Accessing ActiveTigger
ActiveTigger is a web application designed to be deployed on a server, but it can also run locally on your machine (mind hardware requirements, including GPU).
There are two ways to get started:
- Get an account on an existing instance
- Install it locally
Option 1: Use an existing instance
Existing instance
- CREST development instance
CREST development instance
A development instance of ActiveTigger is running on the Groupe ENSAE-ENSAI servers (with GPU support), accessible here.
Development instance
This instance is primarily intended for testing and reporting bugs. Service continuity is not guaranteed. We reserve the right to modify, restrict, or revoke access at any time.
Who can request access?
- Researchers who have textual data to annotate
- Open source contributors who want to help improve the software
How to get access
- Fill this form.
- Once your account is created, you can:
- Upload data and annotate text
- Train classification models
- Report bugs or suggest features via GitHub issues
Account policy
Accounts without any activity may be deactivated after 2 months.
How to get in touch
- GitHub issues — for bug reports and feature suggestions
- Discord — for questions and discussions
User agreement
Terms of use
- Access to ActiveTigger is currently temporary
- Data storage and security are not guaranteed — do not upload sensitive data
- The service is hosted on ENSAE servers and intended solely for research purposes
- We reserve the right to modify, restrict, or revoke access at any time without notice
Option 2: Install locally
You can deploy your own instance of ActiveTigger on your machine or on a cloud server (e.g., OVH).
GPU recommended
ActiveTigger can run on CPU, but a GPU will significantly speed up model training and predictions, especially with large models (BERT and sentence-bert).
For installation instructions, see the deployment guide.