The XIR based compiler takes the quantized TensorFlow or Caffe model as the input ... depthwise-conv2d, Kernel size, w, h: [1, 16], w, h: [3], Not supported.
確定! 回上一頁