Features
- • Convolution
- • Pooling
- • Fully connected
- • Quantization
- • SIMD
- • Optimized kernels
Contingency mode: public pages and downloads remain available; login, community, and server data may be unavailable.
SpecializedLibraryC/C++
ESP-NN
Espressif Neural Network Library
Optimized kernels for neural-network operations, used as an acceleration block for inference.
Industrial applications
- • Anomaly detection
- • Classification
- • Pattern recognition
- • Edge AI
Requirements
- • Runtime/model integration
- • Quantized model
- • Memory and latency measurement
Recommended families: ESP32, S3, P4
Limitations
- • Not a complete AI solution by itself
- • Compatibility must be confirmed by version