Setup Guide - Zero to AI¶

Quick Start with UV (Recommended)¶

UV is an extremely fast Python package installer and resolver written in Rust. It’s 10-100x faster than pip!

1. Run the installation script¶

./install_dependencies.sh

This will:

Install UV if not already installed
Create a virtual environment at .venv/
Install all dependencies using UV
Activate the environment

2. Activate the environment¶

source .venv/bin/activate

3. Start learning!¶

# Start Jupyter
jupyter notebook

# Or open a specific notebook
jupyter notebook 04-token/tiktoken_example.ipynb

Manual Installation with UV¶

If you prefer manual control:

Install UV¶

curl -LsSf https://astral.sh/uv/install.sh | sh

Create virtual environment¶

uv venv .venv
source .venv/bin/activate  # On macOS/Linux
# .venv\Scripts\activate   # On Windows

Install dependencies¶

# Core dependencies
uv pip install -r requirements.txt

# Or install specific packages from the curriculum stack
uv pip install torch transformers tiktoken sentence-transformers jupyter

Alternative: Traditional pip/venv¶

If you prefer not to use UV:

# Create virtual environment
python -m venv .venv
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Verify Installation¶

# Check Python version
python --version  # Should be 3.11+

# Test imports
python -c "import torch; print(f'PyTorch {torch.__version__}')"
python -c "import transformers; print(f'Transformers {transformers.__version__}')"
python -c "import tiktoken; print('tiktoken ✅')"
python -c "import sentence_transformers; print('sentence-transformers ✅')"

# Start Jupyter
jupyter notebook

Dependencies Overview¶

Core ML/DL¶

PyTorch - Deep learning framework
Transformers - HuggingFace library (BERT, GPT, etc.)
Torchvision - Vision models and datasets

Tokenization¶

tiktoken - OpenAI’s tokenizer
tokenizers - HuggingFace tokenizers library
sentencepiece - Google’s tokenizer

Embeddings & Vector DBs¶

sentence-transformers - Sentence embeddings
chromadb - Vector database
qdrant-client - Qdrant vector DB
weaviate-client - Weaviate vector DB
pymilvus - Milvus vector DB
pgvector - PostgreSQL vector extension

Data Science¶

numpy, pandas, scikit-learn, scipy - Data science essentials
matplotlib, seaborn, plotly - Visualization

Development¶

jupyter - Interactive notebooks
pytest, black, flake8 - Testing and formatting (dev extras)

Learning Path¶

Once setup is complete, follow this path:

23-glossary/GLOSSARY.md
MASTER_STUDY_GUIDE.md
checklist.md
Start with the first phase for your chosen track
Use 00_START_HERE.ipynb inside each module when available

Which Requirements File to Use?¶

File	Use When
`requirements.txt`	Local development (pip)
`environment.yml`	Local development (Conda/Mamba)
`colab_requirements.txt`	Google Colab — skips pre-installed packages
`kaggle_requirements.txt`	Kaggle notebooks — skips pre-installed packages, requires Internet toggle enabled
`requirements-replit.txt`	Replit — auto-installed on run
`pyproject.toml`	`uv pip install -e .` or editable installs

Troubleshooting¶

UV installation fails¶

# Try with pip instead
pip install uv

PyTorch GPU support¶

For CUDA support, install PyTorch separately:

# See https://pytorch.org for your specific CUDA version
uv pip install torch torchvision --index-url https://download.pytorch.org/whl/cu118

PostgreSQL dependencies¶

For pgvector to work, ensure PostgreSQL development headers are installed:

# macOS
brew install postgresql

# Ubuntu/Debian
sudo apt-get install postgresql-server-dev-all

# Then reinstall psycopg2-binary
uv pip install --force-reinstall psycopg2-binary

Jupyter kernel not found¶

python -m ipykernel install --user --name=aiml

Environment Variables¶

Some notebooks require API keys or database connections. Create a .env file:

# OpenAI (for embeddings notebooks)
OPENAI_API_KEY=your_key_here

# Aurora PostgreSQL (for 6-vector-databases/06_aurora_pgvector_guide.ipynb)
AURORA_HOST=your-cluster.region.rds.amazonaws.com
AURORA_PORT=5432
AURORA_DATABASE=vectordb
AURORA_USER=postgres
AURORA_PASSWORD=your_password

# AWS (optional, for AWS services)
AWS_ACCESS_KEY_ID=your_key
AWS_SECRET_ACCESS_KEY=your_secret
AWS_DEFAULT_REGION=us-east-1

Why UV?¶

UV is a next-generation Python package manager:

✅ 10-100x faster than pip
✅ Better dependency resolution (like Poetry/PDM)
✅ Drop-in replacement for pip commands
✅ Written in Rust for maximum performance
✅ Creates reproducible environments

Example speed comparison:

pip: ~45 seconds
UV: ~2 seconds

Need Help?¶

Check notebook 00_START_HERE.ipynb files in each section
Review README.md in each folder

Happy learning! 🚀