Build an LLM from Scratch 1: Set up your code environment

2025-02-26 · Source: Sebastian Raschka · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Novice, long

Summary

Sebastian Raschka, author of "Build a Large Language Model from Scratch," provides a detailed guide for setting up a Python development environment to run the book's code examples. While the book's code is designed to run on various hardware, including older machines and GPUs, Raschka emphasizes using a dedicated Python installation to avoid conflicts with system-level Python. He recommends installing Python 3.11 or 3.12, slightly older than the latest 3.13, to ensure compatibility with libraries like PyTorch. The guide covers installing Python via Homebrew or python.org, setting up the fast UV package manager, creating isolated virtual environments, and installing all necessary packages from a `requirements.txt` file. For users facing local setup issues, cloud alternatives like Lightning Studio and Google Colab are suggested, with specific instructions for Colab.

Key takeaway

For AI Engineers setting up a development environment for LLM projects, prioritize creating isolated Python virtual environments using tools like UV. Install a stable Python version (e.g., 3.11 or 3.12) to ensure compatibility with deep learning libraries like PyTorch, rather than the absolute latest. This approach minimizes dependency conflicts and provides a clean, reproducible setup, allowing you to easily manage project-specific packages without affecting your system or other projects.

Key insights

Isolate Python environments for LLM development to ensure stability and package compatibility.

Principles

• Avoid modifying system Python installations.
• Use slightly older Python versions for library compatibility.
• Virtual environments prevent dependency conflicts.

Method

Install Python (e.g., 3.11/3.12), then UV package manager, create a virtual environment, activate it, and install all project dependencies using `uv pip install -r requirements.txt`.

In practice

• Install Python via Homebrew or python.org.
• Use UV for faster package management.
• Consider Google Colab for cloud execution.

Topics

• Python Environment Setup
• UV Package Manager
• Virtual Environments
• Large Language Models
• PyTorch

Best for: AI Student, Machine Learning Engineer, AI Engineer

Related on AIssential

• How to build your own AI developer tools with Claude Code | CJ Hess (Tenex) — Custom AI-native tools enhance developer workflows by transforming AI-generated plans into interactive visual diagrams.
• From Trainee to Trainer: LLM-Designed Training Environment for RL with Multi-Agent Reasoning — LLMs can automate reinforcement learning environment design by analyzing policy failures and proposing configuration changes.
• 10 Python Libraries for Building LLM Applications — Building LLM applications requires specialized Python libraries for control over model loading, data integration, serving, and evaluation.
• EnvSimBench: A Benchmark for Evaluating and Improving LLM-Based Environment Simulation — LLMs struggle with simultaneous state updates in environment simulation, requiring structured approaches for reliability.
• The “wow demo” trap is killing LLM projects. Here’s the exit. — Transitioning from LLM demos to production requires structured learning in architecture, reliability, and full-stack implementation.
• What is your current local LLM setup? — The community actively explores diverse local LLM setups, balancing hardware, software, and model choices for varied applications.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Sebastian Raschka.