CLI Overview - Neurenix

The Neurenix CLI provides a comprehensive set of commands for managing your machine learning projects, from initialization to deployment.

Installation

The Neurenix CLI is included when you install the neurenix package:

pip install neurenix

Basic Usage

All Neurenix CLI commands follow this pattern:

neurenix <command> [<args>]

Available Commands

The Neurenix CLI provides the following commands:

Command	Description
`init`	Initialize a new Neurenix project with folder structure and config
`run`	Run a model training script with specified configuration
`eval`	Evaluate a trained model with specific metrics
`export`	Export a trained model to various formats (ONNX, TorchScript, etc.)
`serve`	Serve a trained model as a RESTful API
`save`	Save the current project state
`predict`	Make predictions using a trained model
`hardware`	Manage hardware settings and device selection
`preprocess`	Preprocess input data for model training
`monitor`	Monitor model training in real-time
`optimize`	Optimize a model with hyperparameters or quantization
`dataset`	Manage datasets for training and evaluation
`help`	Display help information about commands

Getting Help

To get help for any command:

# Show all available commands
neurenix help

# Get help for a specific command
neurenix help <command>
neurenix <command> --help

Common Workflows

Quick Start

# Initialize a new project
neurenix init --name my-project --template basic

# Navigate to project directory
cd my-project

# Train a model
neurenix run train.py

# Evaluate the model
neurenix eval --model models/model.nrx --data data/test.csv

# Export for deployment
neurenix export --model models/model.nrx --format onnx

# Serve the model
neurenix serve --model models/model.nrx --port 8000

Advanced Workflow

# Initialize with advanced template
neurenix init --name advanced-project --template advanced

# Train with custom configuration
neurenix run train.py --config config.json --epochs 50 --batch-size 64

# Monitor training
neurenix monitor --log-dir logs

# Optimize the model
neurenix optimize --model models/model.nrx --method quantization

# Export optimized model
neurenix export --model models/model.nrx --format onnx --optimize --quantize int8

Project Structure

When you initialize a Neurenix project, the CLI creates the following structure:

my-project/
├── config.json          # Project configuration
├── train.py            # Training script
├── data/               # Dataset directory
├── models/             # Saved models
├── configs/            # Additional configurations
└── logs/               # Training logs

​Installation

​Basic Usage

​Available Commands

​Getting Help

​Common Workflows

​Quick Start

​Advanced Workflow

​Project Structure

​Next Steps