No description
Find a file
2023-08-26 17:33:31 +02:00
level_1 quick fix to the bug 2023-08-17 08:41:30 +02:00
level_2 Updates to the api + refactor 2023-08-25 12:11:12 +02:00
.gitignore Merge branch 'main' into code_review 2023-08-25 12:12:46 +02:00
infographic_final.png Add files via upload 2023-08-16 18:23:40 +02:00
LICENSE Initial commit 2023-08-16 18:16:33 +02:00
README.md Update README.md 2023-08-26 17:33:31 +02:00

PromethAI-Memory

Memory management for the AI Applications and AI Agents

Infographic Image

The Motivation

Browsing the database of theresanaiforthat.com, we can observe around 7000 new, mostly semi-finished projects in the field of applied AI, whose development is fueled by new improvements in foundation models and open-source community contributions.

It seems it has never been easier to create a startup, build an app, and go to market… and fail.

AI apps currently being pushed out still mostly feel and perform like demos.

To address this issue, dlthub and prometh.ai will collaborate on a productionizing a common use-case, progressing step by step. We will utilize the LLMs, frameworks, and services, refining the code until we attain a clearer understanding of what a modern LLM architecture stack might entail.

Read more on our blog post prometh.ai

PromethAI-Memory Repo Structure

The repository contains a set of folders that represent the steps in the evolution of the modern data stack from POC to production

  • Level 1 - CMD script to process PDFs We introduce the following concepts:
    1. Structured output with Pydantic
    2. CMD script to process custom PDFs
  • Level 2 - Memory Manager implemented in Python We introduce the following concepts:
    1. Long Term Memory
    2. Short Term Memory
    3. Episodic Buffer
    4. Attention Modulators The code at this level contains:
    5. Simple PDF ingestion
    6. FastAPI
    7. Docker Image
    8. Memory manager
    9. Langchain-based Agent Simulator
    10. Data schema

How to use

Each of the folders contains a README to get started.