https://github.com/schoobani/persian-generative-chatbot

A repo dedicated to different approaches in building a Persian Generative Chatbot.
https://github.com/schoobani/persian-generative-chatbot

chatbot persian text-generation

Last synced: about 2 months ago
JSON representation

A repo dedicated to different approaches in building a Persian Generative Chatbot.

• Host: GitHub
• URL: https://github.com/schoobani/persian-generative-chatbot
• Owner: schoobani
• Created: 2022-07-16T10:05:24.000Z (over 2 years ago)
• Default Branch: master
• Last Pushed: 2022-09-07T18:58:23.000Z (over 2 years ago)
• Last Synced: 2024-10-26T23:53:50.696Z (3 months ago)
• Topics: chatbot, persian, text-generation
• Language: Jupyter Notebook
• Homepage:
• Size: 507 KB
• Stars: 9
• Watchers: 1
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
### Persian Generative Chatbot

A repo dedicated to different approaches in building a Persian Generative Chatbot.

### Data

We need a dataset of conversation pairs to build a chatbot.

### 1- Ninisite

[ninisite](ninisite.com) is a persian forum with millions of conversation pairs on different life-style topics. We wrote a [simple script](/data/ninisite/) to crawl the conversation pairs, like the following sample:

```

{

    "topic": "بیمارستان هدایت تهران برای زایمان چطوره؟",

    "question": [

        "سلام خانما",

        "من تو بیمارستان میلاد برای بارداری و زایمان پرونده دارم.اما بیمار کرونایی بستری داره و فعلا هم تعطیله.",

        "شنیدم بیمارستان هدایت فقط زنان و زایمانه؟کسی چکاپ های بارداری و زایمانش رو تو هدایت انجام داده؟راضی بودین؟",

        "لطفا راهنماییم کنید"

        ],

    "answer": [

        "آره من 26سال پیش اونجا بدنیا اومدم مامانم ک خیلی تعریف میکرد ."

        ]

}

```

The following table shows some stats on the process of crawling the dataset (The table is getting updatad constantly).

| Total Targeted Topics |  Crawled   | Crawled Conversation Pairs |  Size  |

| :-------------------: | :--------: | :------------------------: | :----: |

|        636921         | 48708 (7%) |          1221025           | 778 MB |