These datasets were collected during ConvAI2 competition. Every
json file contains dialogues for PersonaChat task.
- data_tolokers.json – data collected during DeepHack.Chat hackathon in July 2-8 2018 via Yandex.Toloka service (paid workers). 3127 dialogues.
- data_intermediate.json – dialogues by the bots from DeepHack.Chat and volunteers collected from July 9 to October 29, 2018. 291 dialogues
- data_volunteers.json – wild evaluation for ConvAI final round from October 29 to December 17, 2018. Dialogues by bots from ConvAI finals and volunteers. 1111 dialogues
The whole dataset collected during DeepHack.Chat and ConvAI2 Wild Evaluation round: summer_wild_evaluation_dialogs.json.
The data collected during other evaluation is available though Parl.ai framework here.
The 2017 dataset can be found here.