Open Issues Need Help
View All on GitHubAI Summary: This issue proposes adding an 'epsilon' parameter to the `EvaluationConfig` class with a default value. The goal is to centralize epsilon usage and remove hardcoded epsilon values from the `ErrorCalculation` class, making the configuration more manageable and consistent.
A simple, from-scratch C++ Feedforward Neural Network with zero dependencies, built to demystify the fundamentals of deep learning.
AI Summary: This issue proposes to move the activation momentum setting from a direct function call to the constructor, making it a configurable parameter. It suggests a static helper function to estimate momentum if not explicitly provided, improving user experience and code organization.
A simple, from-scratch C++ Feedforward Neural Network with zero dependencies, built to demystify the fundamentals of deep learning.
A simple, from-scratch C++ Feedforward Neural Network with zero dependencies, built to demystify the fundamentals of deep learning.
A simple, from-scratch C++ Feedforward Neural Network with zero dependencies, built to demystify the fundamentals of deep learning.