How to evaluate and select ML models + other resources
Here are the best MLOps articles, case studies, and interviews that we published (or refreshed or came across) this month. Hope you'll find something interesting here.?
---
Case studies & practical MLOps
>?Managing Computer Vision Projects ?- If you work in the Computer Vision space, check this Q&A with Michal Tadeusiak , Director of AI at deepsense.ai. He's a very experienced practitioner combining technical savvy and managerial skills in the field of data analytics and ML.?
>?Time Series Forecasting Example Project ?- Here's an example project we prepared at neptune.ai to showcase how people can use Neptune in time series projects. We showcase the web app, but there's also a link to a GitHub repo, so you can dig into the code as well.?
One of the example dashboards |?Source
>?SwirlAI Newsletter:?The Data Value Chain, Data Contracts in the Data Pipeline, The 4 types of ML Model Deployment. ?- Another solid edition of the SwirlAI newsletter run by Aurimas Griciūnas , Neptune's Solutions Architect. Definitely worth following.
---
Guides & tutorials
>?The Ultimate Guide to Evaluation and Selection of Models in Machine Learning ?- We recently reviewed and updated this one, s make sure to check it. We expanded the section on how to evaluate models and added some suggested further reading.?
>?Distributed Training: Errors to Avoid ?- If you train models in a distributed environment, here's a post for you. You'll find there 10?of the most common errors in distributed model training?and suggested solutions to each of them.
>?Training Models on Streaming Data [Practical Guide] ?- And here, something for people who want to learn more about using machine learning for streaming data analysis (hands-on example included).?
---
领英推荐
Tools
---
Q&As with ML practitioners
In March, we have two?MLOps Live Q&As ?planned! We'll talk about:
>?Navigating organizational barriers by doing MLOps?with Leanne Kim Fitzpatrick ,?Director of Data Science at Financial Times?- March 14, 2023
>?Tackling MLOps challenges in computer vision?with Marcin Tuszyński ,?Data Scientist w ReSpo.Vision - March 28, 2023?
You can?register here ?to have all of them on the calendar. Or you can just watch them live on?Neptune's Linkedin profile .?Also, to catch up with previous episodes, go to?YouTube ,?Spotify , or?Apple Podcasts .?
---
Okay, that's it for today. If you want to talk about these recommendations, send me an email or?join the MLOps community here , and find the?#neptune -ai channel?there.
Feel free to forward this newsletter to your friends and communities if you find it useful!