llama-cpp-python==0.1.78 
numpy==1.23.4
huggingface_hub
langchain-text-splitters
lmqg
openai

# for fine tunning
transformers
datasets
wandb
rouge_score
trl
peft
accelerate


