Attend the Women in ML Symposium on December 7 Register now

ro_sts

References:

ro_sts

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:ro_sts/ro_sts')
  • Description:
The RO-STS (Romanian Semantic Textual Similarity) dataset contains 8628 pairs of sentences with their similarity score. It is a high-quality translation of the STS benchmark dataset.
  • License: CC BY-SA 4.0 License
  • Version: 1.0.0
  • Splits:
Split Examples
'test' 1379
'train' 5749
'validation' 1500
  • Features:
{
    "score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    },
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}