Navigates and documents the NeuroScript standard library for composing complex neural architectures.
The NeuroScript Standard Library (ns-stdlib) skill provides a comprehensive, live-queried catalog of primitive and composite neurons within the NeuroScript ecosystem. It acts as an intelligent reference guide for developers building neural models, offering detailed signatures, parameter requirements, and structural patterns. Whether you are implementing high-level Transformer blocks or fine-tuning low-level activation functions, this skill facilitates rapid architectural decision-making by providing categorized indices and logic-based recommendations for choosing the right components.
주요 기능
01Categorized index covering activations, normalization, convolution, and attention
02Documentation for composite library neurons like TransformerStacks and FFNs
03Decision tree logic for selecting optimal neurons based on architectural needs
04Live lookup of primitive neurons with full signatures and parameters
05Guidance on structural routing patterns and implicit fork/concat operations
060 GitHub stars
사용 사례
01Selecting appropriate normalization or activation functions for specific model types
02Verifying tensor shapes and layer configurations during model development
03Building and debugging complex neural networks using NeuroScript DSL
