abapteachablemachine | basic flash card game developed in ABAP Javascript

Train and test your own ML audio model on Teachable Machine - It's really FUN!
Export the model : Either download the model (ZIP file) or get the shareable link on Teachable Machine.
If you've downloaded the model locally, unzip your model and upload 'metadata.json', 'model.json' and 'weights.bin' to ABAP server. I have created a 'models' folder under the BSP application ZZATM.
Run transaction ZZATMSETUP to create a new player, game and questions.
Player name - Up to 40 characters is supported but avoid using any special characters such as &,%,*,$ etc...
Profile picture - URL to a small avatar picture. I've included four sample images in the demo_assets folder.
Debug - If this option is checked, during gameplay, the class labels and scores will be displayed below the screen.
Sound settings - configure different background music during gameplay and sound effects for when getting a correct answer and when the game ends.
Game name - Up to 40 characters is supported but avoid using any special characters such as &,%,*,$ etc...
Game type - Only one option is available now but this gives the flexibility to create more games and route them to different pages.
Game model - The full and complete URL (https) to the ML model that contains model.json, metadata.json and weights.bin
Match probability - The minimum score match to consider that it's a correct answer. When the ML model is 100% sure, the score is 1.00. Recommended config could be something like 0.80 (easy difficulty - not really accurate), 0.90 (medium difficulty) or 1.00 (hard difficulty - highly accurate match). You'll just need to tweak this and it also depends on your ML model as well.
Time to answer - The time in seconds to display the flash card image and listen for an answer. Rcommended config can be something like 10 seconds (easy difficulty) - 5 seconds (medium difficulty) or 3 seconds (hard difficulty).
Array index - this is a little tricky - you need to match the wordLabels array in the metadata.json file of your model. Typically, "Background noise" is the first entry so you can ignore array[0], and you just need to note what the other class labels are and their array index.
Image URL - The flash card image to display when showing this question and remember to match the array index correctly. See array index 4,5,6 and look at the matching answer - 1, 3,2 - so, don't always assume that it is in order.