PerceptiLabs
Search…
PerceptiLabs
v0.13
Welcome
⚡
Getting Started
Quickstart Guide
Load and Pre-process Data
Build Models
Train Models
Evaluate Models
Export and Deploy Models
Manage, Version, and Track Models
🙏
How to Contribute
Datasets
Models
Components
👨🍳
Tutorials
Basic Image Recognition
Basic Image Segmentation
🛠
Advanced
Common Testing Issues
Common Training Issues
Components
CSV File Format
Debugging and Diagnostic Features
How PerceptiLabs Works With TensorFlow
Included Packages
Types of Tests
UI Overview
Data Wizard
Overview Screen
Model Training Settings
Modeling Tool
Training View
Evaluate View
Deploy View
Using the Exported/Deployed Model
💡
Use Cases
General
Environmental
Healthcare & Medical
Industrial IoT & Manufacturing
📖
Support
FAQs
Changelog
Code of Conduct
Marketing Site
Powered By GitBook
Modeling Tool
The Modeling Tool encompasses the workspace where you edit your models. This includes the toolbar, Component dropdowns, Components Pane, and the Code Editor.

Modeling Tool Toolbar

Key elements of the Modeling Tool's toolbar.
Key elements of the Modeling Tool Toolbar:
  1. 1.
    Undo/redo: performs an undo or redo of the last action performed.
  2. 2.
    Components: categories of Components that you can drag and drop onto your model.
  3. 3.
    Save: saves the changes for the current model.
  4. 4.
    Run: displays the Model Training Settings popup where you can customize the training settings to use prior to starting the model training process. Note: you can return to the Modeling Tool during training and run training for another model in parallel.
  5. 5.
    Data Settings: allows you to modify settings for your data that were originally configured in the Data Wizard.
  6. 6.
    Weights: Load and use your most recent checkpoint.
  7. 7.
    Preview: toggles previews in the Modeling Tool on or off. Previews show what the transformed data looks like for each component.
  8. 8.
    Settings: toggles display of the Components and Settings panes.
Note: When you run your model after having previously trained it, you may be presented with a popup that says Start training from latest epoch?. Clicking Yes means that PerceptiLabs will continue the training from where you last stopped it. For example, if you had to stop the training for any reason you can easily pick it back up. Or if you notice that the number of epochs that you trained on were not enough, you can continue training it a bit longer.

Components

The Component tabs above the workspace categorizes all of the components that you can drag and drop onto your model.
Component Dropdowns
See Components for a description of each component.

Components and Settings Panes

The Components pane on the right-hand side of PerceptiLabs lists each Component that has been added to your model. You can click on each Component listed to quickly locate and highlight that Component in the Modeling Tool.
Below that is the Settings pane that provides the settings for the currently-selected component. If the Preview button is enabled, a Preview pane will be shown at the bottom corresponding to the selected component:
​
​
​

Code Editor

Click Open Code in the Settings pane to show the Code Editor which displays the code for the currently selected component in your model. Using this editor you can modify the default code that was generated by PerceptiLabs for that component:
The Open code button on the right opens a component's code.
For more information about the code, see: How PerceptiLabs Works with TensorFlow.

​

Previous
Model Training Settings
Next
Training View
Last modified 10mo ago
Copy link
On this page
Modeling Tool Toolbar
Components
Components and Settings Panes
Code Editor