ковост2

Ссылки:

ru_de

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_de')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_tr

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_tr')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_fa

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_fa')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_sv-SE

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_sv-SE')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_mn

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_mn')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_zh-CN

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_zh-CN')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_cy

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_cy')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_ca

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_ca')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_sl

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_sl')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_et

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_et')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_id

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_id')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_ar

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_ar')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_ta

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_ta')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_lv

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_lv')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

en_ja

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/en_ja')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 15531
'train' 289430
'validation' 15531
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

fr_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/fr_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 14760
'train' 207374
'validation' 14760
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

де_ен

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/de_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 13511
'train' 127834
'validation' 13511
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

es_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/es_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 13221
'train' 79015
'validation' 13221
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ca_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/ca_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 12730
'train' 95854
'validation' 12730
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

it_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/it_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 8951
'train' 31698
'validation' 8940
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ru_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/ru_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 6300
'train' 12112
'validation' 6110
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

zh-CN_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/zh-CN_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 4898
'train' 7085
'validation' 4843
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

pt_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/pt_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 4023
'train' 9158
'validation' 3318
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

fa_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/fa_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 3445
'train' 53949
'validation' 3445
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

et_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/et_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1571 г.
'train' 1782 г.
'validation' 1576 г.
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

mn_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/mn_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1759 г.
'train' 2067
'validation' 1761 г.
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

nl_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/nl_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1699 г.
'train' 7108
'validation' 1699 г.
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

tr_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/tr_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1629 г.
'train' 3966
'validation' 1624 г.
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ar_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/ar_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1695 г.
'train' 2283
'validation' 1758 г.
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

sv-SE_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/sv-SE_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1595 г.
'train' 2160
'validation' 1349
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

lv_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/lv_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 1629 г.
'train' 2337
'validation' 1125
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

sl_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/sl_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 360
'train' 1843 г.
'validation' 509
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ta_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/ta_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 786
'train' 1358
'validation' 384
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

ja_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/ja_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 684
'train' 1119
'validation' 635
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

id_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/id_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 844
'train' 1243
'validation' 792
  • Функции : 
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

cy_en

Используйте следующую команду, чтобы загрузить этот набор данных в TFDS:

ds = tfds.load('huggingface:covost2/cy_en')
  • Описание :
CoVoST 2, a large-scale multilingual speech translation corpus covering translations from 21 languages into English and from English into 15 languages. The dataset is created using Mozilla’s open source Common Voice database of crowdsourced voice recordings.

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .mp3 format and is not converted to a float32 array. To convert, the audio
file to a float32 array, please make use of the `.map()` function as follows:


python
import torchaudio

def map_to_array(batch):
    speech_array, _ = torchaudio.load(batch["file"])
    batch["speech"] = speech_array.numpy()
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • Лицензия : Нет известной лицензии.
  • Версия : 1.0.0
  • Расколы :
Расколоть Примеры
'test' 690
'train' 1241
'validation' 690
  • Функции :
{
    "client_id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}