Privacy terms

Company & Product


Postal Address

Stampfenbachstrasse 52, 

8006 Zürich, Sitzerland

  • 600px-Octicons-mark-github.svg
  • Twitter
  • LinkedIn

© 2020 by WhatToLabel

Data Filtering

for Deep Learning 

Use different filters and parameters to reduce data annotation costs, improve generalization and find outliers in your raw dataset

Supporting computer vision tasks for


Object Detection


Use WhatToLabel to

Save Cost

Use WhatToLabel to save 50% of data annotation costs by filtering data before annotation

Reduce Overfitting

Use WhatToLabel on your annotated training data to remove semantic redundancies which otherwise result in overfitting

Find Novel Samples

Every time you filter a dataset you also get a dataset analysis report. Use the provided information to improve your data collection pipeline

Easy Filtering based on your needs

Our Blogposts

Introducing WhatToLabel

Are you curious about research areas such as active, self-supervised, and semi-supervised learning and how we can optimize datasets rather than optimizing deep learning models? You’re in good company, and this blog post will tell you all about it!

Data Labeling: AI’s Human Bottleneck

Customers are increasingly demanding smart products such as autonomous cars or home assistants. This leads to the expected growth of the AI market to over $100Bn by 2025 (image below). But what does it take to make a product smart?

The Data You Don‘t Need: Removing Redundant Samples

In ML there is the saying garbage in, garbage out. But what does it really mean to have good or bad data? In this post, we will explore data redundancies in the training set of fashion-MNIST and how it affects test set accuracy.