README.md

metadata

language:
  - en
task_categories:
  - summarization

AutoTrain Dataset for project: test-summarization

Dataset Description

This dataset has been automatically processed by AutoTrain for project test-summarization.

Languages

The BCP-47 code for the dataset's language is en.

Dataset Structure

Data Instances

A sample from this dataset looks as follows:

[
  {
    "feat_id": "13829542",
    "text": "Kasia: When are u coming back?\r\nMatt: Back where?\r\nKasia: Oh come on\r\nKasia: you know what i mean\r\nMatt: I really don't \r\nKasia: When are you coming back to Warsaw\r\nMatt: I have no idea\r\nMatt: maybe around easter\r\nKasia: will you let me know\r\nMatt: sure if I know something then I will let you know asap\r\nKasia: ok \r\nMatt: are you mad?\r\nKasia: a bit\r\nMatt: oh come on\r\nMatt: this is not my fault \r\nMatt: there is no way that I can answer that question\r\nMatt: not now\r\nKasia: Fine",
    "target": "Matt doesn't know when he's coming back to Warsaw. He might come around Easter. When he knows more, he will let Kasia know. Kasia is a bit upset."
  },
  {
    "feat_id": "13862523",
    "text": "Oliver: Have you beaten the game yet?\nTom: Not yet\nOliver: Ok... what mission are you playing?\nTom: The one before the final one, it's pretty hard\nOliver: I didn't find it particularly hard\nTom: I mean, combat is easy at this point in the game but the puzzles are difficult\nOliver: Ok, I got it\nTom: It's fun how most horror action games let you get really powerful by the end of the story\nOliver: Well, you know, being a pussy from start to finish ain't my idea of fun even in a horror game\nTom: I know... but do you remember Bioshock? You turned into some sort of superhero and the final boss was pretty much a joke\nOliver: Don't you dare talk like that about my favorite game\nTom: I know, I love it too, but it had its flaws\nOliver: No it didn't XD\nTom: Lol\nOliver: Well, keep playing\nTom: I'll let you know when I finish this one and my overall impressions\nOliver: Ok",
    "target": "Tom will contact Oliver after finishing new horror action game."
  }
]

Dataset Fields

The dataset has the following fields (also called "features"):

{
  "feat_id": "Value(dtype='string', id=None)",
  "text": "Value(dtype='string', id=None)",
  "target": "Value(dtype='string', id=None)"
}

Dataset Splits

This dataset is split into a train and validation split. The split sizes are as follow:

Split name	Num samples
train	11785
valid	2947