:probabl.
@probabl_ai
Open-source data science and machine learning
With skore v0.10, you now have a data accessor in the EstimatorReport! It consists in a @skrub_data TableReport that allows you to interactively explore your data and gain precious insights before your modelling! 🎬 Check out our short demo video: eu1.hubs.ly/H0mhFMN0
youtube.com
YouTube
Introducing the data accessor of the EstimatorReport
🚀 Save time on ML evaluations with skore! 🚀 No more recomputing predictions for each metric. Skore caches predictions, enabling instant metric calculations and fast plots. Check out this example for more information: eu1.hubs.ly/H0md0-f0
☀️ It’s summer time! Isn’t it a great moment to take a step back, and learn new things? Check out our whiteboard videos to learn about the quantile trick, why tree gradients give you a boost, or the optimizer curse, etc: eu1.hubs.ly/H0m4QqP0
What is the difference between good and bad data scientists? Good ones know the data they work with. Perfectly understanding the data can be time-consuming. Yet, some basic analysis can pave the way for high-value insights. Now in skore’s release 0.10, with a data accessor!
Born from the scikit-learn initiative, Skore is designed to support not just scikit-learn models but a wide array of machine learning models, including foundational models, such as TabICL, a tabular foundation model! Check our full example: eu1.hubs.ly/H0lXszB0
“Let’s improve this model in production !” Skore can help you to know how the release of a new model will impact the business with its objects EstimatorReport and ComparisonReport. ⭐ eu1.hubs.ly/H0lz0c30 📜 eu1.hubs.ly/H0lz1BD0
Last week, the team organized a workshop for Saint-Gobain. We presented skore and skrub in the morning, and did an open source sprint in the afternoon. Thank you Mojdeh Rastgoo for preparing everyone! Check skore and skrub's good first issues to participate too!
Launching Skolar: a new platform for hands-on, structured and certified training in open-source data science. eu1.hubs.ly/H0lk5kY0
🖌️ There is that plot provided by skore that you like. It quickly gives an idea of the performances. You would like to see the underlying numbers to investigate further. 🔍 That’s why we created the Displays, that can either return a frame, or a plot. 📚 eu1.hubs.ly/H0lfgz60
Did you ever felt lost among all the experiments tried in the model development phase? Skore has your back! 🎉 Check out the demo video of our new release with the search feature: eu1.hubs.ly/H0kZ0cK0 🔗 GitHub repo: eu1.hubs.ly/H0kZ0vv0 📘 Docs: eu1.hubs.ly/H0kZ00F0
⌛ Enriching our metrics! In the new version 0.9 of Skore, we added some metrics: you can now access the confusion matrix! Just a couple of lines, everything easily accessible from the same object. To know more about it, you can check the docs: eu1.hubs.ly/H0kVC040
We have an event every week these days! 🍕 After the machine learning challenge, we organized a sprint with Unaite to onboard students to open source. Open source is a great way to showcase their skills, as would a personal project!
🎤 Next week, our product engineer Marie Sacksick will be presenting how to extend scikit-learn with skore, but also with skrub and skops. Thanks Pyladies Paris for this opportunity! To book your seat: eu1.hubs.ly/H0j-Qpq0
United States Tendencias
- 1. Wemby 37.7K posts
- 2. Steph 77.3K posts
- 3. Spurs 33.3K posts
- 4. Draymond 16.7K posts
- 5. Warriors 56.3K posts
- 6. Clemson 11.3K posts
- 7. Louisville 11K posts
- 8. Zack Ryder 16.7K posts
- 9. #Truedtac5GXWilliamEst 95K posts
- 10. Massie 58.6K posts
- 11. #DubNation 2,133 posts
- 12. WILLIAMEST CALLME MELODY 95.7K posts
- 13. Aaron Fox 2,476 posts
- 14. #SmackDown 53.5K posts
- 15. Harden 15.4K posts
- 16. Bill Clinton 194K posts
- 17. PERTHSANTA JOY KAMUTEA 394K posts
- 18. #NEWKAMUEVENTxPerthSanta 388K posts
- 19. Marjorie Taylor Greene 49.7K posts
- 20. Bubba 59.1K posts
Something went wrong.
Something went wrong.