Clone repo (recommended for developers)

1. from home dir, git clone git@github.com:bnb32/modbot.git

2. Create modbot environment and install package

   1. Create a conda env: conda create -n modbot

   2. Run the command: conda activate modbot

   3. cd into the repo cloned in 1.

   4. prior to running pip below, make sure the branch is correct (install from main!)

   5. Install modbot and its dependencies by running: pip install . (or pip install -e . if running a dev branch or working on the source code)

   6. To train on a GPU run pip install tensorflow-gpu and conda install -c conda-forge cudnn

   7. If using a BERT model download a preprocessing model and encoder and put in data/bert_preprocess and data/bert_encoder, respectively. Example downloads: https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3 and https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/4