Hailo monitor in Windows OS

General
1

Hello,

i have a simple question .

Is Hailort monitor not supported in Windows OS?

2

Hey @Yu_Eunsang ,

Welcome to the Hailo Community!

You’re correct— the hailortcli monitor command is only supported on Linux.

1 Like
4

@omria , Thank you for your quick response

Then, is there any other tool to monitor Hailo on Windows OS?

I want to see the NPU usage during the Inferencing Test.

Should I get it Through the Hailo API?

Best Regards

5

Hey @Yu_Eunsang,

Yep, I’d definitely recommend going this route for monitoring NPU utilization during your inference runs.

Both our C/C++ and Python APIs have built-in runtime statistics that you can poll directly from your application. Here’s a quick Python example of how I’ve been doing it:

from hailo_platform.pyhailort import VDeviceParams, VDevice, InferModelParams, ConfiguredInferModel

# After you've loaded your HEF and set up your VDevice
device = VDevice(VDeviceParams(device_ids=[0], scheduler=True))
model = ConfiguredInferModel(device, InferModelParams(hef_path="your_model.hef"))

# Run your inferences
outputs = model.run(inputs)

# Then grab the stats
stats = device.get_statistics()
print(f"Device Utilization: {stats.device_utilization}%")
print(f"Model Utilization: {stats.model_utilization}%")

If you’re working with the C API instead, there are equivalent functions in the “runtime statistics” section - stuff like hailo_vdevice_get_runtime_statistics() that’ll give you the same metrics.

For more detailed info, check out the HailoRT User Guide - specifically the Runtime statistics section starting around page 229 for the C++ API, and pages 365-367 for the Python API documentation.