Neuron Runtime return codes

Code	Name	Description
0	NERR_OK	Successful completion.
1	NERR_FAIL	Non specific failure, e.g. low level interface to the hardware error.
2	NERR_INVALID	Invalid/corrupted NEFF file, bad Neuron instruction, invalid access to Neuron memory, etc.
3	NERR_INVALID_HANDLE	Invalid handle was passed to the request, e.g. for a Neural Network that has not been loaded or has been previously unloaded.
4	NERR_RESOURCE	Failed to allocate a resource for requested operation, for example: not enough TDRAM to load a Neural Network, not enough Host memory to perform an operation.
5	NERR_TIMEOUT	Request timed out.
6	NERR_HW_ERROR	RT failed to initialize, after a number of failed attempts to start RT stays up and returns the error in response to every GRPC call. Common causes: bad hugepages configuration (insufficient number of reserved pages on the system), failure to initialize Neuron device.
7	NERR_QUEUE_FULL	Not enough space in the inference input queue. The number of submitted and not completed inference requests is greater that what’s been configured during NN load. This is a transient error, inference requests can be submitted after some of the in-flight inferences have completed.
9	NERR_RESOURCE_NC	Not enough available NCs to load a Neural Network.
10	NERR_UNSUPPORTED_VERSION	NN load failed because the version of NEFF is not supported.
1000	NERR_INFER_PENDING	Inference has not completed yet.
1002	NERR_INFER_BAD_INPUT	Invalid input has been submitted to infer, e.g. missing some of the input tensors, incorrect input tensor sizes.
1003	NERR_ INFER_COMPLETED_WITH_NUM_ERR	Inference is completed. Numerical errors were encountered while executing the inference (NaN).
1004	N ERR_INFER_COMPLETED_WITH_ERR	Inference is completed. Non-numerical errors were encoutered while executing the inference. Usually indicates problems with compiled NN (compiler errors) or the hardware issues.