Aim CLI

Aim CLI offers a simple interface to easily organize and record your experiments. Paired with the Python Library, Aim is a powerful utility to record, search and compare AI experiments. Here are the set of commands supported:

Command

Description

init

Initialize the aim repository.

version

Displays the version of aim cli currently installed.

up

Runs Aim web UI for the given repo.

upgrade

Upgrades legacy Aim repository from 2.x to 3.0.

reindex

Process runs left in ‘in progress’ state and optimized finished runs.

init

**This step is optional.** Initialize the aim repo to record the experiments.

$ aim init

Creates .aim directory to save the recorded experiments to. Running aim init in an existing repository will prompt the user for re-initialization.

version

Display the Aim version installed.

$ aim version

up

Start the Aim web UI locally.

$ aim up [ARGS]

Args

Description

-h &#124; --host <host>

Specify host address.

-p &#124; --port <port>

Specify port to listen to.

--repo <repo_path>

Path to parent directory of .aim repo. Current working directory by default

--dev

Run UI in development mode.

upgrade

Upgrade Aim repository containing data logged with older version of Aim.

$ aim upgrade [ARGS] SUBCOMMAND

Args

Description

--repo <repo_path>

Path to parent directory of .aim repo. Current working directory by default

upgrade subcommands

Upgrade aim repository from 2.x to 3.0.

$ aim ugrade 2to3 [ARGS]

Args

Description

--skip-failed-runs

Use this flag to skip runs which are failed/have missing or incomplete data.

--skip-checks

Use this flag to skip new repository consistency checks.

--drop-existing

Use this flag to clear old .aim directory. By default old data is kept in .aim_legacy.

reindex

Update index to include all runs in Aim repo which are left in progress.

$ aim reindex [ARGS]

Args

Description

--repo <repo_path>

Path to parent directory of .aim repo. Current working directory by default

--finalize-only

Only finalize runs left in ‘in progress’ state. Do not attempt runs optimization.