Text Generation On Cnn Daily Mail

ROUGE-L

Results

Performance results of various models on this benchmark

Comparison Table

Model Name	ROUGE-L
palm-pre-training-an-autoencoding	41.41