I try to generate the two int8 engine with two gpus.
The second gpu will show the error and the one is successful.
When I retry the second gpu, and it is successful.
Do you know the reason?
rayhong
1
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Cask Error when using int8, fp32 works fine. | 1 | 457 | March 9, 2020 | |
tensorRT3: is it possible to create two engines? | 2 | 479 | October 18, 2021 | |
using samepleINT8 and GEMM error | 0 | 554 | July 27, 2017 | |
GPU isn't used when serializing a int8 engine model from UFF | 1 | 609 | October 16, 2019 | |
Building TensorRT int8 engine fails | 1 | 342 | January 20, 2021 | |
GPU isn't used when serializing a int8 engine model from UFF | 0 | 264 | October 10, 2019 | |
Int8 get the same result, but in FP16 the result is correct | 1 | 403 | December 1, 2021 | |
Multiple threads execution with different engines in tensorrt | 3 | 2525 | December 13, 2022 | |
Building TensorRT int8 for batch greater than 1 fails | 1 | 429 | January 26, 2021 | |
TensorRT YOLO Int8 on GTX 1080ti | 9 | 3983 | October 12, 2021 |