TVM简介: https://blog.csdn.net/tlzhatao/article/details/93630910
TVM: An Automated End-to-End Optimizing Compiler for Deep Learning
知乎 https://www.zhihu.com/topic/20179059/hot
Apache TVM 是一个开放源代码的机器学习编译器框架,用于 CPU,GPU 和机器学习加速器。它旨在使机器学习工程师能够在任何硬件后端上高效地优化和运行计算。 TVM 提供以下主要功能: 尽量少地将深度学习模型编译为可部署的模块。 后端以更好的性能自动生成和优化模型的基础架构。
Open Deep Learning Compiler Stack Documentation |Contributors |Community |Release Notes Apache TVM is a compiler stack for deep learning systems. It is designed to close the gap between theproductivit