Because I updated OpenAI Whisper I want to try its Cantonese The text on the top is the original the text just transcribed using WhisperAI The text on the bottom is human-edited Ok, let's go Ok, let's get started Ok, actually, I'm not feeling very well today But I... um ... Maybe my voice isn't very good I mainly want to try... Since i've already updated this... Whisper ... Um... OpenAI Whisper I want to try its Cantonese Let me show you what it is Let's type 'Whisper' ...um... hyphen Because I mainly... Why is the sound like this? Is it due to my voice? my sick voice or ... whatever reason? Let me see Let me change it a little bit Vocoder Ok, if this... I think it should be due to my voice My voice is not good, it doesn't matter It's not related to the recording problem Huh Let's go back to this WhisperAI first That is... Um... This one OpenAI Whisper Whisper Haha, out of tune Ok Hum... Hum... It's actually quite new You can see it Version 2023-1117 That is, yesterday Just updated 18 hours ago It updated the code and... and it has a... Cantonese database I mean the (Cantonese) model It should be here Here Added Cantonese here This one Ok, now I want to try ... In this video right at the bottom of the screen Sorry, I mean You see in this video the caption shown on top is ... the original WhisperAI's transcription that was created using the WhisperAI's Cantonese model The caption below is the human-edited version the human-edited transcription They'll be put together for comparison I mean, the top white caption... is the original caption ... the one that's transcribed using whisperAI's Cantonese model The one that uses... um ... um ... the one below ... Oops, the yellow bottom one is ... Sorry, I mean ... if there's human-edited caption below, it'll be... displayed ... em... in yellow Ok Um ... see if ... understand (my Cantonese) This is just testing the Cantonese transcription Alright let's go Oops, it should be almost there Ok, done
