This repo is dedicated to building apps with LLMs. Currently they are based on OpenAI API. In the future, I plan to also use open-source LLMs like those in Hugging Face.
Large Language Models (LLMs) are powerful tools that can be used for a variety of tasks, including question answering, summarization, and translation. However, LLMs can see their performance decrease when generating responses, especially for open-ended, challenging questions, or questions requiring information not publicly available.
ChromaDB is an open-source embedding database that can be used to improve the performance of LLMs.
The process of retrieving relevant documents and using them as context for a generative model is known as retrieval-augmented generation (RAG), and it allows LLMs to make inferences using information that wasn’t included in their training dataset.