seaborn scipy datasets==2.14.5 gradio huggingface-hub==0.18.0 numpy==1.24.2 APScheduler==3.10.1 pandas==1.3.4 nervaluate==0.2.0 evaluate==0.4.0 rouge_score sacrebleu==2.4.2 transformers==4.17.0 scikit-learn==1.0.2 py-rouge==1.1 bert-score==0.3.13 sentencepiece==0.1.99 tokenizers==0.19.1