I am using the nn_quantizer.py file provided by ARM on Github (cifar10 example) for quantization to 8 bits. For my dataset it requires the Q2.5 ...
確定! 回上一頁