NVIDIA
the result on TensorRT10.3 is difference from TensorRT10.8 and is wrong
The results of tensorRT10.3 are different from those of TensorRT 10.8. The results of 10.8 are as expected, so the results of 10.3 are wrong. I want to use 10.3 because jetpack supports the version of 10.3. My network contains transformers with dynamic dimensions. I don't know if there is any bug in version 10.3.
The results of tensorRT10.3 are different from those of TensorRT 10.8. The results of 10.8 are as expected, so the results of 10.3 are wrong. I want to use 10.3 because jetpack supports the version of 10.3. My network contains transformers with dynamic dimensions. I don't know if there is any bug in version 10.3.
I try the version of TensorRT 10.5, the results is OK , so I think it is a problem with version 10.3. Is there any solution?
Based on what you're describing it sounds like a bug that was in 10.3 that was fixed in subsequent versions. Our recommendation would be to switch to the newer versions if possible.
