Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Issues with Transfer Learning for Default Labeler #1155

Open
DylanVig opened this issue Jun 25, 2024 · 0 comments
Open

Issues with Transfer Learning for Default Labeler #1155

DylanVig opened this issue Jun 25, 2024 · 0 comments
Assignees

Comments

@DylanVig
Copy link

I attempted to create my own custom labeler by using the transfer learning example in the documentation. I attempted to add three labels to the labeler: Name, Datetime (which is already a category in the default labeler), and Nationality. This is my code for training my labeler, as well as the csv I used to train it.
large_fake_data.csv
Screenshot 2024-06-25 at 3 06 44 PM

I attempted to use this custom labeler on a csv that had 6 columns: Name, Datetime, Phone Number, SSN, Email, and Nationality. When I ran it with the custom labeler, it correctly identified the names, datetimes, and nationalities. However, it also falsely identified the phone numbers, SSNs, and emails incorrectly (usually identifying email and SSN and nationality and phone number as datetime). When I run it with the default labeler, it seems to pick up on those three fields just fine. Is there a problem with how I am programming my labeler, how I'm training it, etc? Here is my code for testing my labeler, as well as the csv I used to get these results:
three_cat_labeler_test_data.csv
Screenshot 2024-06-25 at 3 11 02 PM

Thank you!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

5 participants