Skip to content

Latest commit

 

History

History
60 lines (44 loc) · 1.78 KB

README.md

File metadata and controls

60 lines (44 loc) · 1.78 KB
Raw

privateLLM - managing private data for LLM

hacked during codecamp 2023.

concept of privateLLM

Hint: All function comments generously generated by chatGPT

Getting started

# install dependencies
poetry install

# start postgres vectorstore and jaeger trace collector
docker compose up -d

# copy environment template and adjust the settings
cp .env.template .env
vi .env

# run texttitan
poetry run python privatellm/main.py

Access the OpenAPI UI at http://localhost:8000/docs

The valid default credentials are

  • rrr:rrr
  • ttt:ttt
  • yyy:yyy

Use OpenAI models (required for most features)

Add some credit balance to your account and create an API key.

Place the API key in apikey.txt or provide it using the environment variable OPENAI_API_KEY.

Hint: Free accounts are not suffient.

Use gpt4All

  • install GPT4All
  • download a model
  • create a link to the model you want to use at gpt4all/model.bin

Use LLama2

  • Download the llama-2-7b-chat model from Meta (email registration required)
  • Convert the model using ggerganov/llama.cpp to GGUF format
  • Place it in llama-2-7b-chat/ggml-model-f16.gguf

presentation

Setup

for Linux

cd doc
wget https://github.com/yusukebe/revealgo/releases/download/v1.2.2/revealgo_1.2.2_Linux-64bit.tar.gz
tar xvfz revealgo_1.2.2_Linux-64bit.tar.gz
chmod +x revealgo
./revealgo

all others please find your corresponding revealgo version at https://github.com/yusukebe/revealgo/releases/tag/v1.2.2