Transfer rate from GPU to CPU with pytorch on Xavier NX

Robotics & Edge Computing Jetson & Embedded Systems Jetson Xavier NX

amir6 February 12, 2022, 8:25pm 1

I’m trying to run a segmentation network, the result of which is a 8x3x224x224 tensor.
This takes ~500ms, which seems excessive (this is the time it takes the .cpu() function to run, as measured by cProfile).

Is there a way to reduce this?
Thanks

Topic		Replies	Views
PyTorch's cpu() function call takes a lot of time on Jetson Xavier Jetson AGX Xavier	8	1281	July 16, 2019
transport prediction to cpu lost too much time Jetson AGX Xavier	1	374	September 4, 2019
Transfer data from GPU to CPU takes too much times on TX2 TensorRT	1	1302	August 9, 2019
Transfer data from GPU->CPU takes too much time. TensorRT	0	293	May 23, 2019
Transfer data from GPU->CPU takes too much time. TensorRT	0	256	May 23, 2019
Transfer data from GPU->CPU takes too much time. TensorRT	0	570	May 23, 2019
Transfer data from GPU->CPU takes too much time. TensorRT	0	328	May 23, 2019
Transfer data from GPU->CPU takes too much time. TensorRT	0	265	May 23, 2019
transfer data from GPU->CPU takes too much time. TensorRT	0	309	May 23, 2019
Transfer data from GPU to CPU takes too much times on TX2 Jetson TX2	5	1002	October 18, 2021

Transfer rate from GPU to CPU with pytorch on Xavier NX

Related topics