Pipelines for convenience, *and* safety

  Рет қаралды 635

:probabl.

:probabl.

Күн бұрын

There are many reasons to appreciate pipelines in scikit-learn. They are relatively easy to declare, they are expressive ... but they are also typically safer! This is an under appreciated aspect of pipelines which is why this video is all about exploring what might go wrong if you don't use pipelines.
00:00 Introduction
00:47 Notebook
03:17 pd.get_dummies
05:17 First Error
07:00 Production
12:13 Pipelines!
The code for this video can be found on our youtube appendix:
github.com/probabl-ai/youtube...

Пікірлер: 3
@nicop175
@nicop175 2 ай бұрын
I have been fighting with the team the whole 2023 about the usage of pipelines and columntransformers. Happily they finally agree. Your series would have helped a lot to try and convince them.
@probabl_ai
@probabl_ai 2 ай бұрын
That was a very satisfying message to read. Thanks for that!
@jonathanjayes
@jonathanjayes 2 ай бұрын
Thank you - this was very informative - big ups pipelines!
Scaling Datasets in Pipelines
14:35
:probabl.
Рет қаралды 715
Durable Workflows with Effect Cluster by Mattia Manzati (Effect Days 2024)
26:25
Effect | TypeScript at Scale
Рет қаралды 1,2 М.
Did you find it?! 🤔✨✍️ #funnyart
00:11
Artistomg
Рет қаралды 125 МЛН
КАРМАНЧИК 2 СЕЗОН 5 СЕРИЯ
27:21
Inter Production
Рет қаралды 597 М.
How to Set Up Nextcloud on Ubuntu 24.04 LTS
41:20
Learn Linux TV
Рет қаралды 9 М.
5 Tips For Object-Oriented Programming Done Well - In Python
16:08
Rest API - Best Practices - Design
15:50
High-Performance Programming
Рет қаралды 95 М.
Rethinking the Pipelines API
51:51
:probabl.
Рет қаралды 460
Introduction to the Prometheus Monitoring System | Key Concepts and Features
10:38
Prometheus Monitoring with Julius | PromLabs
Рет қаралды 42 М.