Closed kgarg8 closed 3 years ago
Hi @kgarg8, for training the baselines, we use the open-source code provided by Chan et al. (2019): https://github.com/kenchan0226/keyphrase-generation-rl
Once we have the output keyphrases on the testset, we run it through our evaluation scripts to obtain the quality and diversity metrics.
Makes sense, thanks Hareesh for the reply.
Hi,
Do you have have baseline model files e.g.
catSeq
,catSeqD
,catSeqCorr
,catSeqTG
,catSeqTG2RF1
? You must be using them for diversity evaluation.