ML Container Creator¶
ML Container Creator (MCC) is a Yeoman generator that produces bring-your-own-container (BYOC) projects for deploying ML models on Amazon SageMaker. You select a model, a serving framework, and a deployment target — MCC generates the Dockerfile, serving code, deployment scripts, and tests.
What It Supports¶
Serving Architectures¶
| Architecture | Backends | Use Case |
|---|---|---|
| HTTP | Flask, FastAPI | Predictive models (sklearn, XGBoost, TensorFlow) |
| Transformers | vLLM, SGLang, TensorRT-LLM, LMI, DJL | Large language models |
| Triton | FIL, ONNX Runtime, TensorFlow, PyTorch, vLLM, TensorRT-LLM, Python | Multi-framework model serving via NVIDIA Triton |
| Diffusors | vLLM-Omni | Diffusion models (image generation) |
Deployment Targets¶
| Target | Description |
|---|---|
| Managed Inference | SageMaker real-time endpoints |
| Async Inference | S3-based asynchronous processing with SNS notifications |
| Batch Transform | S3-to-S3 dataset processing |
| HyperPod EKS | Kubernetes deployment on SageMaker HyperPod clusters |
Quick Start¶
npm install -g yo
git clone https://github.com/awslabs/ml-container-creator.git
cd ml-container-creator
npm install && npm link
yo @aws/ml-container-creator
See Getting Started for prerequisites, installation details, and a full walkthrough.
Documentation¶
User Guide¶
- Getting Started — Install MCC and deploy your first model
- How It Works — Architecture, prompt flow, and generated project structure
- Configuration — CLI flags, environment variables, config files, and MCP
- Deployment & Inference — Build paths, deployment targets, and lifecycle scripts
- Examples — End-to-end walkthroughs for each architecture
- Troubleshooting — Common issues and solutions
Developer Guide¶
- Contributing — Development setup and contribution workflow
Links¶
License¶
Apache-2.0. See CONTRIBUTING for security issue reporting.