Xin Lin

Results 4 comments of Xin Lin

转换为 tensorRT 后，后处理时间增加

> pytorch版本在3090上的后处理时间一般在1-2ms，转为tensorrt后再进行后处理时间增加到几十毫秒 请问作者这是哪里的问题？ 请问你转动态（多batch）tensorrt会报错吗？我只能转静态的

Cannot convert the onnx model into the tensorrt model with dynamic shape

Whatever I try, the batchsize for the onnx model is fixed which is one. ![image](https://user-images.githubusercontent.com/41569547/179932801-1d4a14f3-0b3e-4f74-a522-9e12099eefab.png)

【官方】QQ交流群

这个群不开放了吗 一直请求 没有回应鸭

Can i freeze the class based loss so that i can train only class what i want?

> Hi, FateScript. Can you do me a favor in adding me on the official wechat group?