Cuda failure with cached engine

NVIDIA / object-detection-tensorrt-example

Running object detection on a webcam feed using TensorRT on NVIDIA GPUs in Python.

211 stars 61 forks source link

Open merry-cooperation opened 5 years ago

merry-cooperation commented 5 years ago

Hello. Example runs great with freshly built TensorRT engine, but when trying to do inference with cached one I get

Cuda failure: 1
Aborted (core dump)

after

# Transfer input data to the GPU.
[cuda.memcpy_htod_async(inp.device, inp.host, stream) for inp in inputs]

in do_inference() at line 281 in inference.py

Deleting .buf file and rebuilding enigine helps.

pjkasper commented 5 years ago

!Same here..seg fault -at line 281 inference.py ( running GTX 1060 CUDA 10.1 NVIDIA 418.56)

HazemAbdelhafez commented 4 years ago

I have the same problem.

bl6594 commented 4 years ago

any progress solving this issue?