Simple Audio-Transcriber

A simple implementation of the Whisper-AI from OpenAI

Check for Cuda and PyTorch

Simple run the file "test_for_cuda.py", if you want to check if PyTorch is setup correct.

Install Cuda and PyTorch witch Cuda-Support:

1. Check what version of Cuda you need. For that look at and follow the accepted Answer in the Stackoverflow-Thread
2. After that, download the appropriate Cuda-Version from their Archive
3. Install PyTorch with Cuda-Support. You can generate an Installation-command via the official PyTorch-Site
   Copy and run the command as it is. You shouldn't have to the the Cuda-Version in the command itself.
4. Run "test_for_cuda.py" to see if installation was successful.

Transcribing X files

To transcribe an x amount of files simply insert them into the Input-folder and run "transcribe_cuda_multiple_files.py". After it ran through you can find the corresponding txt-files in the Output-Folder.