“A demonstration of WNB Training Serverless SFT used the QEN3 14B model as its base model, served via WNB Inference.”