Customize annotation and training data tooling.
Diffgram is the most customizable software in this space. The following items can be customized:
- Look, feel, and branding of the UI
- Processes for Human Tasks
- Pipelines for data flow
- Hardware configuration,
- The software itself through open source pull requests. can be customized.
Diffgram is the most customizable training data platform.
Workflows (Human and MLOps)
White Label Customization
White Label UI Customization Introduction
Get the exact branded interface you want. Don't want an annotator seeing a certain button? Want to embed Diffgram in your software or process? Control it all with Diffgram.
With Diffgram, using a What You See is What You Get Interface you can easily customize core visual aspects of the application itself. What buttons are shown to annotators, layouts and spacing, what the buttons look like etc.
Most off the shelf options cover less than 2% of the potential annotation automation space.
How do you make automations work for you?
Use Diffgram to easily write your own automations and run them right in Diffgram.
Interactive Annotation Automation
The labels and attributes are one of the most common customizations. You can build your annotation schema and interface to your exact needs.
Create Your First Label Template
Construct custom data access layers using custom queries. Includes Boolean logic and more.
Access Data - Queries Overview
Do you have recurring spatial types or want to define your own?
You can even create customized spatial types:
Custom Instance Templates
Because diffgram is open source you can even inspect and customize the code. Your team can directly implement features and optionally propose them as pull requests to become part of the mainstream.
Generate custom reports and pin them to your custom dashboard. Customize per project and more.
Diffgram automatically chooses the interface type (image, video, 3D etc) based on the file type. You can still customize layouts, sizing etc. within this context.
You can run diffgram how you see fit on your hardware.
Updated 18 days ago