quantization aware trainingの論文「Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference」を読みました
MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']], displayMath: [ ['$$','$$'], ["\\[","\\]"] ] } }); はじめに CTOの幅野です。 Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferenceを読みました。 この論文はTensorflow/Tensorflow Liteで実装さ…
