Local Models

Running LLMs on your own hardware β€” from first installation to advanced tuning.

← Awesome AI Handbook Β· πŸ‡·πŸ‡Ί Русский

Where to Start β€” Choose Your Scenario

Absolute Beginner

Never ran a model, not familiar with the terminal.

First: basics/README.md β€” what AI is, how models work, what hardware you need.

Then install: Environment setup and installation β€” Homebrew, Ollama, first model in 10 minutes.

Or skip the terminal entirely: LM Studio (described in running-models.md).

I Want to Understand

Already installed Ollama, ran a model, but want to know how it works.

  1. How to find and run a model β€” hands-on guide
  2. Choosing a model for your task β€” what to pick for coding / chat / RAG
  3. Memory and context β€” why models dont fit and how to fix it

Advanced User

Want maximum performance, API tuning, tool comparison.

  1. Advanced Ollama setup β€” Modelfile, API, env vars
  2. Tool comparison β€” Ollama vs LM Studio vs MLX vs llama.cpp
  3. Quantization β€” Q4, Q5, Q8 β€” what to choose
  4. Apple Silicon benchmarks β€” tok/s on M1–M4

Reference (for Everyone)

Section About
Model catalog 50+ models with specs
Common problems Diagnostics and fixes

Section Files

# File Audience Time
1 getting-started.md Beginners 10 min
2 running-models.md Everyone 15 min
3 models.md Everyone 10 min
4 memory-and-context.md Everyone 10 min
5 advanced-setup.md Advanced 15 min
6 tools.md Everyone 20 min
7 quantization.md Advanced 10 min
8 benchmarks/apple-silicon.md Everyone 5 min
9 troubleshooting.md When needed 5 min
10 catalog.md Reference β€”

Navigation: ← Back to main Β· Catalog