I really like the feature to test the speed (inference time) of a custom model at specific CPUs available at the workbench.
I wonder if such a functionallity is available without the workbench. Via a normal cli command or p.ex. in a jupyter lab ( if necessary also with an authentication via the intel dev account)
Thanks for reaching out.
There is Benchmark Tool in OpenVINO that estimate deep learning model inference performance on supported device. Performance can be measured for two inference modes: synchronous (latency-oriented) and asynchronous (throughput-oriented).
Latency measures inference time (ms) required to process a single input. When it comes to batch input need to measure throughput (images per second or frames per second, FPS).
This thread will no longer be monitored since we have provided a solution. If you need any additional information from Intel, please submit a new question.