Transcribe and Translate Audio with Whisper Using Ubuntu

preview_player
Показать описание
Welcome to our comprehensive guide on installing and using OpenAI's Whisper, an advanced speech recognition model. This tutorial is perfect for those looking to dive into the world of AI-powered speech recognition. Follow these steps to get started:

Requirements:

- Python: Ensure you have Python version 3.8 to 3.11 installed on your system. You can download Python from official Python website.

- Pip3: This is Python's package installer, which should come with your Python installation.

- Pytorch: A machine learning library for Python, critical for running Whisper. Install it by following instructions on the Pytorch website.

- FFmpeg: This tool handles multimedia data and is necessary for audio processing. Download it from the FFmpeg official site.

- OpenAI-whisper: Install this package via pip3 using the command pip3 install openai-whisper.
Installation Steps:

1. Verify Python & Pip: Check your Python and Pip versions using python --version and pip --version in your command line.

2. Install Pytorch: Follow the instructions on the Pytorch website to install the correct version for your system.

3. Install FFmpeg: After downloading, follow the installation guide on the FFmpeg site to set it up.

4. Install Whisper: Run pip3 install openai-whisper in your command line.
Usage Guide:

Once installed, you can use Whisper to transcribe audio files. We'll cover basic usage in this video, showing how to input an audio file and get a text transcript.
Troubleshooting Tips:

Make sure all components are installed correctly.
Verify that your Python environment is properly set up, especially if using virtual environments.
Additional Resources:

For more detailed information, check out the links below

Conclusion:
We hope this guide helps you successfully install and use Whisper for your speech recognition needs. For more tech tutorials, don't forget to like, share, and subscribe to our channel!
Рекомендации по теме
Комментарии
Автор

The transcription and translation are very good, I'm from Brazil, and I watched this video with active subtitles (not YouTube's automatic subtitles, which are pretty bad by the way), and the quality of the subtitles in Portuguese is excellent! If possible, you can put subtitles in Portuguese on all your videos that I will watch! I'm going to install Whisper on my Linux Mint now.

CaioTechLabs
Автор

I love the way you made this video. Especially the way you made mistakes and corrected them without cutting. Since I make the same mistakes constantly this is extremely helpful. I really mean it.

gregorrick
Автор

I didn't understand a bit of this. Fantastic video.

StephenWilds
Автор

Great job! Your install went much better than mine. I wish I had seen this video before hand. Oh well, we both have functioning whisper now.

TimothyHuey
Автор

Nice job, thanks for posting. Years ago, tried Dragon, but did not like it, trying to punctuate while maintain a though flow. Last 7 years, full time Linux. Currently bouncing between Nerd Dictation and also considering Whisper. The video is helpful. I think I'll try using both programs. 😎

craigw
Автор

Thank you so much, very informative 👍🙏

nouche
Автор

Very interesting, please continue to make these videos.

iraislandii
Автор

Whisper will work on any hardware. If you don't have a NPU or GPU, it will crunch that audio with your CPU. I've run it on a Raspberry Pi and the transcription speed was actually pretty impressive.

arthurpizza
Автор

Great video! I was about to switch to Linux, I'm glad I saw this instead. I'll have too look into if there is a way to make a new version of Windows secure instead it seems.. This is too much for me.

twinesper
Автор

Does whisper work with lyrical songs? Or just spoken word?

gregvisioninfosoft
Автор

looks like everything is installed but when I they type 'whisper --version or --help' nothing happens - says command not found.

X1:~$ whisper --version

Command 'whisper' not found, did you mean:
command 'whipper' from deb whipper (0.9.0-7build1)
Try: sudo apt install <deb name>

X1:~$ whisper
Command 'whisper' not found, did you mean:
command 'whipper' from deb whipper (0.9.0-7build1)
Try: sudo apt install <deb name>

I've been on Ubuntu about 15 years - installs are often challenging but rarely nothing happens

johnmiglautsch
Автор

after pip3 install torch
i got :
error:

thegames
Автор

excellent vidéo merci beaucoup sa marche bien

spectre-CA
Автор

Can we convert audio directly while speaking without saving the audio file?

pallawikumari
Автор

Unfortunately, this is the path only for the Chosen Ones. Fortunately for ordinary people there is Buzz with UI.

Bora_na_bora