I have been given an Avnet Tria Vision AI kit fitted with a QCS6490 SoM. Whilst there is documentation for this kit on the Edge-Impulse site, the same kit is not present in the drop-down list of Target hardware when I start a new E-I project. However, the Thundercom RGB3-Gen2 DK is listed in the target hardware drop-down list, so I wonder if I can target my project to the RGB3-Gen2 board, but deploy the results on the Tria board?
Will the Trai board be added to the target HW list?
Hi DJE666,
The target selection dropdown in the top right of your screen in an Edge Impulse project is for choosing the RAM/ROM/Latency estimation target. This doesn’t have any impact on your end deployment, just gives an indication of the model performance on device. As such I reccomend choosing the RB3-Gen2 board as your target in this case as it is the same chip as your Tria kit so the results should be the same.
For deployment as per the documentation you can either use the linux cli to run your model and automatically deploy the correct type, or download the Linux (AARCH64 with Qualcomm QNN) executable and use either the linux runner or one of our Linux SDKs e.g. Python to run the model in an end application