 So now let's go ahead and get started with the hands-on. So take the IoT board that you have in front of you and connect the USB connector to the STLink programmer. So then you'll want to open up STM32Q programmer and go ahead and connect your board, open file, and browse to the directory that's here. See ST Euro 2019, hands-on. Workshop 5, STM32L4, QBAI. Then you want to look for the function pack, Sensing 1, Version 2.2. This is the same one that you can download on ST.com and look for the binaries for your board. So the reason why we have so many binaries is that there's one for each neural network implementation. So there's one flavor for the Acuse 16 classification and three different flavors for the HAR. They're being trained, it's either a different topology or a different data set for the training. The ones with BL means that they will include a bootloader on top of the user application. The ones without BL need to be programmed at a specific memory region and you need to have programmed the bootloader prior to programming the application. So without BL you will use it for the photo and the ones with BL can be used just out of the box using Q-Programmer. So we'll be using the ASC underscore BL dot bin open, program, download, file down and complete and you can disconnect Q-Programmer. Once your board is programmed we'll open a serial terminal application such as TerraTerm, new connection, ST-Link Virtual Comports and you want to change the BOD rate to 115-200. Once your UART terminal is connected you can go ahead and press the black reset button and you should be able to see the started blog of the application. In the log you should look up for your Bluetooth device address. This will be used to identify which one is your device when we'll go ahead and connect to it using the smartphone. Once your board is programmed you can use the STBLE sensor app on your Android phone and find your device and go to the audio classification page. So other features of the app are just the raw sensor data. You can plot it. You can see your battery level if it's a battery powered device. But the two important tabs for the AI pack is the AI algorithm page where it's either audio classification or activity recognition depending on which binary you've programmed and the AI data log page where you can select which sensor you want to log and enter some labels and store it either on an SD card or the external QSPI flash memory. This is very useful to create new models, new data sets and retrain or either retrain your model.