CrowdTruth/CrowdTruth-core

CrowdTruth framework for crowdsourcing ground truth for training & evaluation of AI systems

40
/ 100
Emerging

When building an AI system, you often need to gather labeled data from many people to train it. This tool helps researchers and AI practitioners process the results from crowdsourcing platforms like Amazon Mechanical Turk and CrowdFlower. It takes the raw crowdsourcing data and applies a specialized methodology to determine the most reliable "ground truth" labels, even when crowd workers disagree.

No commits in the last 6 months.

Use this if you are developing or evaluating AI systems and need to derive high-quality, reliable labeled datasets from potentially noisy or inconsistent crowdsourced annotations.

Not ideal if you are looking for a platform to run crowdsourcing tasks or a general-purpose data labeling tool, as this focuses specifically on processing and evaluating existing crowdsourcing results.

AI training data machine learning evaluation crowdsourcing quality control data annotation ground truth generation
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 16 / 25

How are scores calculated?

Stars

64

Forks

11

Language

Jupyter Notebook

License

Apache-2.0

Last pushed

Apr 08, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/CrowdTruth/CrowdTruth-core"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.