如何固化tensorRT engine,我不想每次启动进程重新转换engine #12653
Unanswered
wangxu372848892
asked this question in
Q&A
Replies: 1 comment
-
哦 木有人回答 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
我用[(https://hub.docker.com/r/paddlepaddle/paddle/tags/)]中的paddle:2.5.2-gpu-cuda12.0-cudnn8.9-trt8.6 容器中成功编译了2.5版本的Inference库(gpu+trt),并拉取paddleocr直接编译了C++版本的demo.我在arg.cpp下没找到可以固化cache trt engine的参数,按下面命令运行时每次都会重新转换engine,有时候会好几分钟,非常浪费时间
系统环境/System Environment:百度官方提供的ubuntu20.04 docker-gpu:2.5.2-gpu-cuda12.0-cudnn8.9-trt8.6
版本号/Version:Paddle:inference推理库版本2.5 、 PaddleOCR版本: release/2.7
运行指令/Command Code:
Beta Was this translation helpful? Give feedback.
All reactions