Environment Installation

First, you could try

conda env create -f environment.yml
conda activate wrench

If it's not working, you can create a new conda enviroment and install the packages as below

conda env create -n wrench python=3.6
source activate wrench

Remember to set channel_priority: true in your .condarc.

or

conda config --set channel_priority true

Install transformers

conda install transformers -c huggingface

If your GLIBC version is old, try

conda install -c huggingface tokenizers=0.10.1 transformers=4.6.1

Then install other conda packages

conda install pytorch torchvision cudatoolkit=10.2 -c pytorch
conda install scikit-learn
conda install spacy==3.1.2
conda install numba==0.43
conda install seqeval
conda install networkx==2.3
conda install nltk==3.6.2
conda install pandas==1.1.5
conda install optuna==2.7.0

Then install pip packages

pip install pytokenizations==0.7.2
pip install sentence-transformers==1.1.1
pip install tensorboardx==1.4
pip install snorkel==0.9.7
pip install skweak==0.2.13
pip install snorkel-metal==0.5.0
pip install git+https://github.com/HazyResearch/flyingsquid.git
pip install pgmpy==0.1.14
pip install numbskull==0.1.1

If you want to use epoxy, you should do

conda install -c pytorch faiss-gpu cudatoolkit=10.2
pip install cytoolz==0.11.0

If you want to use learn_structure, you should do

conda install cvxpy==1.1.13

For meta-learning

pip install higher