TensorFlow Lite, a low latency, smaller footprint inference engine, uses the Eigen library and techniques such as pre-fused activations and quantized ...
確定! 回上一頁