A guide to generating subtitles through Whisper AI

[lordlance]

Whisper AI is a software that allows artificial intelligence to transcribe audio or video files supplied to it. I have no idea how subtitles are made here but someone on Reddit suggested that I share my guide here as people may find it useful. To keep things simple, almost everything is done via powershell and all that is required of you is to copy and paste the commands given below.

Step 1: Write powershell after pressing the windows button on your keyboard and on the right side you should see run as administrator option. If not, right click on powershell and click on run as administrator as seen here (you can choose to open in either of the two ways.)

Step 2: Install chocolatey by pasting the following commands into powershell:

Set-ExecutionPolicy AllSigned

Set-ExecutionPolicy Bypass -Scope Process -Force; [System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072; iex ((New-Object System.Net.WebClient).DownloadString('https://community.chocolatey.org/install.ps1'))

Step 3: Open a new powershell in administrator mode like you did in step 1 and paste the following commands one by one:

choco feature enable -n=allowGlobalConfirmation

choco install python39 github-desktop ffmpeg

pip install git+https://github.com/openai/whisper.git

pip install git+https://github.com/dsymbol/decipher

Step 4: Download this file: https://mega.nz/file/nLwzFIrQ#IamjPi_d9h7faGlTg0U_zurDsMHiTtoQRQyiyfM4oUohttps://mega....toQRQyiyfM4oUo

Step 5: Finally simply drag and drop your file to the BAT file that you just downloaded and the process of creating an SRT file will begin. Depending on how old your PC is, a single play can take the entire night so let it run overnight and be patient. This is a very CPU-intensive task so newer CPUs could do it much faster. I have a quad-core CPU and a single play can take several hours.

[jack_666]

The op needed to upload this "humongous" one line of code to the cloud.

FOR %%A IN (%*) DO decipher transcribe -i %%A --model small.en --language English

smh