datasets
huggingface_hub
nltk
numpy
pandas
peft
replicate
streamlit
torch
transformers==4.36.1
wandb
evaluate
rouge_score
bert_score