This demo showcases Object Detection with SSD and new Async API. Async API usage can improve overall frame-rate of the application, because rather than wait for inference to complete, the app can continue doing things on the host, while accelerator is busy. Specifically, this demo keeps the number of Infer Requests that you have set using
nireq flag. While some of the Infer Requests are processed by IE, the other ones can be filled with new frame data and asynchronously started or the next output can be taken from the Infer Request and displayed.
NOTE: This topic describes usage of C++ implementation of the Object Detection SSD Demo Async API. For the Python* implementation, refer to Object Detection SSD Python* Demo, Async API Performance Showcase.
The technique can be generalized to any available parallel slack, for example, doing inference and simultaneously encoding the resulting (previous) frames or running further inference, like some emotion detection on top of the face detection results. There are important performance caveats though, for example the tasks that run in parallel should try to avoid oversubscribing the shared compute resources. For example, if the inference is performed on the FPGA, and the CPU is essentially idle, than it makes sense to do things on the CPU in parallel. But if the inference is performed say on the GPU, than it can take little gain to do the (resulting video) encoding on the same GPU in parallel, because the device is already busy.
This and other performance implications and tips for the Async API are covered in the Optimization Guide
Other demo objectives are:
-labelsoption) or class number (if no file is provided)
On the start-up, the application reads command line parameters and loads a network to the Inference Engine. Upon getting a frame from the OpenCV VideoCapture it performs inference and displays the results.
NOTE: By default, Open Model Zoo demos expect input with BGR channels order. If you trained your model to work with RGB order, you need to manually rearrange the default channels order in the demo application or reconvert your model using the Model Optimizer tool with
--reverse_input_channelsargument specified. For more information about the argument, refer to When to Reverse Input Channels section of Converting a Model Using General Conversion Parameters.
New "Async API" operates with new notion of the "Infer Request" that encapsulates the inputs/outputs and separates scheduling and waiting for result, next section.
The pipeline is the same for both modes. The difference is in the number of Infer Requests used.
The Inference Engine also offers new API based on the notion of Infer Requests. One specific usability upside is that the requests encapsulate the inputs and outputs allocation, so you just need to access the blob with
More importantly, you can execute a request asynchronously (in the background) and wait until ready, when the result is actually needed. In a mean time your app can continue :
Notice that there is no direct way to measure execution time of the Infer Request that is running asynchronously, unless you measure the Wait executed immediately after the StartAsync. But this essentially would mean the serialization and synchronous execution. That is why the inference speed is not reported. It is recommended to use the Benchmark App for measuring inference speed of different models.
For more details on the requests-based Inference Engine API, including the Async execution, refer to Integrate the Inference Engine New Request API with Your Application.
Running the application with the
-h option yields the following usage message:
Running the application with the empty list of options yields the usage message given above and an error message.
NOTE: Before running the demo with a trained model, make sure the model is converted to the Inference Engine format (*.xml + *.bin) using the Model Optimizer tool.
You can use the following command to do inference on GPU with a pre-trained object detection model:
The only GUI knob is using Tab to switch between "User specified" mode and "Min latency" mode.
The demo uses OpenCV to display the resulting frame with detections (rendered as bounding boxes and labels, if provided). The demo reports