/yolo26
│   ├── seg/                  # Instance segmentation  → r8.im/ultralytics/yolo26-seg
│   ├── sem/                  # Semantic segmentation  → r8.im/ultralytics/yolo26-sem
│   ├── cls/                  # Classification         → r8.im/ultralytics/yolo26-cls
│   ├── pose/                 # Pose estimation        → r8.im/ultralytics/yolo26-pose
│   └── obb/                  # Oriented bounding box  → r8.im/ultralytics/yolo26-obb
│       ├── cog.yaml          #   each task dir has the standard four-file layout:
│       ├── predict.py        #   cog.yaml, predict.py, download.py, README.md
│       ├── download.py
│       └── README.md
├── assets/                   # Sample images for workflow smoke tests
│
├── .github/workflows/        # Automated deployment
│   ├── push.yml              # Model deployment workflow
│   ├── ci.yml                # Code quality checks
│   └── format.yml            # Code formatting
│
├── test_prediction.py        # Local YOLO11n testing utility
├── replicate_metadata.py     # Creates Replicate models and syncs page metadata on deploy
├── requirements.txt          # Dependencies
├── LICENSE                   # AGPL-3.0 license
└── README.md                 # This file

⚡ Quick Start

Deploy a Model

Models deploy to the corresponding Replicate endpoints:

# Clone repository
git clone https://github.com/ultralytics/replicate.git
cd replicate

# Deploy YOLO11n
cd yolo11n
python download.py
cog login
cog push r8.im/ultralytics/yolo11n

# Or deploy another configured model
cd ../yolov8s-worldv2
python download.py
cog push r8.im/ultralytics/yolov8s-worldv2

cd ../yoloe11s
python download.py
cog push r8.im/ultralytics/yoloe-11s

# Or deploy a YOLO26 task (detect/seg/sem/cls/pose/obb)
cd ../yolo26/seg
python download.py
cog push # uses the image: target in cog.yaml (r8.im/ultralytics/yolo26-seg)

Automated Deployment with GitHub Actions

1. Setup secrets:

   • Go to repository Settings → Secrets → Actions
   • Add REPLICATE_API_TOKEN with your Replicate API token

2. Deploy:

   • Manual: Actions tab → "Push YOLO to Replicate" → Run workflow
   • Automatic: Push changes to main builds, tests, and deploys each configured model

🛠️ Installation

Install Cog (Replicate's deployment tool):

sudo curl -o /usr/local/bin/cog -L https://github.com/replicate/cog/releases/latest/download/cog_$(uname -s)_$(uname -m)
sudo chmod +x /usr/local/bin/cog

For local development and testing:

pip install -r requirements.txt

🎯 Available Models

Directory	Replicate model	Predictor	Notes
yolo11n/	r8.im/ultralytics/yolo11n	YOLO	Official YOLO11n object detection model
yolov8s-worldv2/	r8.im/ultralytics/yolov8s-worldv2	YOLOWorld	Open-vocabulary YOLOv8s WorldV2 model
yoloe11s/	r8.im/ultralytics/yoloe-11s	YOLOE	YOLOE-11S segmentation model with class prompt support
yolo26/detect/	r8.im/ultralytics/yolo26	YOLO	YOLO26 detection (COCO), runtime size n/s/m/l/x
yolo26/seg/	r8.im/ultralytics/yolo26-seg	YOLO	YOLO26 instance segmentation (COCO-Seg)
yolo26/sem/	r8.im/ultralytics/yolo26-sem	YOLO	YOLO26 semantic segmentation (Cityscapes)
yolo26/cls/	r8.im/ultralytics/yolo26-cls	YOLO	YOLO26 classification (ImageNet)
yolo26/pose/	r8.im/ultralytics/yolo26-pose	YOLO	YOLO26 pose estimation (COCO-Pose)
yolo26/obb/	r8.im/ultralytics/yolo26-obb	YOLO	YOLO26 oriented bounding box (DOTAv1)

🔧 Model Setup

Each model directory includes a download.py script used by the deployment workflow before cog build:

python yolo11n/download.py
python yolov8s-worldv2/download.py
python yoloe11s/download.py
python yolo26/seg/download.py # YOLO26 task dirs (detect/seg/sem/cls/pose/obb) follow the same pattern

🧪 Local Testing

Test the model locally before deploying:

# Test YOLO11n
python yolo11n/download.py
python test_prediction.py --model yolo11n --image assets/bus.jpg

🚀 Features

• 🏎️ Optimized: PyTorch model for fast inference
• 🤖 Automated: GitHub Actions for CI/CD
• 📦 Ready-to-use: Pre-configured deployments for multiple YOLO models
• 📊 Scalable: Auto-scaling Replicate infrastructure
• 🎯 Focused: One Cog configuration per model

💡 Contribute

Ultralytics thrives on community collaboration, and we deeply value your contributions! Whether it's reporting bugs, suggesting features, or submitting code changes, your involvement is crucial.

• Reporting Issues: Encounter a bug? Please report it on GitHub Issues.
• Feature Requests: Have an idea for improvement? Share it via GitHub Issues.
• Pull Requests: Want to contribute code? Please read our Contributing Guide first, then submit a Pull Request.
• Feedback: Share your thoughts and experiences by participating in our official Survey.

A heartfelt thank you 🙏 goes out to all our contributors! Your efforts help make Ultralytics tools better for everyone.

📄 License

Ultralytics offers two licensing options to accommodate diverse needs:

• AGPL-3.0 License: Ideal for students, researchers, and enthusiasts passionate about open collaboration and knowledge sharing. This OSI-approved open-source license promotes transparency and community involvement. See the LICENSE file for details.
• Enterprise License: Designed for commercial applications, this license permits the seamless integration of Ultralytics software and AI models into commercial products and services, bypassing the copyleft requirements of AGPL-3.0. For commercial use cases, please inquire about an Ultralytics Enterprise License.

📮 Contact

For bug reports or feature suggestions related to this project or other Ultralytics projects, please use GitHub Issues. For general questions, discussions, and community support, join our Discord server!