The dataset used to fine-tune the GPT-2 model used in Anees for the multi-turn dialogue generation.

The dataset is a combination of 4 multi-turn dialogue datasets:

• DailyDialog: a high-quality multi-turn open-domain English dialog dataset. On average there are around 8 speaker turns per dialogue with around 15 tokens per turn.
• EmpatheticDialogues: a large-scale multi-turn empathetic dialogue dataset collected on Amazon Mechanical Turk, containing 24,850 one-to-one open-domain conversations.
• Persona-Chat: crowd-sourced dialogues where each participant plays the part of an assigned persona; and each persona has a word-distinct paraphrase.
• BlendedSkillTalk: an English-language dataset blending three conversation skills in balanced proportions (demonstrating knowledge, empathy, or ability to talk about oneself).

• The English dataset was tokenized using the GPT2 Tokenizer.
• The Arabic dataset was tokenized using the AraGPT2 Tokenizer.
• The translation from English to Arabic was done using Opus-MT on Colab.
• The preprocessing and loading details of the data can be found on Anees repository.

• Li, Y., Su, H., Shen, X., Li, W., Cao, Z., & Niu, S. (2017). Dailydialog: A manually labelled multi-turn dialogue dataset. arXiv preprint arXiv:1710.03957.
• Rashkin, H., Smith, E. M., Li, M., & Boureau, Y. L. (2018). Towards empathetic open-domain conversation models: A new benchmark and dataset. arXiv preprint arXiv:1811.00207.
• Zhang, S., Dinan, E., Urbanek, J., Szlam, A., Kiela, D., & Weston, J. (2018). Personalizing dialogue agents: I have a dog, do you have pets too?. arXiv preprint arXiv:1801.07243.
• Smith, E. M., Williamson, M., Shuster, K., Weston, J., & Boureau, Y. L. (2020). Can You Put it All Together: Evaluating Conversational Agents' Ability to Blend Skills. arXiv preprint arXiv:2004.08449.