How can I get the predicted gap sentences in Pegasus? I notice the mask_token_sent function in PegasusTokenizer in huggingface, should I use it? If so, what 'str' should I put in the argument space? Can someone provide some suggestions on that please?
Hi,
How can I get the predicted gap sentences in Pegasus? I notice the
mask_token_sent
function in PegasusTokenizer in huggingface, should I use it? If so, what 'str' should I put in the argument space? Can someone provide some suggestions on that please?