Papers with Code · CC BY-SA 4.0
Dialogue
Dialogue is notoriously hard to evaluate. Past approaches have used human evaluation.
데이터셋
72개Visual Dialog v1.0 test-stdFluent Speech CommandsVisDial v0.9 valSwitchboard corpusKVRETWizard-of-OzCoSQLLRE07ICSI Meeting Recorder Dialog Act (MRDA) corpusSecond dialogue state tracking challengeSnips-SmartLightsMULTIWOZ 2.0Persona-ChatSIMMC2.0Snips-SmartSpeakerVoxForge EuropeanYouTube News dataset (No Noise)YouTube News dataset (White Noise)irc-disentanglementrt-inod-jailbreakingMULTIWOZ 2.1MULTIWOZ 2.1OpenViDial 2.0Spoken-SQuADUntranscribed mixed-speech datasetVoxForge CommonwealthDeliDataIndicTTSLinux IRC (Ch2 Elsner)Linux IRC (Ch2 Kummerfeld)