Modify this part to use OpenVINO Toolkit perform inference on video stream
Closed this issue · 4 comments
@ahkhalwai @Harkirat155 In the model you use here would not able detect road crack you only can detect car or person because it's a pre-trained model.
Please use the following path to get label_map.pbtxt and read it. You will know which should be able to detect.
Hello @ahkhalwai,
The following is the result I run the last cell encouter error
2020-02-24 02:16:16.590382: W tensorflow/compiler/xla/service/platform_util.cc:256] unable to create StreamExecutor for CUDA:0: failed initializing StreamExecutor for CUDA device ordinal 0: Internal: failed call to cuDevicePrimaryCtxRetain: CUDA_ERROR_OUT_OF_MEMORY: out of memory; total memory reported: 7981694976. 2020-02-24 02:16:16.590651: F tensorflow/stream_executor/lib/statusor.cc:34] Attempting to fetch value instead of handling error Internal: no supported devices found for platform CUDA
Just restart the kernel and it will work. I can't check it for now I'm outside... But I believe, releasing the memory would work.
It's a memory error for hardware acceleration, and generally would have risen from multiple excutions of the detection and inference code.