optuna-dashboard

Web dashboard for Optuna. Code files were originally taken from Goptuna.

Usage

You can install optuna-dashboard via pip.

$ pip install optuna-dashboard

After you installed, optuna-dashboard command is available. Please execute it with Optuna storage URL.

$ optuna-dashboard sqlite:///db.sqlite3
Bottle v0.12.18 server starting up (using WSGIRefServer())...
Listening on http://localhost:8080/
Hit Ctrl-C to quit.

More command line options

$ optuna-dashboard --help
usage: optuna-dashboard [-h] [--port PORT] [--host HOST] [--quiet] storage

A third-party dashboard for optuna.

positional arguments:
  storage      Optuna Storage URL

optional arguments:
  -h, --help   show this help message and exit
  --port PORT  port number (default: 8080)
  --host HOST  hostname (default: 'localhost')
  --quiet      quiet

Features

Interactive realtime graphs

You can check graphs of optimization history, parallel coordinate and intermediate values.

Rich trials data grid

You can walk-through trials with filtering and sorting.

Create and delete study

You can create and delete a study from Dashboard.

Alternatives

optdash: a third-party dashboard for optuna.

1.7 KiB Raw Blame History