Git Product home page Git Product logo

logviz's Introduction

Logviz: Visualize openai/evals logs

This is a simple webapp for visualizing logs from openai/evals, primarily developed by me (Ian McKenzie) and Dane Sherburn (@danesherbs).

image

Running the webapp

  • The webapp can be run from the command line by simply running logviz
  • The webapp will then be accessible via the browser:
    • localhost:5001 (or whichever --port you choose)
  • The old version of logviz can be run with logviz --old
  • Run logviz --help for more information about options

Installation

This app can be installed with pipx, pip, or poetry. Instructions are provided for all three.

  • pipx allows you to run logviz from the command line at any time, regardless of the current directory or active virtual environment.
  • pip installation is pretty straightforward.
  • poetry installation is best for developing, since it's easy to poetry add new dependencies and it manages the virtual environment for you.

Pipx instructions

NOTE: I'm not yet sure how easy it is to update the app when installing via pipx. It might be fine, I'll have to test it.

  1. Clone the repository and cd into it
  2. Install pipx
  3. Install logviz with pipx
    • pipx install .

Pip instructions

  1. Clone the repository and cd into it
  2. Make a virtual environment
    • python -m venv .venv
  3. Activate the virtual environment
    • source .venv/bin/activate
  4. Install the webapp
    • pip install .
    • (For developing, run pip install -e . to install in editable mode instead.)

Poetry instructions

  1. Clone the repository and cd into it
  2. Install poetry
    • curl -sSL https://install.python-poetry.org | python3 -
  3. Make and activate a poetry shell
    • poetry shell
  4. Install the project
    • poetry install

Migrating to database-based logviz

The most recent version of logviz uses an sqlite database to manage logs. This is a breaking change, so logs previously added to logviz will not be accessible by default. I've added a migration script:

  1. Run logviz
  2. Run ./scripts/migrate_logs <dir> <port>, where dir is the log directory you used and port is the port you are running on.
    • By default, this would be:
    • ./scripts/migrate_logs ~/.cache/logviz 5001

logviz's People

Contributors

junshern avatar ianmckenzie-oai avatar naimenz avatar

Stargazers

Juan M Menendez avatar Erik Ritter avatar Giulio Starace avatar Dane avatar  avatar

Watchers

 avatar Kostas Georgiou avatar  avatar

Forkers

conglesolutionx

logviz's Issues

Directly open one log file with a command

Is there a way to directly open one view with a log from the console, without the need of manually upload the file on the web?

something like:
logviz ./logs/my_eval_log_gpt-3.5-turbo.jsonl

That would be very handy.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.