Release Notes: v0.0.4

This release focuses on expanding the repository with comprehensive educational resources and research-oriented examples for deploying deep learning models on the ESP32.

Examples and Documentation

• Comprehensive Deployment Tutorial: Added examples/Tutorial.md, an end-to-end guide covering model training, INT8 quantization, and embedded C++ implementation.

• Deep Context Model Architectures: Introduced implementation examples for multiple neural network architectures:

• Multi-Layer Perceptron (MLP): Baseline classification using synthetic datasets.

• Convolutional Neural Networks (CNN): Spatial-temporal feature extraction for time-series windows (Static Context).

• Recurrent Neural Networks (RNN): Unrolled sequential models for stable TFLM deployment (Dynamic Context).

• Quantization Pipeline: Documented the full Post-Training Quantization (PTQ) process, including representative dataset generation and the configuration of INT8 kernels for non-linear activations.

Summary of Changes

All modifications in this version are localized to the examples/ directory and documentation files to provide a clearer path from research to deployment.