Speed up GridSearchCV using parallel processing

Want your grid search to run faster? Set n_jobs=-1 to use parallel processing with all CPUs!
👉 New tips every TUESDAY and THURSDAY! 👈
🎥 Watch all tips: • scikit-learn tips
🗒️ Code for all tips: github.com/justmarkham/scikit...
💌 Get tips via email: scikit-learn.tips
=== WANT TO GET BETTER AT MACHINE LEARNING? ===
1) LEARN THE FUNDAMENTALS in my intro course (free!): courses.dataschool.io/introdu...
2) BUILD YOUR ML CONFIDENCE in my intermediate course: courses.dataschool.io/buildin...
3) LET'S CONNECT!
- Newsletter: www.dataschool.io/subscribe/
- Twitter: / justmarkham
- Facebook: / datascienceschool
- LinkedIn: / justmarkham

[@dataschool]

I hope this tip helps your work go faster! ⏰ Another way to save time is to use RandomizedSearchCV instead of GridSearchCV. Watch this video to learn more: ua-cam.com/video/Q5dH5mOQ_ik/v-deo.html

[@anefuoche1053]

the truth is the calmness of your voice makes learning exciting too

[@dataschool]

That's great to hear! 🙏

[@philwebb59]

There's more to it than simply setting n_jobs to -1. You also need enough RAM for all those parallel processes to share. Otherwise, you max out your swap space and everything comes to a grinding halt.

[@dataschool]

Thanks Phil! That's why I always recommend timing your process (with and without n_jobs set to -1) to see if it's helping or not. It's hard to know whether you have enough RAM for any given process without just trying it!

[@mdmantejsingh]

Thank you, it's good to know.

[@dataschool]

You're welcome!