Skip to content

User Guide

In-depth documentation of GPUX core concepts and features.

📚 What's in This Guide?

This guide covers core GPUX concepts in depth. Each section builds on previous knowledge from the Tutorial.

📖 Guide Sections

Working with Models

Learn about ONNX models, conversion from popular frameworks, optimization techniques, and model management.

Topics: - Understanding ONNX format - Converting from PyTorch, TensorFlow, scikit-learn, Hugging Face - Model inspection and visualization - Optimization (quantization, compression) - Versioning and debugging

GPU Providers

Deep dive into execution providers and hardware-specific optimization.

Topics: - What execution providers are - Available providers (TensorRT, CUDA, CoreML, ROCm, etc.) - Provider selection logic - Platform-specific setup and configuration - Performance comparison and troubleshooting

Inputs & Outputs

Master input/output handling, data types, and validation.

Topics: - Input data types and shapes - JSON format requirements - Output handling and labels - Automatic validation - Type conversion

Data Preprocessing

Preprocessing pipelines for text, images, and audio data.

Topics: - Text tokenization - Image preprocessing (resize, normalize) - Audio resampling - Configuration options - Custom preprocessing

Batch Inference

Process multiple inputs efficiently for higher throughput.

Topics: - Batch input format - Batch size optimization - Performance gains - Memory considerations - Python API for batching

Python API

Complete reference for using GPUX programmatically.

Topics: - GPUXRuntime class - Inference methods - Batch processing - Benchmarking - Context managers - Testing

Error Handling

Common errors, solutions, and debugging techniques.

Topics: - Common error messages - Debugging methods - Exception handling - Validation techniques - Troubleshooting guide

🎯 How to Use This Guide

For Beginners

Start with the Tutorial first, then return here for deeper understanding.

For Experienced Users

Jump directly to relevant sections based on your needs.

For Reference

Use this guide as a reference when working on specific features.

💬 Need Help?