bert_score>=0.3.13 git+https://github.com/huggingface/evaluate@7d7d81dd3ffec0812e2edb09f86b3b1e31d61118 sacremoses>=0.0.53 absl-py nltk rouge_score>=0.1.2 # below install tensorflow and downgrades numpy, so heavy dependency git+https://github.com/google-research/bleurt.git