subtask_a_monolingual

The model is a CNN layer that takes the output of RoBERTa embeddings to classify a piece of text as machine generated or human generated.

Installation

run the following command:

pip install -r requirements.txt

Running the code

python cnn_roberta.py --training_data=path/to/train/data.jsonl --dev_data=path/to/dev/data.jsonl --output_file=path/to/output.jsonl