yolov8n-seg模型的int8转化报错

[Killuaisaack114514]

Env

• GPU, RTX 3090
• OS, Ubuntu22.04
• Cuda 12.2
• TensorRT version 8.6.1.6

About this repo

• which branch/tag/commit are you using?
• master-branch
• which model?
• yolov8n-seg

Your problem

• what is your command?
  ./yolov8_seg -s yolov8n.wts yolov8n_int8.engine n
• what's your output?
  [06/28/2024-09:34:46] [E] [TRT] 10: Could not find any implementation for node (Unnamed Layer* 351) [Convolution] + (Unnamed Layer* 352) [Scale] + PWN(PWN((Unnamed Layer* 353) [Activation]), (Unnamed Layer* 354) [ElementWise]).
  [06/28/2024-09:34:46] [E] [TRT] 10: [optimizer.cpp::computeCosts::3869] Error Code 10: Internal Error (Could not find any implementation for node (Unnamed Layer* 351) [Convolution] + (Unnamed Layer* 352) [Scale] + PWN(PWN((Unnamed Layer* 353) [Activation]), (Unnamed Layer* 354) [ElementWise]).)
  void serialize_engine(std::string&, std::string&, std::string&, float&, float&, int&): Assertion `serialized_engine' failed.
  Aborted (core dumped)

我还尝试过将fp16的选项，engine转化没有问题，但是int8就会报如上错误

[mpj1234]

我也是

[mpj1234]

I have solved v5-seg and submitted a new PR. You can refer to #1556.

[mpj1234]

It has been fixed. Please check in trt10 branch or #1557.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.