This is example code to show how to train AI on your documents

The prevalent method used by most developers involves building vector stores from documents (including PDFs). These can then be used with ChatGPT 3.5 or 4, depending on the budget. The file named langchain.py demonstrates this approach.

Fine-tuning ChatGPT with your documents is possible with ChatGPT. This approach requires the preparation of data for fine-tuning the AI model, but it yields much better results. As of now, this is only possible with ChatGPT 3.5. However, ChatGPT 4 will be available for fine-tuning by the end of the 2023. Example of this method is in the ai-finetuning.ts file.