코보스트2

참고자료:

en_de

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_de')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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_tr

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_tr')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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_fa

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_fa')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_sv-SE')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_mn')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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_zh-CN

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_zh-CN')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_cy')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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_ca

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_ca')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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_sl

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_sl')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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"
    }
}

ko_et

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_et')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_id')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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"
    }
}

ko_ar

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_ar')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_ta')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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_lv

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_lv')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/en_ja')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 15531
'train' 289430
'validation' 15531
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/fr_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 14760
'train' 207374
'validation' 14760
  • 특징 :
{
    "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"
    }
}

드엔

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/de_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 13511
'train' 127834
'validation' 13511
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/es_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 13221
'train' 79015
'validation' 13221
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/ca_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 12730
'train' 95854
'validation' 12730
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/it_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 8951
'train' 31698
'validation' 8940
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/ru_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 6300
'train' 12112
'validation' 6110
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/zh-CN_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 4898
'train' 7085
'validation' 4843
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/pt_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 4023
'train' 9158
'validation' 3318
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/fa_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 3445
'train' 53949
'validation' 3445
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/et_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1571
'train' 1782년
'validation' 1576년
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/mn_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1759년
'train' 2067
'validation' 1761년
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/nl_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1699년
'train' 7108
'validation' 1699년
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/tr_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1629년
'train' 3966
'validation' 1624년
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/ar_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1695년
'train' 2283
'validation' 1758년
  • 특징 :
{
    "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_ko

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/sv-SE_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1595년
'train' 2160
'validation' 1349
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/lv_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 1629년
'train' 2337
'validation' 1125
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/sl_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 360
'train' 1843년
'validation' 509
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/ta_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 786
'train' 1358
'validation' 384
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/ja_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 684
'train' 1119
'validation' 635
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/id_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 844
'train' 1243
'validation' 792
  • 특징 :
{
    "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

TFDS에 이 데이터세트를 로드하려면 다음 명령어를 사용하세요.

ds = tfds.load('huggingface:covost2/cy_en')
  • 설명 :
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 Mozillas 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"])
  • 라이센스 : 알려진 라이센스 없음
  • 버전 : 1.0.0
  • 분할 :
나뉘다
'test' 690
'train' 1241
'validation' 690
  • 특징 :
{
    "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"
    }
}