Local/Private AI/LLM Ultra Crash Course

Explore how large language models work—from pre-training on vast data with transfer learning to fine-tuning for coding—using transformer architecture, tokens, and high-dimensional vectors.

Learn to download and install Ollama to run large language models locally, connect apps to a local Ollama installation, and configure model storage with environment variables.

Learn to select and run models on Ollama, choose llama3, compare sizes (8b and 70b parameters), and decide between instruct and text models for clear instructions and better chat results.

Practice running a simple chat across multiple AI models, compare the performance and resources of small and large llama3, llava, and deep seek coder, and craft prompts including images.

Explore how leaderboards and benchmarks help you compare leading llm models, assess math-focused performance, and test multiple models with ollama for your tasks.

Learn to install and run the open web user interface for Ollama with Docker, sign up, select models, compare results across multiple models, and manage prompts and settings.

Learn to install anything llm locally with Docker and Ollama. Build workspaces, embed documents, manage vector storage, and explore agents and web sources.

Explore code and text generation and manipulation using Visual Studio Code and ollama, connect local models with the continue.dev extension, and craft a kids story through prompts and edits.

Master code creation and editing with advanced techniques, using coding models, embeddings, and a live preview workflow to configure and refine a coding project.

Explore prompting techniques, from setting an overall goal and global-to-detail reasoning to clear, precise instructions; apply hierarchical steps and example-driven prompts to optimize AI responses.

Practice crafting prompts to outline complex AI presentations, generate slide-friendly content, and explain bullet points clearly using simple language; evaluate results and refine slides to fit each slide.

Wrap up the course by inviting questions, sharing Telegram and Discord channel links in the course materials, and preparing learners for the next course.