lccc

Referências:

grande

Use o seguinte comando para carregar este conjunto de dados no TFDS:

ds = tfds.load('huggingface:lccc/large')
  • Descrição :
LCCC: Large-scale Cleaned Chinese Conversation corpus (LCCC) is a large corpus of Chinese conversations.
A rigorous data cleaning pipeline is designed to ensure the quality of the corpus.
This pipeline involves a set of rules and several classifier-based filters.
Noises such as offensive or sensitive words, special symbols, emojis,
grammatically incorrect sentences, and incoherent conversations are filtered.
  • Licença : MIT
  • Versão : 1.0.0
  • Divisões :
Dividir Exemplos
'train' 12007759
  • Características :
{
    "dialog": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}

base

Use o seguinte comando para carregar este conjunto de dados no TFDS:

ds = tfds.load('huggingface:lccc/base')
  • Descrição :
LCCC: Large-scale Cleaned Chinese Conversation corpus (LCCC) is a large corpus of Chinese conversations.
A rigorous data cleaning pipeline is designed to ensure the quality of the corpus.
This pipeline involves a set of rules and several classifier-based filters.
Noises such as offensive or sensitive words, special symbols, emojis,
grammatically incorrect sentences, and incoherent conversations are filtered.
  • Licença : MIT
  • Versão : 1.0.0
  • Divisões :
Dividir Exemplos
'test' 10.000
'train' 6820506
'validation' 20.000
  • Características :
{
    "dialog": [
        {
            "dtype": "string",
            "id": null,
            "_type": "Value"
        }
    ]
}