librispeech_asr

सन्दर्भ:

साफ

इस डेटासेट को TFDS में लोड करने के लिए निम्नलिखित कमांड का उपयोग करें:

ds = tfds.load('huggingface:librispeech_asr/clean')
  • विवरण :
LibriSpeech is a corpus of approximately 1000 hours of read English speech with sampling rate of 16 kHz,
prepared by Vassil Panayotov with the assistance of Daniel Povey. The data is derived from read
audiobooks from the LibriVox project, and has been carefully segmented and aligned.87

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .flac 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 soundfile as sf

def map_to_array(batch):
    speech_array, _ = sf.read(batch["file"])
    batch["speech"] = speech_array
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • लाइसेंस : कोई ज्ञात लाइसेंस नहीं
  • संस्करण : 2.1.0
  • विभाजन :
विभाजित करना उदाहरण
'test' 2620
'train.100' 28539
'train.360' 104014
'validation' 2703
  • विशेषताएँ :
{
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "audio": {
        "sampling_rate": 16000,
        "mono": true,
        "decode": true,
        "id": null,
        "_type": "Audio"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "speaker_id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "chapter_id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

अन्य

इस डेटासेट को TFDS में लोड करने के लिए निम्नलिखित कमांड का उपयोग करें:

ds = tfds.load('huggingface:librispeech_asr/other')
  • विवरण :
LibriSpeech is a corpus of approximately 1000 hours of read English speech with sampling rate of 16 kHz,
prepared by Vassil Panayotov with the assistance of Daniel Povey. The data is derived from read
audiobooks from the LibriVox project, and has been carefully segmented and aligned.87

Note that in order to limit the required storage for preparing this dataset, the audio
is stored in the .flac 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 soundfile as sf

def map_to_array(batch):
    speech_array, _ = sf.read(batch["file"])
    batch["speech"] = speech_array
    return batch

dataset = dataset.map(map_to_array, remove_columns=["file"])
  • लाइसेंस : कोई ज्ञात लाइसेंस नहीं
  • संस्करण : 2.1.0
  • विभाजन :
विभाजित करना उदाहरण
'test' 2939
'train.500' 148688
'validation' 2864
  • विशेषताएँ :
{
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "audio": {
        "sampling_rate": 16000,
        "mono": true,
        "decode": true,
        "id": null,
        "_type": "Audio"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "speaker_id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "chapter_id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

सभी

इस डेटासेट को TFDS में लोड करने के लिए निम्नलिखित कमांड का उपयोग करें:

ds = tfds.load('huggingface:librispeech_asr/all')
  • विवरण :
LibriSpeech is a corpus of approximately 1000 hours of read English speech with sampling rate of 16 kHz,
prepared by Vassil Panayotov with the assistance of Daniel Povey. The data is derived from read
audiobooks from the LibriVox project, and has been carefully segmented and aligned.87
  • लाइसेंस : कोई ज्ञात लाइसेंस नहीं
  • संस्करण : 2.1.0
  • विभाजन :
विभाजित करना उदाहरण
'test.clean' 2620
'test.other' 2939
'train.clean.100' 28539
'train.clean.360' 104014
'train.other.500' 148688
'validation.clean' 2703
'validation.other' 2864
  • विशेषताएँ :
{
    "file": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "audio": {
        "sampling_rate": 16000,
        "mono": true,
        "decode": true,
        "id": null,
        "_type": "Audio"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "speaker_id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "chapter_id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}