I thought I would share how I got local setup working without a lot of a hassle.
I am assuming one has already cloned the repo https://github.com/fastai/course-v4 and it is on the repo folder.
To get it working I have updated my environment.yml
to the snippet below , (There is PR to get this merged).
name: fastai
channels:
- fastai
- pytorch
- defaults
dependencies:
- jupyter
- pytorch>=1.3.0
- torchvision>=0.5
- matplotlib
- pandas
- requests
- pyyaml
- fastprogress>=0.1.22
- pillow
- python>=3.6
- scikit-learn
- scipy
- spacy
- pip
- pip:
- graphviz
- fastbook
Now one can create an environment using :
conda env create --file environment.yml
To activate the environment:
conda activate fastai
And to run the jupyter notebooks:
jupyter notebook
Hope this helps out those that want to run this locally.
In case you have conflict because fastai environment already exists, you can give it a new name, or remove it. I recommend removing the older environment (if you can).
To replace and old fastai environment with the new suggest one follow these steps:
conda deactivate
conda env remove --name fastai
conda env create --file environment.yml
conda activate fastai