July 24, 2023

From Python to Elixir Machine Learning

125

clicks

Source: thestackcanary.com

The author, Andrés Alejos, details the process of transitioning from machine learning in Python to utilizing the growing ecosystem of machine learning libraries in Elixir. He posits that the concurrency support and distributed system capabilities of Elixir make it suitable for ML applications, despite Python's historical dominance in the field. Andrés shares his experience in porting two libraries, EXGBoost and Mockingjay, originally implemented in Python, to Elixir counterparts, highlighting Elixir-Nx's 'Nx.Serving' as a viable distributed model serving solution. He provides an overview of his workflow in the porting process, including understanding the macro system, reading documentation, studying source code in detail, and finally, implementing new code. Through examples, Andrés shows the transition of code idioms from Python's object-oriented approach to Elixir's functional and pattern-based paradigms, demonstrating Elixir's capabilities with the Nx library for tensor computations and strategies like GEMM, Tree Traversal, and Perfect Tree Traversal. He concludes by encouraging Elixir developers to take on the challenge of bringing machine learning into their ecosystem, citing its potential and the support provided by the community.

Simplifying Recursion with Elixir's Stream Module

In his blog post, James Edward Gray II discusses the challenges of using recursion in coding and introduces a simpler alternative using the Stream module in Elixir. He demonstrates how to convert a Fibonacci sequence generator from a recursive approach to a stream-based approach, and highlights the benefits of using Stream to control the flow of iteration and generate potentially infinite sequences of data.

Jul 23, 2023

Supavisor: A Postgres Connection Pooler for Huge Scale

Supavisor is a scalable, cloud-native Postgres connection pooler written in Elixir and developed by Supabase. It can handle millions of connections without significant overhead or latency, making it the optimal solution for developers who want to create applications that can effortlessly scale under extreme workloads.

Aug 11, 2023

Insights on Using Elixir and Phoenix for Commercial Projects

Alex Korban shares his thoughts on using Elixir, Phoenix, and LiveView for a commercial project over the past 18 months. He finds Elixir enjoyable with a good combination of functional programming and concurrency, and while there are some areas that could be improved, both Elixir and Phoenix are solid tools with LiveView being an impressive addition for more complex applications.

Aug 11, 2023