Closed NasonZ closed 4 months ago
A work around for this is to install the dependencies via the given requirements.txt
instead of using the code given in the notebook.
@NasonZ I'm having same issues, can you paste what you edited in the notebook to get it working?
I uploaded the repo to colab as a zip file (cloning it would also work I guess), then ran.
!pip install cython
!sudo apt update && sudo apt install ffmpeg
!unzip /content/whisper-diarization-main.zip
!cd whisper-diarization-main/
%cd whisper-diarization-main/
!ls
!pip install -r requirements.txt
Managed to run the rest of the notebook without issue after this
Has not helped. The original notebook could not import (download) a model.
After I added the lines, as per above, to the notebook :
...
!pip install numba==0.58.0
!pwd
!pip install cython
!git clone https://github.com/MahmoudAshraf97/whisper-diarization
#!sudo apt update && sudo apt install ffmpeg
%cd whisper-diarization/
!ls
!pip install -r requirements.txt
i get
[<ipython-input-4-126e61cacd2c>](https://localhost:8080/#) in <cell line: 6>()
4 import json
5 import shutil
----> 6 from faster_whisper import WhisperModel
7 import whisperx
8 import torch
9 frames
[/usr/local/lib/python3.10/dist-packages/torch/__init__.py](https://localhost:8080/#) in <module>
227 if USE_GLOBAL_DEPS:
228 _load_global_deps()
--> 229 from torch._C import * # noqa: F403
230
231 # Appease the type checker; ordinarily this binding is inserted by the
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xb0 in position 0: invalid start byte
Notebook updated and confirmed working on colab
I'm attempting to run the notebook but it fails when trying to run the imports.
I've tried installing and uninstalling numpy, any ideas on how to resolve this?