Error On Install Dependencies; expected? #127

Closed
opened 2023-03-13 03:52:13 +00:00 by deadeye-1123 · 7 comments

I'm getting this error on install - the voice generation works, but I get a ffmpeg failure on generate transcript (ffmpeg.exe is in ./bin) not sure if they are related? Are these errors ignorable and should I dig else where to resolve the ffmpeg call issues? Thanks!

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
voicefixer 0.1.2 requires librosa<0.9.0,>=0.8.1, but you have librosa 0.10.0 which is incompatible.
pyannote-audio 2.1.1 requires einops<0.4.0,>=0.3, but you have einops 0.6.0 which is incompatible.
pyannote-audio 2.1.1 requires soundfile<0.11,>=0.10.2, but you have soundfile 0.12.1 which is incompatible.
tortoise 2.4.4 requires transformers==4.19, but you have transformers 4.26.1 which is incompatible.

I'm getting this error on install - the voice generation works, but I get a ffmpeg failure on generate transcript (ffmpeg.exe is in ./bin) not sure if they are related? Are these errors ignorable and should I dig else where to resolve the ffmpeg call issues? Thanks! ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. voicefixer 0.1.2 requires librosa<0.9.0,>=0.8.1, but you have librosa 0.10.0 which is incompatible. pyannote-audio 2.1.1 requires einops<0.4.0,>=0.3, but you have einops 0.6.0 which is incompatible. pyannote-audio 2.1.1 requires soundfile<0.11,>=0.10.2, but you have soundfile 0.12.1 which is incompatible. tortoise 2.4.4 requires transformers==4.19, but you have transformers 4.26.1 which is incompatible.
Owner

but I get a ffmpeg failure on generate transcript (ffmpeg.exe is in ./bin) not sure if they are related?

You'll need to uninstall and install ffmpeg-python, for some reason.

Windows:

.\venv\Scripts\activate
pip uninstall ffmpeg-python
pip install ffmpeg-python
deactivate

Linux:

source ./venv/bin/activate
pip3 uninstall ffmpeg-python
pip3 install ffmpeg-python
deactivate
> but I get a ffmpeg failure on generate transcript (ffmpeg.exe is in ./bin) not sure if they are related? You'll need to uninstall and install ffmpeg-python, for some reason. Windows: ``` .\venv\Scripts\activate pip uninstall ffmpeg-python pip install ffmpeg-python deactivate ``` Linux: ``` source ./venv/bin/activate pip3 uninstall ffmpeg-python pip3 install ffmpeg-python deactivate ```

I have similar error, using setup-cuda.sh.
Tried the ffmpeg-python uninstall/intsall, didn't do anything.

Installing collected packages: librosa
  Attempting uninstall: librosa
    Found existing installation: librosa 0.8.0
    Uninstalling librosa-0.8.0:
      Successfully uninstalled librosa-0.8.0
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
pyannote-audio 2.1.1 requires einops<0.4.0,>=0.3, but you have einops 0.6.0 which is incompatible.
Successfully installed librosa-0.8.1
rm: cannot remove '*.bat': No such file or directory

I tested this on ubuntu LTS 18, 20, 22, RockyLinux9 and their supported python versions, every time same message.

The setup-cuda.sh (i added set -e at the beginning so it'd stop at errors):

#!/bin/bash
set -e
# get local dependencies
git submodule init
git submodule update --remote
# setup venv
python3 -m venv venv
source ./venv/bin/activate
python3 -m pip install --upgrade pip # just to be safe
python3 -m pip install wheel
# CUDA
pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu117
# install requirements
python3 -m pip install -r ./requirements.txt # install local requirements
python3 -m pip install -r ./modules/tortoise-tts/requirements.txt # install TorToiSe requirements
python3 -m pip install -e ./modules/tortoise-tts/ # install TorToiSe
python3 -m pip install -r ./modules/dlas/requirements.txt # instal DLAS requirements, last, because whisperx will break a dependency here

python3 -m pip install -U einops==0.6.0
python3 -m pip install -U librosa==0.8.1

rm *.bat

deactivate

While performing installation manually, I get the error after:

python3 -m pip install -r ./modules/tortoise-tts/requirements.txt
I have similar error, using setup-cuda.sh. Tried the ffmpeg-python uninstall/intsall, didn't do anything. ``` Installing collected packages: librosa Attempting uninstall: librosa Found existing installation: librosa 0.8.0 Uninstalling librosa-0.8.0: Successfully uninstalled librosa-0.8.0 ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. pyannote-audio 2.1.1 requires einops<0.4.0,>=0.3, but you have einops 0.6.0 which is incompatible. Successfully installed librosa-0.8.1 rm: cannot remove '*.bat': No such file or directory ``` I tested this on ubuntu LTS 18, 20, 22, RockyLinux9 and their supported python versions, every time same message. The setup-cuda.sh (i added set -e at the beginning so it'd stop at errors): ``` #!/bin/bash set -e # get local dependencies git submodule init git submodule update --remote # setup venv python3 -m venv venv source ./venv/bin/activate python3 -m pip install --upgrade pip # just to be safe python3 -m pip install wheel # CUDA pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu117 # install requirements python3 -m pip install -r ./requirements.txt # install local requirements python3 -m pip install -r ./modules/tortoise-tts/requirements.txt # install TorToiSe requirements python3 -m pip install -e ./modules/tortoise-tts/ # install TorToiSe python3 -m pip install -r ./modules/dlas/requirements.txt # instal DLAS requirements, last, because whisperx will break a dependency here python3 -m pip install -U einops==0.6.0 python3 -m pip install -U librosa==0.8.1 rm *.bat deactivate ``` While performing installation manually, I get the error after: ``` python3 -m pip install -r ./modules/tortoise-tts/requirements.txt ```
Author

My symptom was unrelated to the einops version error message and solved by mrq's suggestion above. Many thanks!

However I'm still getting the version conflict message on install/update, fyi

My symptom was unrelated to the einops version error message and solved by mrq's suggestion above. Many thanks! However I'm still getting the version conflict message on install/update, fyi

I get this error message in Google Colab. Why?
!source ./venv/bin/activate
/bin/bash: ./venv/bin/activate: No such file or directory
Tried it another way
%cd /content/ai-voice-cloning/venv/bin
!source ./activate
/bin/bash: ./activate: No such file or directory
And this way
!python -m venv venv
Error: Command '['/content/ai-voice-cloning/venv/bin/python3', '-Im', 'ensurepip', '--upgrade', '--default-pip']' returned non-zero exit status 1.

I get this error message in Google Colab. Why? `!source ./venv/bin/activate` `/bin/bash: ./venv/bin/activate: No such file or directory` Tried it another way `%cd /content/ai-voice-cloning/venv/bin` `!source ./activate` `/bin/bash: ./activate: No such file or directory` And this way `!python -m venv venv` `Error: Command '['/content/ai-voice-cloning/venv/bin/python3', '-Im', 'ensurepip', '--upgrade', '--default-pip']' returned non-zero exit status 1.`
Owner

You're using an extremely old version of the notebook.

You're using an extremely old version of the notebook.

Where can I find the latest version? I downloaded this file here.
https://git.ecker.tech/mrq/ai-voice-cloning/raw/branch/master/notebook_colab.ipynb
Edit: Nevermind I think I found it.
https://git.ecker.tech/mrq/ai-voice-cloning.git
Edit: I noticed if I install this instead. Then I'll be able to use the source command
!apt install python3.9-venv
!source ./venv/bin/activate

Where can I find the latest version? I downloaded this file here. https://git.ecker.tech/mrq/ai-voice-cloning/raw/branch/master/notebook_colab.ipynb Edit: Nevermind I think I found it. https://git.ecker.tech/mrq/ai-voice-cloning.git Edit: I noticed if I install this instead. Then I'll be able to use the source command `!apt install python3.9-venv` `!source ./venv/bin/activate`
Owner

That should be it. It doesn't have a line for !source ./venv/bin/activate, so whatever you're using is most definitely not the most recent version.

That should be it. It doesn't have a line for `!source ./venv/bin/activate`, so whatever you're using is most definitely not the most recent version.
mrq closed this issue 2023-03-21 21:40:36 +00:00
Sign in to join this conversation.
No Milestone
No project
No Assignees
4 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: mrq/ai-voice-cloning#127
No description provided.