Did someone manage to force the Android device to use NPU in the app: compiled_model_api/image_segmentation/kotlin_npu/android ?

[rafalfirlejczyk]

I tried on two devices S22 with "SM8450" and CT70 with "QCS6690". I adopted the device_target_configuration.xml for the QCS6690. I compiled the model using the colab: https://colab.research.google.com/github/google-ai-edge/litert-samples/blob/main/compiled_model_api/colab/LiteRT_AOT_Compilation_Tutorial.ipynb#scrollTo=I_MF7QfsICSL

I could not however make the NPU work. CPU and GPU work OK.
Did someone make it work? On what device? What .so library have been used for that?
thanks

Modified part of the device_target_configuration.xml:

<config:device-group name="Qualcomm_SM8450">
    <config:device-selector>
        <config:system-on-chip manufacturer="Qualcomm" model="SM8450"/>
    </config:device-selector>
    <config:device-selector>
        <config:system-on-chip manufacturer="QTI" model="SM8450"/>
    </config:device-selector>
    <config:device-selector>
        <config:system-on-chip manufacturer="Qualcomm" model="QCS6690"/>
    </config:device-selector>
    <config:device-selector>
        <config:system-on-chip manufacturer="QTI" model="QCS6690"/>
    </config:device-selector>
</config:device-group>

[kmnaseef]

i tried on samsung flip 5 (Snapdragon 8 gen 2 ) and Xiomi 14 (Snapdragon 8 gen 3). but no success