## Live Coding: GenAI Multimodal, RAG, RAPTOR, Agent, and Finetuning

Here, a total of 6 Generative AI demos were run and posted online: RAG, RAPTOR, Agent, and Finetuning Colab notebooks, as well as Multimodal image prompts given to OpenAI ChatGPT 4o and Google Gemini. [Recorded Discussion](https://youtu.be/qc5_MzJG_QY)

Generations from notebooks typically reflected their desired purpose, with Finetuning being applied with a biochemistry structures dataset and tested with general and scientific questions, and then pushed to Hugging Face as downloadable 16 bit and 4 bit GGUF versions during the live discussion. Hugging Face Models: [Model 1](https://huggingface.co/kevinkawchak/gradientai-Llama-3-8B-Instruct-Gradient-1048k-16Molecule), [Model 2](https://huggingface.co/kevinkawchak/gradientai-Llama-3-8B-Instruct-Gradient-1048k-Molecule-q4-k-m-GGUF)

Models adapted to Colab environments and additional resources mentioned above are available through the GitHub Repository.

Original Models or Chatbots:

1) RAG, Sakil786 GitHub: https://github.com/Sakil786/RAG_using_Llama-3/blob/main/RAG_using_Llama_3.ipynb

2) Multimodal, OpenAI ChatGPT 4o: https://chatgpt.com/ 

3) Multimodal, Google Gemini: https://gemini.google.com/

4) RAPTOR, Langchain GitHub: https://github.com/langchain-ai/langchain/blob/master/cookbook/RAPTOR.ipynb

5) AGENT, Langchain GitHub: https://github.com/langchain-ai/langgraph/blob/main/examples/multi_agent/agent_supervisor.ipynb

6) Finetuning, unsloth: https://colab.research.google.com/drive/135ced7oHytdxu3N2DNe1Z0kqjyYIkDXp?usp=sharing