This Colab notebook transcribe audio using Whisper by OpenAI. It uses another python package for identifying and separating speakers. In addition, it creates a TXT and SRT files do download.

In order to use:

1. Run all cells and follow the order
2. Upload an audio file
3. Select a language (English / Hebrew / Auto mode)
4. Select number of speakers
5. Run rest of cells
6. The transcription (including the diarization - separation by speakers) will be shown in the last cell
7. The SRT and TXT files will be available for download in the sidebar (under main folder)

Please note: GPU units may be neccessary for running this! If you don't have any GPU units, you may get error of CUDA: Out of memory! Make sure to select GPU Runtime, and if possible - turn on "high RAM".

If using please make sure to citate Yuval Avidani @HACKIT.CO.IL