ESP-NN release v1.0 Beta
Supported/Tested Espressif Chipsets:
- ESP32-S3 (Assembly optimizations)
- ESP32 (Generic optimizations)
- ESP32-C3 (Generic optimizations)
ESP-NN is tested to work with following ML framework:
- TFlite-Micro
This release provides optimizations for the following kernels (layers) from TFlite-Micro:
- Convolution
- Depthwise Convolution
- Fully Connected
- MaxPool, AvgPool
- SoftMax
- Add
- Mul
- Relu
Please find the details on how to use ESP-NN, and other useful info here.