Speaker 1: Hello everybody and welcome to a new video. In today's video I am going to show you how to install Whisper by OpenAI, the speech recognition application on a call app without any kind of complications and long videos. Without further ado let's jump straight into the call app that I've prepared for you guys. This is the call app, it's really really simple and all you have to do is to run two or three sounds and you'll see exactly what I mean. In this moment we do not have the Whisper installed so let's just press on the on the link up here and install Whisper and the required application which is ffmpeg and as you've seen I already have most of these things installed so for you it will take slightly longer than it took for me like 15-20 seconds longer. Cool now that we've installed it all we need to do is to add our mp3 or mp4 file to the Colab. now for this i have two ways that you can do it one is to download a video from youtube of course i suggest it's your own for this example i'm going to use one that's already existing and we are going to use this wget here however before doing that we are going to use a youtube 2mp3 website it's free i i have no affiliation whatsoever with this website but i use it frequently by myself and what I do is save the file to Dropbox right and all you have to do is go to Dropbox and basically get your file there you go and copy the link nice and once you have the link you're going to go back into the Colab paste it and just run this out and as you see here it's going to get downloaded. You also have a way to do this from downloading it on your PC by pressing download, saving it on your PC and then uploading to the Colab by yourself manually which takes slightly longer as WCAG uses the Colab resources to download files. Now the third step and one of the most important one is to rename this to something easier. We're going to name this to sound.mp3 since I already have it here as this then you are going to see something here a summation mark whisperer sound.mp3 the name of the file model medium.en which is the model that we are going to use they do have different kind of models you can see details about it here I consider that medium is the best if you want to transcribe really fast and and very accurate let's just press on it and we are going to see in real time when everything is being downloaded so the first thing that's being done is the model medium.en which says for english and english only if you want to use this in another language i do suggest you use the large model however it is considerably slower so please keep that in mind when you are doing so and also keep in mind that you cannot use .en on large if you were to use the large model you will either use it in this format so you will rather go something like model large v3 which is the latest and probably the best and so on you cannot use the .en termination here is the transcript being transcribed in a live manner and pretty fast as you see it's going to take a very short amount of time before this is downloaded as the video is not too long itself only three minutes or so and it has been transcribed for me and here is a video you can refresh this and you can get your transcribed file here in SRT json, tcv, txt and vdt and all you have to do is to well either right click on it and press download and that's pretty much it that's how you have it guys that's how you use whisper really really fast and if you want to see a video how you install it on your local PC please let me know in the comments and I will do that too. Until the next time see you guys thank you for watching and please don't forget to subscribe