Transcribe Audio Files in your Google Drive with OpenAI's Whisper

In this tutorial I show how to use OpenAI's Whisper automatic speech recognition model, Google Drive and Colab to transcribe all the audio files in a Google Drive folder for free.

[bartholdrotgans]

Thank you Andrew! Very clear and works perfectly!

[ezelleze]

I need assistance, i followed the instructions carefully, my original audio had Arabic voice over. so i removed the hashtag from the large model and added it to the others, that's the only difference from your tutorial but for some reason the exported txt file contains random English text. how to fix please..

[jussimirpasold]

Really cool, it is possible to add speakers identification (diarization) in this model?

[jagdale]

I want to translate Spanish audio file in English can you send me Code

[steveb]

For some reason it does not work for me. the transcripts resulting from this method do not make sense at all. My guess is that the 16 Khz mono setting is too low quality.

[PremiumLeo]

Hey andrew, is there any way to use the new Whisperai using the API key, to transcribe from my google drive?

Thank you.

[laurentm]

Hi. I have an error in Step 2: MessageError: Error: credential propagation was unsuccessful

[vladip.]

Hey Andrew, thank you for that wonderfull piece 🙂
How to change the code to other language recognition - or is it working only defined with loading the right "model" without ".en"??

[naturals]

I am trying to figure out how to add translation to this. Still working on it. Anyone do this yet?

[olegchernov]

What a great video! Thank you for all of your hard work. I was wondering if it is possible to set up some kind of trigger, maybe a time trigger, to run this script periodically, let's say every couple of hours.

[marwanalnimri]

Andrew, thanks for this awesome video. I need to add time code per line to use as subtitles, how would I do that. With your instruction I was able to transcribe a video to text, all I am missing is the time code that is needed for subtitle burner App. Thanks

[TF-qtjh]

❤Great video! So many similar videos everywhere these days, but you were the one of the fists and easiest to follow. Thank you! I was looking for this! By the way, I'm trying to add "Summarize" function after this transcribing step. I'd appreciate it if you could give me (us viewers) some idea?💙

[hoopstalkcentral]

is Whisper more accurate than Playground?

[MuralesenExilio]

This is very helpful, thank you! How can I implement the language detection to Spanish?

[MondSemmel]

Thanks a ton, this is a wonderfully simple tutorial. When I looked up how to use OpenAI Whisper before, the instructions usually involved setting up a docker server and/or required a very powerful GPU.

[miriamclo]

Hi!
I must say your explanation was clear, but I did not understand the part about language. My transcriptions are in Spanish, and, in this case, Which model do I have to use? Should the code be the following: model = whisper.load _model ("medium.sp")?

[invisiblelovehaha]

Thanks for your useful tutorial. The hardware accelerator list only includes CPU, T4, TPU, A100 GPU, and V100 GPU. How can I make use of M1 Max? Thank you.

[dephinus]

This is brilliant, thank you. Not going to lie, I didn't realise I should open the Python Notebook link and then open in Colaborate as this is a very new world to me, but once I finally figured it out and let the system do it's thing, it has been a game changer. I record audio notes on my watch (Samsung Classic Watch 6), share them to my Google Drive into the WhisperAudio folder, and your magic code does the rest. You rock!

[kabirmjoshi]

Hi, thiswas super helpful. However I had a query. I followed your process to begin transcribing my interviews. These are a mix of English and Hindi. As I ran the transcription, I realised that It was taking really long. A 45 min interview took more than 90 min. is this normal or is there a way I can improve the speed

[chrismalone]

Thanks so much for this! Running my first file now.