SingScribe is a fine-tuned version of Whisper, a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification. I have fine-tuned it to be used on Singlish.