Everything You Need to Ship Models
A full Mac with administrator access: install the ML tools you use
Model Training
Prototype and train small models on Apple Silicon with MLX, PyTorch and Metal acceleration
Unified Memory
16GB shared by a 10-core CPU and a 10-core GPU on a dedicated Mac mini M4
Dataset Management
Keep datasets on the 256GB SSD, or fetch them from your own cloud storage
Your Own Tools
Install the experiment tracking you use, such as MLflow or TensorBoard
Reproducible Research
Version your environments, data, and models with DVC and Git
HuggingFace Ready
Fine-tune transformers, run inference, and push to the Hub
Why ML Engineers Choose Orbit
Mac mini M4: 10-core CPU, 10-core GPU, 16GB unified memory, 256GB SSD
Install PyTorch, TensorFlow or MLX with Metal acceleration
Jupyter, VS Code, and any macOS tool you install
Core ML and MLX run natively on Apple Silicon
256GB SSD for your checkpoints
Full macOS desktop with administrator access