sst

References:

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:sst')
  • Description:
The Stanford Sentiment Treebank, the first corpus with fully labeled parse trees that allows for a
complete analysis of the compositional effects of sentiment in language.
  • License: No known license
  • Version: 1.0.0
  • Splits:
Split Examples
'test' 2210
'train' 8544
'validation' 1101
  • Features:
{
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    },
    "tokens": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "tree": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

dictionary

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:sst/dictionary')
  • Description:
The Stanford Sentiment Treebank, the first corpus with fully labeled parse trees that allows for a
complete analysis of the compositional effects of sentiment in language.
  • License: No known license
  • Version: 1.0.0
  • Splits:
Split Examples
'dictionary' 239232
  • Features:
{
    "phrase": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}

ptb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:sst/ptb')
  • Description:
The Stanford Sentiment Treebank, the first corpus with fully labeled parse trees that allows for a
complete analysis of the compositional effects of sentiment in language.
  • License: No known license
  • Version: 1.0.0
  • Splits:
Split Examples
'test' 2210
'train' 8544
'validation' 1101
  • Features:
{
    "ptb_tree": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}