LLaMa for Developers
1h 35mIntermediate2024-04-17
Authors
Denys Linkov
Course details
In this course, learn how to customize open-source AI models with one of the most common open-source models, LLaMa (Large Language Model Meta AI). Instructor Denys Linkov shares a hands-on approach to working with LLaMa, explaining LLaMa architecture, prompting, deploying, and training models. He uses a series of Python notebooks to show you how to adapt LLaMa to your use cases and employ it in an enterprise or startup environment.
Skills covered
LlamaNatural Language Processing (NLP)MetaProgramming FoundationsGenerative AIArtificial Intelligence (AI)Software DevelopmentOne-Off
Concepts
0. Introduction
- 01 - Developing AI models using LLaMA
1. Introduction to LLaMA
- 02 - Using LLaMA online
- 03 - Running LLaMA in a notebook
- 04 - Accessing LLaMA in an enterprise environment
2. LLaMA Architecture
- 05 - The LLaMA architecture
- 06 - The LLaMA tokenizer
- 07 - The LLaMA context window
- 08 - Differences between LLaMA 1 and 2
3. Fine-Tuning LLaMA
- 09 - Fine-tuning LLaMA with a few examples
- 10 - Fine-tuning LLaMA and freezing layers
- 11 - Fine-tuning with LLaMA using LoRa
- 12 - Reinforcement learning with RLHF and DPO
- 13 - Fine-tuning larger LLaMA models
4. Serving LLaMA
- 14 - Resources required to serve LLaMA
- 15 - Quantizing LLaMA
- 16 - Using TGI for serving LLaMA
- 17 - Using VLLM for serving LLaMA
- 18 - Using DeepSpeed for serving LLaMA
- 19 - Explaining LoRA and SLoRA
- 20 - Using a vendor for serving LLaMA
5. Prompting LLaMA
- 21 - Difference between LLaMA with commercial LLMs
- 22 - Few shot learning with LLaMA
- 23 - Chain of thought with LLaMA
- 24 - Using schemas with LLaMA
- 25 - Optimizing LLaMA prompts with DSPy
- 26 - Challenge - Generating product tags
- 27 - Solution - Generating product tags
Conclusion
- 28 - Continue your LlaMA AI model development journey