May 05, 2022

Tutorial on Recognizing Handwritten Digits with Elixir ML

122

clicks

Tutorial on Recognizing Handwritten Digits with Elixir ML

Source: fly.io

Machine Learning Nx Phoenix LiveView

Philip Brown demonstrates the power of Elixir for building a machine learning application from scratch. The tutorial serves as an end-to-end guide, starting with setting up the Phoenix framework, preparing the project for machine learning with libraries like Nx and Axon, and processing training data from the MNIST database. Key aspects involve transforming images and labels into a suitable format for the model, defining the model's structure, and training the model with the categorical cross-entropy loss function and accuracy metric. The application provides a LiveView frontend, where users can draw digits, and the system predicts the handwritten numbers using the trained model. Finally, the process includes testing, saving, and loading the model state. Predictions are displayed interactively, leveraging the power of Elixir's live updates without needing external machine learning tools. The full code is available on GitHub.

Continue reading

Related posts

Building a Note-Taking Application with Elixir's LiveView and GenServer

In this blog post, AbulAsar explains how to build a simple notetaking app using Liveview and GenServer in two parts. In part 1, AbulAsar focuses on generating the application, adding an API layer, and creating the GenServer process to manage the application state.

Nov 01, 2021

Understanding Phoenix LiveView as an Elixir Process

Jason Stiebs from Fly.io explains how Elixir Processes work in Phoenix LiveViews. He emphasizes the importance of understanding that a LiveView is a process and discusses the benefits and implications of using processes in building scalable applications.

Jun 14, 2023

Integrating Dialyzer into Existing Elixir Projects Smoothly

This blog post by Noah Betzen discusses how to successfully implement Dialyzer, a static code analysis tool, in an existing Elixir project without causing inconvenience to the team. The post includes tips on ignoring existing errors, preventing new errors, fixing errors, and integrating Dialyzer into the CI pipeline to catch type errors before merging code.

Jun 12, 2023

© HashMerge 2024