{% extends shell('app') %} {% block title %}Generate feature models{% endblock %} {% block head %} {% endblock %} {% block breadcrumb %} {% endblock %} {% block content %}

Generate feature models

Pick how you want to build your feature models. Both generators produce UVL files ready to download or upload to the hub.
{{ icon('cpu', size=24) }}

Random generator

Builds random feature models from the numbers you choose, such as how many features, how deep the tree and which constraints. It runs in your browser, and you download the models as a ZIP.

Open the random generator {{ icon('arrow-right', size=24, class_='align-middle') }}
{{ icon('message-square', size=24) }}

Language model generator

Experimental

Writes feature models from an idea you describe, or from a model of the hub, with an open language model that runs in your browser. Each one is checked before you see it, and nothing leaves your computer.

{#- What it takes, said before entering (UXA-41). -#}

{{ icon('info', size=14, class_='me-1') }}It needs a recent Chrome or Edge with WebGPU, and downloads a model of 1 to 5 GB the first time.

{{ icon('triangle-alert', size=14, class_='me-1') }}This browser has no WebGPU, so this generator cannot run here.

Open the language model generator {{ icon('arrow-right', size=24, class_='align-middle') }}
{% endblock %} {% block scripts %} {# Icons, which the generator draws itself (see ui.js). #} {{ feature_script('generator', 'ui') }} {#- Whether this browser can run the language model generator at all. -#} {% endblock %}