MahmoudAshraf97/whisper-diarization

Google Colab script fails

Closed this issue · 2 comments

Trying to run the below, which uses the scripts here:
https://colab.research.google.com/github/Transcripts4All/tools4all/blob/main/whisper-diarization.ipynb
https://github.com/Transcripts4All/tools4all/
Transcripts4All/tools4all#6

import os
!git clone https://github.com/MahmoudAshraf97/whisper-diarization
os.chdir("/content/whisper-diarization")
!git checkout a7635077d76a5288a27998ab3a13d519541f063d
!pip install git+https://github.com/m-bain/whisperX.git@78dcfaab51005aa703ee21375f81ed31bc248560
!pip install "nemo-toolkit[asr]>=2.dev"
!pip install --no-deps git+https://github.com/facebookresearch/demucs#egg=demucs
!pip install git+https://github.com/oliverguhr/deepmultilingualpunctuation.git
!pip install git+https://github.com/MahmoudAshraf97/ctc-forced-aligner.git

It fails when I try to run it. It worked about 30 days ago, but not until issues were resolved first, similar to this time.

Using PasteBin to get around Github's max character limit:

Step 1 (above): https://pastebin.com/H8D1XXuw

Runtime -> Restart Session

Step 2: https://pastebin.com/RnHYQMKk

I noticed this from Step 1:

Line 335:
sentence-transformers 3.2.1 requires transformers<5.0.0,>=4.41.0, but you have transformers 4.39.3 which is incompatible.

Line 606:
gcsfs 2024.10.0 requires fsspec==2024.10.0, but you have fsspec 2024.9.0 which is incompatible.

Line 607:
sentence-transformers 3.2.1 requires transformers<5.0.0,>=4.41.0, but you have transformers 4.39.3 which is incompatible.

This notebook is not a one that I maintain, please use the one in this repo

Thanks for the fast reply. The linked script looked like it was mainly using yours, but I probably got that wrong.

I tried using your notebook, but it hangs when I start the "Processing" step, the green arrow seems to be stuck at line 3, return_code = os.system(.

I have audio_path = "Example.wav" where Example.wav is in the content directory.

I restarted the runtime after cell 1 under "Installing Dependencies" had completed.

Edit - It worked, Google Colab was just running slowly! 😛