exo-explore
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
Installed exo on my two Mac pros 6,1. They're connected by thunderbolt and communicate perfectly well. After running exo I get a notification saying none of PyTorch or Tensorflow have been found. Ive installed PyTorch on both computers and still have the same problem. I can see the two nodes, but not getting anything to work. Llama runs on terminal just fine. I even openwebui working through docker. Any help would be greatly appreciated.
I countered same problem at Ubuntu 18.04. I change the python version to 3.12 and install pyTorch (it only support python 3.12).
Thank you. I've installed various versions of python but it still says version 3.13. How can I uninstall all pythons and just install version 3.13?
What is your OS? how do you install python? from source code or yum/apt package manager? you can check the /usr/local/bin, /usr/bin, /usr/lib, /usr/local/lib. Or using "python3.12" instead of "python" command.
Python 3.12.8 on Debian (SID) here.
First I installed it with pipx and didn't manage to get torch to work with it... But I could at least run exo globally.
But it seems the correct way to install is (with venv) like this?:
git clone https://github.com/exo-explore/exo.git
cd exo
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
pip install torch --index-url https://download.pytorch.org/whl/cpu
But then the path of exo is not in the path and hence not globally accessible. So what is the right way to have it globally accessible and still run within the env?
EDIT: I ended up making this script to run it:
cd /home/joakim/Software/exo/
source .venv/bin/activate
exo
Just put it in e.g. ~/.local/bin/ as exo and make sure it can be executed...
