Model Serving Techniques - Learn Data Science with Travis

Travis chat

🍮 Edge#147: MLOPs – Model Serving

TheSequence

In this issue: we explain what model serving is; we explore the TensorFlow serving paper; we cover TorchServe, a super simple serving framework for PyTorch. 💡 ML Concept of the Day: Model Serving Con...

🌀Edge#12: The challenges of Model Serving~

TheSequence

In this issue: we explain the concept of model serving; we review a paper in which Google Research outlined the architecture of a serving pipeline for TensorFlow models; we discuss MLflow, one of the ...

Several Ways for Machine Learning Model Serving (Model as a Service)

Towards AI

No matter how well you build a model, no one knows it if you cannot ship model. However, lots of data scientists want to focus on model building and skipping the rest of the stuff such as data…

Machine-learned model serving at scale

Towards Data Science

Imagine you have a machine-learned model that you would like to use in some application, for instance, a transformer model to generate vector representations from text. You measure the time it takes…

101 For Serving ML Models

Pratik’s Pakodas 🍿

Learn to write robust APIs Me at Spiti Valley in Himachal Pradesh → ML in production series Productionizing NLP Models 10 Useful ML Practices For Python Developers Serving ML Models My love for unders...

Gliding into Model-Based

Towards Data Science

Reinforcement Learning (RL) can be a daunting space to those new to the field due to terminology and complex mathematics formula. However, the principles underlying it are more intuitive than first…

Model serving architectures on Databricks

Marvelous MLOps Substack

Many different components are required to bring machine learning models to production. I believe that machine learning teams should aim to simplify the architecture and minimize the amount of tools th...

Models as Serverless Functions

Towards Data Science

I recently published Chapter 3 of my book-in-progress on leanpub. The goal with this chapter is to empower data scientists to leverage managed services to deploy models to production and own more of…

Models as Web Endpoints

Towards Data Science

In the second chapter of Data Science in Production, I discuss how to set up predictive models as web endpoints. This is a useful skill, because it enables data scientists to shift from batch model…

9. Model persistence

Scikit-learn User Guide

After training a scikit-learn model, it is desirable to have a way to persist the model for future use without having to retrain. The following sections give you some hints on how to persist a scik......

Model Deployment — Conclusion

Becoming Human: Artificial Intelligence Magazine

This is the concluding article of the Model Deployment Series. In this series we saw various techniques which can be used to deployed any ML model. We also touched the upper layer of the CICD via Git…...

Models and databases

Django documentation

A model is the single, definitive source of information about your data. It contains the essential fields and behaviors of the data you’re storing. Generally, each model maps to a single database tabl...

Two is better than one: Ensembling Models

Towards Data Science

Ensembling sounds like a very intimidating word at first but it’s actually deceptively simple….lemme explain ensembling with an analogy So basically ensembling/combining two or more algorithms could…

Serving ML Models with gRPC

Towards Data Science

gRPC APIs are fast, efficient, and type-safe. Next time you need to create an ML prediction service, ditch REST and give gRPC a shot!

Model-Agnostic Methods

Christophm Interpretable Machine Learning Book

Separating the explanations from the machine learning model (= model-agnostic interpretation methods) has some advantages (Ribeiro, Singh, and Guestrin 2016 27 ). The great advantage of model-agnostic...

Model choice

Towards Data Science

If models are cheap, don’t choose. The more models the merrier. But we mustn’t waste our time on rubbish and there’s more to a good model than the accuracy of its predictions. At catastrophic…

Beginner’s guide to Model Deployment

Analytics Vidhya

Are you a beginner in the field of machine learning and wondering how to bring your project to live. Deploy Machine learning models using Flask and Heroku.

Mastering the Many Models Approach

R-bloggers

Intro Setup Fundamentals Extensions Endgame Wrap-up Intro The tidyverse “many models” approach was formally introduced in the first edition of R for Data Science (R4DS) in 2017. Since then, the tidyve...

Parameter Servers

Dive intro Deep Learning Book

As we move from a single GPU to multiple GPUs and then to multiple servers containing multiple GPUs, possibly all spread out across multiple racks and network switches, our algorithms for distributed ...

Demystifying training and serving models on cloud

Analytics Vidhya

Set up, deploy and serve a machine learning model on the cloud(Azure Kubernetes Service)

Polyaxon, Argo and Seldon for model training, package and deployment in Kubernetes

Analytics Vidhya

n it’s simplest form, model management can be seen as training one machine learning model, then repeating this tens, hundreds, or thousands of times with different data, parameters, features and…

Offsetting the Model — Logic to Implementation

Towards Data Science

In our property and casualty insurance world very often we use a term called ‘offset’ which is widely used for modeling rate (count/exposure) such as the number of claims per exposure unit. This…

How Models Work

Kaggle Learn Courses

Introduction We'll start with an overview of how machine learning models work and how they are used. This may feel basic if you've done statistical modeling or machine learning before. Don't worry, w...

Model optimization techniques

Analytics Vidhya

When I used to look at my model sizes once training gets over I always gets frustrated because of the number it shows up. But no more worries as the machine learning world has developed immensely…