• NCSDK Python API 2.x
• Overview
• Enumerations
  • DeviceHwVersion
  • DeviceOption
  • DeviceState
  • FifoDataType
  • FifoOption
  • FifoState
  • FifoType
  • GlobalOption
  • GraphOption
  • GraphState
  • LogLevel
  • Status
• Structures
  • TensorDescriptor
• Global Functions
  • enumerate_devices()
  • global_get_option()
  • global_set_option()
• Classes
  • Device
    • Overview
    • Device.close()
    • Device.destroy()
    • Device.get_option()
    • Device.open()
    • Device.set_option()
  • Fifo
    • Overview
    • Fifo.allocate()
    • Fifo.destroy()
    • Fifo.get_option()
    • Fifo.read_elem()
    • Fifo.remove_elem()
    • Fifo.set_option()
    • Fifo.write_elem()
  • Graph
    • Overview
    • Graph.allocate()
    • Graph.allocate_with_fifos()
    • Graph.destroy()
    • Graph.get_option()
    • Graph.queue_inference()
    • Graph.queue_inference_with_fifo_elem()
    • Graph.set_option()

Fifo.read_elem()

Overview

This method returns an element from the Fifo, usually the result of an inference, along with the associated user-defined object that was passed to Fifo.write_elem() or Graph.queue_inference_with_fifo_elem().

This will also remove the element from the queue.

Syntax

output_tensor, user_obj = fifo.read_elem()

Parameters

None.

Return

tensor, user_obj

A numpy.ndarray with output tensor data of the type specified by the FifoDataType option, and the user-defined data that was passed to Fifo.write_elem() or Graph.queue_inference_with_fifo_elem().

Raises

Exception with a status code from Status if underlying function calls return a status other than Status.OK.

Notes

• The FifoType set during initialization must allow read access for the API.
• The Fifo cannot be written to or read from until it is allocated with Graph.allocate_with_fifos() or Fifo.allocate(). The FifoState must be ALLOCATED.
• The API can only read each element once even if FifoOption.RW_CONSUMER_COUNT is greater than 1.
• This is a blocking call if FifoOption.RW_DONT_BLOCK is false. If the Fifo is empty this method will not return until there is an element to read.
  • You can check the current fill level of the Fifo with Fifo.get_option() for FifoOption.RO_READ_FILL_LEVEL.
  • After an inference is queued there is a delay while device communication occurs and the inference is performed before the element is removed from the input Fifo and placed into the output Fifo.

Example

from mvnc import mvncapi

#
# Open a Device, create a Graph, and load graph data from file...
#

# Allocate the Graph and create and allocate two associate Fifos for input and output
input_fifo, output_fifo = graph.allocate_with_fifos(device, graph_buffer)

#
# Write elements to an input Fifo with Fifo.write_elem() and initiate inferences with Graph.queue_inference()
#

# Read the result to the output Fifo
result_tensor, user_obj = output_fifo.read_elem()

#
# Do something with the result...
#

# Destroy the Fifos
input_fifo.destroy()
output_fifo.destroy()

#
# Perform other clean up...
#