LogFlow installation¶
How to install it ?
You can do it using pip, docker or pulling from Github. The recommended way is to use Docker. Note that nvidia-docker is recommended if you plan to use GPU for the learning step.
Docker¶
Using the Dockerfile at the root of this git.
docker build -t logflow .
Then, start the docker using nvidia-docker
nvidia-docker run -it --pid=host -v $(pwd):/home/code logflow /bin/bash
Now, you are inside your docker
cd /home/code
Create a data and a model directory
mkdir data
mkdir model
Pip¶
pip install -i https://test.pypi.org/simple/ --no-cache-dir --extra-index-url=https://pypi.org/simple/ LogFlow-Atos-Marc-Platini
Note that it’s a pre-alpha pip version. Usage of docker is recommended instead of pip.
Now, you are ready to use LogFlow !