Installation
This page guides you through installing Neural Magic's core products – DeepSparse, SparseML, and SparseZoo. We provide streamlined installation commands for common use cases and links for in-depth instructions.
Prerequisites
Ensure your system meets the following requirements before proceeding with the installation based on your use case:
Deployment Hardware Requirements
- CPU: x86 with AVX2 instructions (Intel Haswell or newer, AMD Zen 2 or newer) or ARM v8.2 or newer.
- RAM: Minimum 1GB (model and configuration dependent)
Training Hardware Requirements
- CPU: Intel, AMD, or ARM
- RAM: 4GB+ recommended (depends on your models and workflows)
- GPU: NVIDIA (recommended, 16GB+ VRAM); AMD (limited support)
Software Requirements
- OS: Linux (e.g., Ubuntu, CentOS, Red Hat); MacOS (limited support)
- Python: 3.8 - 3.11
- PyTorch: >= 1.1.0, < 2.2
Task-Specific Installation
Choose the installation command tailored to your primary deep learning task.
LLMs - Causal Language Modeling
pip install deepsparse[llm,server] sparseml[llm]
Computer Vision: Object Detection
pip install deepsparse[yolov8,server] sparseml[yolov8]
Computer Vision: Image Classification
pip install deepsparse[torchvision,server] sparseml[torchvision]
Natural Language Processing
pip install deepsparse[transformers,server] sparseml[transformers]
Product-Specific Installation
For comprehensive installation guides and customization options, explore the dedicated pages for each product:
📄️ DeepSparse
Install DeepSparse, Neural Magic's high-performance inference engine, for optimized deep learning model deployment on CPUs.
📄️ SparseML
Install SparseML, Neural Magic's toolkit for optimizing deep learning models through state-of-the-art sparsification techniques.
📄️ SparseZoo
Install SparseZoo, Neural Magic's repository of pre-sparsified models, or learn how to access it through SparseML and DeepSparse.