alakxender/dhivehi-similarity-tasks

Dhivehi STS

The Dhivehi Semantic Textual Similarity dataset is a translated version. The English sentence pairs were translated into Dhivehi, while keeping the original similarity scores. Each pair is annotated with a similarity score normalized between 0 and 1. Dataset Details

• Columns: "sentence1", "sentence2", "score"
• Column types: str, str, float

Examples:

{
  "sentence1": "މިއީ ރަނގަޅު ޚިޔާލެއް ނޫނެވެ.",
  "sentence2": "މިއީ ވަރަށް ނުބައި ޚިޔާލެއް ކަހަލަ އަޑެކެވެ.",
  "score": 0.8
}