SARCAST-X - Copenhagen Business School thesis project on sarcasm and fake-news detection in climatechange-related tweets (X)

Folders:

TO DO BEFORE FIRST RUN:

TO DO IF YOU WANT TO RETRAIN THE MODELS:

• DELETE the sarcasm_model and/or fake_news_model folders in the models folder
• DELETE the results folder

Environment Variables

In the params.py file, set the following variables:

• SARCASM_MODEL_PATH: Path to the sarcasm model on huggingface
• FAKE_NEWS_MODEL_PATH: Path to the fake news model on huggingface
• CUDA: boolean, False if you have a GPU and do not want to use it, True otherwise
• SARCASM_MODEL_SAVED: boolean, True if the sarcasm model is saved locally, False otherwise
• FAKE_NEWS_MODEL_SAVED: boolean, True if the fake news model is saved locally, False otherwise

Model Architectures

• model_functions.py contains all the model finetuning functions that are imported in main.py

Model Save and MLFLOW logic

• registery.py contains the logic to save the model in the MLFLOW registry and load existing models from the registry.

Running the Code

To run the code, execute python ml_logic/main.py or make run main