FREE & OFFLINE Audio to Text | Whisper: Install Guide | OpenAI Whisper | ASR

OpenAI has done some fantastic things. Whisper is a great project open to the public. Transcribe (Turn audio into text) for MANY languages, all completely for free and all from your computer. No subscription, no fees, nothing. Your data, your computer, your free unlimited transcription.

Timestamps:
0:00 - Intro
0:20 - Whisper AI is good
1:30 - Check if Python is installed
2:37 - Download Python 3.9.9
3:54 - What to do when you have multiple Python versions
4:40 - Install CUDA (NVIDIA users)
5:44 - Installing PyTorch
6:15 - Installing FFMPEG
7:50 - Installing Whisper
8:35 - Whisper: Get help/Command reference
9:19 - Basic Whisper usage
11:08 - Transcribe MP3 to text using Whisper (FREE & Offline)
14:20 - English model variants

#Whisper #AI #OpenAi
-----------------------------
-----------------------------

Everything in this video is my personal opinion and experience and should not be considered professional advice. Always do your own research and ensure what you're doing is safe.

[thanatos]

My father-in-law routinely has a lot of interviews to transcribe. Recently stuck with an increased workload, I went on a quest to find a tool, and stumbled upon whisper, and then your video. I wouldn't have known where to start with python and the cmd window, despite knowing computers well enough. This was of immense help and extremely easy to follow (the only missing thing for me was that I was missing Git. But adding it and also adding it to path was very easy (a bonus thing I learned thanks to this video)). Thank you so much. You just saved an old man a week worth of work and many more in the future.

[morbid.curiosity]

Incredibly thorough, broke down an "intimidating" task into managable chunks, edited well with fantastic audio. I'm glad I was linked this video! Subscribed!!!

[BenjaminK]

Great video, i have been using the wisper API for testing but that costs i didnt know this was open source, your video got me up and running in no time at all, thank you for taking the time to make this its very cool stuff

[IR]

This was wonderful, I followed along using my Macbook Air, and it worked. Thank you so much. Brilliant!

Edit -

I could transcribe a 45-minute audio recording of a meditation I had to learn. Another idea is to record your own voice for a project and transcribe it. This was so cool, and I thank you for the effort in making this video!

[MondSemmel]

10:38: Note that the "large" Whisper model is now "large-v2", and the old "large" model has been renamed to "large-v1" or something.

[weekendmakeit]

you sir are the best. I had to install "GIT" as well

[TradTalk]

Thanks a lot! I was having a lot of troubles with the PowerShell running of the program, but after this tutorial all my problems were solved.

[usernotfound.]

i have seen so many tutorial videos in my life, this one, is on my top 3 list.
just WOW..
perfect everything perfect, thanks for the great work

[MRAHBB]

Thanks. This took me about 2 hours and some extra installs for missing scripts, moving scripts and paths, but I got it.

[anthologyofheroespodcast]

Very useful, particularly the multiple Python instances. Thanks!

[lazyhomeowner]

2023. things have still not gotten simpler. this was like a speed run on how to set up an ide.

[MrFerocius]

Awesome quality! Suscribed! Keep them coming.

[klauser]

You only missed the 'git' installation but everything else was great. I got it to work. Thanks.

[sergiomendez]

Thank you very much!!! Never stops sharing please 🙏

[Grapegum]

I would've preferred the venv or even conda approach but still, it was a pretty good tutorial. Thank you!

[BirjloxuBR]

Sorry, but this tutorial is confusing. First the missing git thing, that you didn't explain. Second, it now refuses to use my NVIDIA GPU as the renderer thing or whatever. Even though I installed CUDA. So something is missing. For someone who is not into coding, python or anything related this tutorial is just lacking and does not provide all the correct info and troubleshoots, no pun intended.

[rich]

one obstacle to clarify is that the "path" folder isn't actually in his "ESD" folder, i messed that up

[gamer-mtpe]

after 5:44 i got so confused since it just got confusing. great video i just couldn't understnad past that time period

[andreydmitriyev]

Tnx! And if the video card is built-in, do you need to install nvidia software from the section?

[neoneuron]

what should i choose, i have 1650 amp gpu - does CUDA (pytorch) works ?