hush66 / DNN-Partition-demo

A DNN model partition demo
30 stars 5 forks source link

复现edgent:测量每层的时延及输出数据大小方面(疑问) #3

Open szpIlove opened 4 years ago

szpIlove commented 4 years ago

O73J47_CVM(K)1G`{~@HSKQ 大佬,在复现edgent文章:测量每层的时延及输出数据大小方面,您用了什么样的方法或您有什么好的建议?谢谢您啦~

hush66 commented 4 years ago

输出数据根据模型结构和输入数据不是可以算出来了嘛  每层时延我是自己写脚本记录的

--------------原始邮件-------------- 发件人:"szpIlove "<notifications@github.com>; 发送时间:2020年8月13日(星期四) 晚上8:33 收件人:"hush66/DNN-Partition-demo" <DNN-Partition-demo@noreply.github.com>; 抄送:"Subscribed "<subscribed@noreply.github.com>; 主题:[hush66/DNN-Partition-demo]复现edgent:测量每层的时延及输出数据大小方面(疑问) (#3)

大佬,在复现edgent文章:测量每层的时延及输出数据大小方面,您用了什么样的方法或您有什么好的建议?谢谢您啦~

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

szpIlove commented 4 years ago

谢谢您,因为我目前在用keras+tensorflow复现,所以在这方面还是才疏学浅,您可以分享一下脚本吗?谢谢您的帮助

hush66 commented 4 years ago

哦哦  我之前用pytorch写的  可以交流交流

--------------原始邮件-------------- 发件人:"szpIlove "<notifications@github.com>; 发送时间:2020年8月13日(星期四) 晚上10:38 收件人:"hush66/DNN-Partition-demo" <DNN-Partition-demo@noreply.github.com>; 抄送:"hush66 "<hush26@mail2.sysu.edu.cn>;"Comment "<comment@noreply.github.com>; 主题:Re: [hush66/DNN-Partition-demo] 复现edgent:测量每层的时延及输出数据大小方面(疑问) (#3)

谢谢您,因为我目前在用keras+tensorflow复现,所以在这方面还是才疏学浅,您可以分享一下脚本吗?谢谢您的帮助

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Calithy commented 2 years ago

脚本文件可以分享一下吗?想学习一下