CMSIS-NN Quantization specific registration for SVDF
int8 input data type registration is split into its own function. The PR is a work towards RFC https://github.com/tensorflow/tflite-micro/blob/main/tensorflow/lite/micro/docs/rfc/002_16x8_quantization_port.md BUG=quantization specific registration for SVDF to reduce library size Change-Id: I0a77d98cd3fa643b530386a6ca59e41fdf2b3ce9
Showing
想要评论请 注册 或 登录