Yuva AI
Synthetic Data for Computer Vision AI.
LABELLING INSTRUCTIONS
Best Practices
USE-CASE
- Specify the use-case (eg. named entity recognition)
- Specify the desired output (eg. all entities identified)
- Give examples of good/bad annotations
class definition
Be very specific and define the class clearly. To do so:
List all the class types
Give at least 3 examples of the class
Give most common counter examples or errors
edge cases
Define common edge cases
Define how you would like us to deal with edge cases. For example, should the labeller defer to the manager or should they not annotate that case.
examples
The more examples you can provide the better our annotators can understand how to label to meet your standards.
Create Your Labelling Instructions
Please login to submit the post.
Login here
Creating a better future, one piece of data at a time.
Address : 200 Adelaide St, Brisbane City QLD 4000, Australia
Phone:+61 424 466 454
Email: yash@yuva.ai