Models Table

This page describes the Trains Web-App Profile page, Models tab, models table which is a list of models in a project or all projects. The Overview section of this page provides details about the information in the models table, and the Features section highlights features and actions you can perform using the models table.


List of models

View a list of all models in a project or all projects by selecting a recent project or all projects from the Home page or Projects pages. The list shows the following default columns:

  • MODEL - Model name. Use this field to sort the list.
  • TAGS - Tags which are added to models to group and organize them, as well as indicate the type of trained model. For example, tags include: packages (model is a zip or tar files that must be unzipped and can also contain additional network configuration information in the zip), OpenImages (trained models), Boxes (trained models, such as COCO trained models), and TF_ModelZoo (trained models).
  • CREATED - Elapsed time since the model was created. Hover over the elapsed time to view the date and time.
  • FRAMEWORK - The model framework, such as Caffe, TensorFlow, PyTorch, Keras, Custom. Use this field to sort and filter the list.
  • PUBLISHED - The model status which can be Published (read-only) or Draft (editable). Use this field to sort and filter the list.
  • USER - The user name. This field is neither sortable nor a filter.
  • TASK - The Task name which is the name for internal Trains representation and Python experiment script Task of the experiment creating the model. This field is neither sortable nor a filter.

Additionally, to organize your current, active work more easily, Trains provides an archive for models (and experiments). Models in the archive only appear when you open the archive, not when you open the Projects page in the "active" models table. You can restore archived models from the archive to the "active" experiments table.


View a model

  • In the models table, click the model. The model details panel opens.

Manage models

The Trains Web-App supports features allowing you to manage your models. These features are available in the model context menu, and the models details panel (see Manage models on the Model Details page).

Model context menu

The model context menu, which appears by right clicking a model, shows the options you can perform on the model.

To manage a model (perform a model action), do the following:

On the models table, right click a model and then on the model context menu click one of the following options:

  • Publish - Publish a model to make it read-only and prevent changes. If the model is already Published, then this option is disabled.
  • Archive - Archive a model to make it visible in the Archive, only, and better organize your current work. Archived models can be restored. See the Archive section below.
  • Download - Download the model as compressed file.
  • Move To Project - Change the model's project to better organize your work.