harthur/brainArchived
Brain
Brain is a JavaScript library for building, training, and running feed-forward neural networks. It implements a multilayer perceptron model designed for pattern recognition and function approximation.
The library includes a standalone inference engine that converts trained models into portable JavaScript functions. This allows predictions to be executed in browser or Node.js environments without requiring the original library dependencies.
The system supports persistent model management through JSON serialization for saving and loading network weights. It also provides a streaming mechanism for incremental training to process large datasets.
Features
- Feed-Forward Neural Networks - Implements a feed-forward multilayer perceptron for pattern recognition and function approximation.
- Client-Side Inference - Enables predictions to be executed in browser or Node.js environments without requiring original library dependencies.
- Inference Execution - Computes predictions for given inputs based on the learned weights of a trained network.
- Model-to-Code Generators - Converts a trained model into a portable JavaScript function for predictions without dependencies.
- Functional Model Exports - Generates a standalone JavaScript function that encodes trained weights directly into the code for dependency-free inference.
- Training Weight Adjustments - Adjusts internal connection strengths during training to minimize the difference between predicted and expected outputs.
- Neural Network Implementations - Provides a core implementation of feed-forward neural networks for pattern recognition within JavaScript.
- Neural Network Training - Adjusts network weights using input-output patterns to enable classification and pattern approximation.
- Neural Networks - Implements a multilayer perceptron model designed for function approximation and pattern recognition.
- Pattern Recognition - Teaches a model to classify data or approximate functions by processing sets of input and output patterns.
- Neural Network Libraries - Provides a complete JavaScript library for building, training, and running feed-forward neural networks.
- Portable Inference Functions - Generates a portable JavaScript function from a trained network for inference without the original codebase.
- Machine Learning Model Formats - Uses a JSON-serializable format for saving and loading trained weights to ensure model persistence.
- Incremental Training - Processes datasets in small chunks via streaming to train models without exhausting system memory.
- Model State Serializers - Provides the ability to serialize trained model weights and biases into JSON for persistent storage.
- Large Dataset Streaming - Reads training patterns from a write stream to process large datasets without exhausting system memory.
- Training Sample Streaming - Feeds training patterns incrementally into the network via a write stream to handle large datasets.
- Neural Network Checkpointing - Saves and loads trained model weights via JSON to avoid retraining the network from scratch.
- AI & Machine Learning - Neural network library for JavaScript.
- Deep Learning Frameworks - Simple neural network library for JavaScript.
- General Machine Learning - Neural network implementation for JavaScript.
- Machine Learning - Neural network implementation for JavaScript environments.
- Machine Learning and Vision - Neural network implementation.
