The following instructions are for installing and executing the roma-llm case study on a linux/ubuntu machine with a GPU properly configured.
- Create a virtual environment using conda:
conda create -n roma-llm python=3.10
- Activate the environment:
conda activate roma-llm
- Install the requirements file
pip install -r requirements.txt
Pre-requisite - you should download the GLUE dataset to a folder named data.
python create_full_sentence_dataset.py
python predict_sentence_dataset_full.py
python analysis_full_prediction.py
Good Luck!