AimStack logo

Supercharged UI for PyTorch Ignite

Run beautiful UI on top of your PyTorch Ignite logs and get powerful run comparison features.
Aim UI interface
Easy to useopen source
Just add the following to your PyTorch Ignite
training config file

[training.logger]
@loggers = "spacy.AimLogger.v1"
repo = "path/to/your/save/directory"
experiment_name = "name_of_your_experiment"
Compare runs easily
to build models faster
7 major features to speed up your work

1.

Powerful pythonic search to select the runs you want to analyze.

Live demo this feature

2.

Group metrics by hyperparameters to analyze hyperparameters’ influence on run performance.

Live demo this feature

3.

Select multiple metrics and analyze them side by side.

Live demo this feature

4.

Aggregate metrics by std.dev, std.err, conf.interval.

Live demo this feature

5.

Align x axis by any other metric.

Live demo this feature

6.

Scatter plots to learn correlations and trends.

Live demo this feature

7.

High dimensional data visualization via parallel coordinate plot.

Live demo this feature
Extend tracking via Aim’s SDK
Track and visualize images, audios, text, plotly and matplotib figures
Aimlflow brings the power of Aim Explorers to MLflow so you can use MLflow and compare lots of runs. Sign up for early access.
Oops! Something went wrong while submitting the form.
Schedule