Difficulties to reproduce summarization results (CNNDM) about pegasus HOT 2 CLOSED

Sorry, my mistake.
I just found out from here that rougeLsum is a summary-level ROUGE-L score.
rougeLsum-F,0.397665,0.406826,0.415280

Assuming from here, the second score looks matching with the reported score.

Scores reproduced! Once again, my apologies for the misinterpretation.
Best,
Wonjin

from pegasus.

Hi Wonjin, I am also trying to replicate the results in the paper. You say that you reproduced the scores, However, I cannot find the score that you posted anywhere in the paper. Could you please specify (with e.g. a screenshot) which score you have replicated with your code:

python pegasus/bin/evaluate.py --params=cnn_dailymail_transformer \
> --param_overrides=vocab_filename=ckpt/pegasus_ckpt/c4.unigram.newline.10pct.96000.model,batch_size=1,beam_size=8,beam_alpha=0.8 \
> --model_dir=ckpt/pegasus_ckpt/cnn_dailymail/model.ckpt-210000 --evaluate_test

from pegasus.