Modular

AMD ROCm-optimized MAX for MI200/MI300 series. Fast GenAI inference on AMD hardware.

Complete MAX stack: CUDA, PyTorch, cuDNN included. Production-ready NVIDIA GenAI deployment.

Lightweight MAX essentials. Faster downloads, minimal dependencies. Needs NVIDIA drivers.

AMD ROCm-optimized MAX for MI200/MI300 series. Fast GenAI inference on AMD hardware.

Full-stack recipes built upon MAX + FastAPI + React. Runs on AMD + NVIDIA.

Universal MAX image supporting both NVIDIA & AMD GPUs. Cross-platform GenAI deployment solution.

Deploy GPU-optimized generative AI models using the MAX container Helm chart for Kubernetes.