Skip to main content

​
Installation & Setup

The AITraining CLI is included when you install the aitraining package.

​
Installation

​
Using pip

pip install aitraining torch torchvision torchaudio

uv pip install aitraining torch torchvision torchaudio

​
Verify Installation

Check that the CLI is working: 
aitraining --version
You should see the version number printed.

​
First Run

Running aitraining without arguments launches the interactive wizard: 
aitraining
This displays the Monostate banner and starts the configuration wizard.

​
GPU Setup

For GPU training, ensure you have CUDA installed and PyTorch can detect your GPU: 
import torch
print(torch.cuda.is_available())  # Should print True
print(torch.cuda.device_count())  # Number of GPUs

​
Apple Silicon (M1/M2/M3)

On Apple Silicon Macs, MPS is used automatically: 
import torch
print(torch.backends.mps.is_available())  # Should print True

​
Environment Variables

Common environment variables:
VariableDescription
HF_TOKENHugging Face token for private models/datasets
WANDB_API_KEYWeights & Biases API key for logging
CUDA_VISIBLE_DEVICESSpecify which GPUs to use
Set these before running training: 
export HF_TOKEN="your_token_here"
export WANDB_API_KEY="your_key_here"

​
Next Steps

Command Structure

Learn the CLI command syntax

LLM Training

Train language models