Evaluating models without test data (Practical AI #194)
Changelog Master Feed
00:00
XGBoose - Should I Add More Features or Should I Tune More?
When we started this project, there were maybe 50 open source pre-trained models. You go to a hugging face now, there are over 50,000. It's really hard to know which model to pick. These models are very hard to improve. We're 20 years behind where software engineering is.
Transcript
Play full episode
