machine_learning_workbench
Framework including pure-Ruby implementation of both feed-forward and recurrent neural networks (fully connected). Training available using neuroevolution (Natural Evolution Strategies algorithms).
- from
- ML with Ruby
- added
- 2026-10-10
- likes
- 0
similar
-
machine_learning_workbench github.com
Framework including pure-Ruby implementations of Natural Evolution Strategy algorithms (black-box optimization), specifically Exponential NES (XNES), Separable NES (sNES), Block-Diagonal NES (BDNES) and more. Applications include neural network search/training (neuroevolution).
-
-
machine_learning_workbench github.com
Growing machine learning framework written in pure Ruby, high performance computing using Numo, CUDA bindings through Cumo. Currently implementating neural networks, evolutionary strategies, vector quantization, and plenty of examples and utilities.
-
Deep NeuroEvolution github.com
Experimental setup based on the machine_learning_workbench towards searching for deep neural networks (rather than training) using evolutionary algorithms. Applications to the OpenAI Gym using PyCall.
-
neural-net-ruby github.com
Neural network written in Ruby.
-
Machine Learning Ruby github.com
ML with Ruby › Machine Learning Libraries > Neural networks: “Framework including pure-Ruby implementation of both feed-forward and recurrent neural networks (fully connected). Training available using neuroevolution (Natural Evolution Strategies algorithms).”