AI PLAYGROUND

This is where I place code dealing with AI in different forms and shapes.

It contains different projects depending on the idea and finding of the day. Code for most of the projects already exists and url are provided in the .py or .ipynb files.

The majority of the work here is adapting the code to my current environment or solving the question with a new LLM or approach.

To access most models it is necesary to have api keys for each LLM provider. Some original videos or code had the instructions for that purpose. I did not add that here.

1. HTML_PARSER --> Web Scrapping methods that include AI to analyze the data

2. HUGGING_FACE --> Testing different functionality using HuggingFace interphase. There is IBM granite code there.

3. IMAGE_TO_CODE --> Results from interactions with CLAUDE model, in which an image was provided and it generated code for different apps.

4. IMPROVE_CODE --> A couple of ways to generate or improve code with LlamaCode and IBM's Granite Code (The latest is very slow*)

5. LANGCHAIN --> Several notebooks utilizing Langchain tools one way or another.

6. LLM_PDF_PARSER --> Methods to Parse PDF to Text and ask questions about the documents.

7. TALK_TO_LLAMA_STREAMLIT --> Experiment with both llama and llamacode to QA and document parsing (pdf and txt)

Most of the tests are done using LLama3 on Ollama

*slownest may be the result of my system not having a GPU.