Today, I continued analyzing the performance of the Whisper model, focusing on the quality of its transcriptions. While the model's output was accurate, I noticed that it tends to produce formal, written language, which might not be ideal for the conversational and accessible tone needed for the Elderly Care System. Recognizing this limitation, I decided that next week I will start fine-tuning the Whisper model to better suit the specific needs of my project, aiming to make the output more natural and conversational for the elderly users.