oscar

References:

• Code
• Huggingface

unshuffled_deduplicated_af

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_af')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	130640

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_als

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_als')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4518

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_arz

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_arz')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	79928

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_an

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_an')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2025

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ast

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ast')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5343

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ba

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ba')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	27050

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_am

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_am')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	43102

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_as

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_as')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	9212

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_azb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_azb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	9985

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_be

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_be')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	307405

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	15762

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bxr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bxr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	36

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ceb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ceb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	26145

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_az

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_az')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	626796

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bcl

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bcl')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_cy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_cy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	98225

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_dsb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_dsb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	37

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1114481

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bs

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bs')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	702

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ce

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ce')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2984

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_cv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_cv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	10130

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_diq

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_diq')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_eml

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_eml')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	80

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_et

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_et')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1172041

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bg

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bg')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3398679

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bpy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bpy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1770

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ca

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ca')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2458067

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ckb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ckb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	68210

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ar

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ar')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	9006977

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_av

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_av')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	360

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bar

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bar')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_bh

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_bh')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	82

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_br

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_br')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	14724

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_cbk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_cbk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_da

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_da')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4771098

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_dv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_dv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	17024

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_eo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_eo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	84752

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_fa

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_fa')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	8203495

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_fy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_fy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	20661

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_gn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_gn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	68

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_cs

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_cs')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	12308039

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_hi

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_hi')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1909387

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_hu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_hu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6582908

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ie

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ie')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	11

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_fr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_fr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	59448891

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_gd

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_gd')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3883

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_gu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_gu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	169834

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_hsb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_hsb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3084

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ia

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ia')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	529

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_io

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_io')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	617

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_jbo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_jbo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	617

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_km

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_km')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	108346

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ku

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ku')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	29054

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_la

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_la')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	18808

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_lmo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_lmo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1374

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_lv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_lv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	843195

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_min

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_min')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	166

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	212556

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mwl

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mwl')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_nah

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_nah')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	58

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_new

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_new')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2126

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_oc

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_oc')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6485

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_pam

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_pam')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ps

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ps')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	67921

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_it

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_it')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	28522082

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ka

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ka')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	372158

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ro

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ro')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5044757

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_scn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_scn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	17

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ko

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ko')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3675420

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_kw

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_kw')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	68

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_lez

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_lez')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1381

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_lrc

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_lrc')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	72

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mg

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mg')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	13343

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ml

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ml')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	453904

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ms

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ms')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	183443

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_myv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_myv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_nds

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_nds')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	8714

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_nn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_nn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	109118

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_os

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_os')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2559

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_pms

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_pms')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2859

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_qu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_qu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	411

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sa

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sa')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7121

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2820821

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sh

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sh')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	17610

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_so

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_so')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	42

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	645747

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ta

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ta')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	833101

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_tk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_tk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4694

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_tyv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_tyv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	24

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_uz

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_uz')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	15074

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_wa

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_wa')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	677

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_xmf

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_xmf')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2418

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	11014487

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_tg

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_tg')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	56259

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_de

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_de')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	62398034

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_tr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_tr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	11596446

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_el

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_el')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6521169

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_uk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_uk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7782375

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_vi

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_vi')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	9897709

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_wuu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_wuu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	64

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_yo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_yo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	49

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_als

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_als')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7324

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_arz

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_arz')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	158113

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_az

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_az')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	912330

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bcl

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bcl')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1675515

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bs

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bs')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2143

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ce

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ce')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4042

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_cv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_cv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	20281

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_diq

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_diq')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_eml

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_eml')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	84

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_et

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_et')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2093621

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_zh

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_zh')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	41708901

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_an

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_an')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2449

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ast

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ast')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6999

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ba

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ba')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	42551

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bg

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bg')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5869686

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bpy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bpy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6046

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ca

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ca')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4390754

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ckb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ckb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	103639

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_es

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_es')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	56326016

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_da

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_da')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7664010

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_dv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_dv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	21018

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_eo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_eo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	121168

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_fi

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_fi')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5326443

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ga

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ga')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	46493

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_gom

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_gom')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	484

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_hr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_hr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	321484

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_hy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_hy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	396093

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ilo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ilo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1578

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_fa

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_fa')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	13704702

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_fy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_fy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	33053

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_gn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_gn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	106

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_hi

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_hi')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3264660

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_hu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_hu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	11197780

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ie

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ie')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	101

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ja

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ja')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	39496439

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_kk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_kk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	338073

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_krc

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_krc')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1377

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ky

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ky')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	86561

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_li

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_li')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	118

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_lt

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_lt')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1737411

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mhr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mhr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	2515

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	197878

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mt

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mt')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	16383

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_mzn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_mzn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	917

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ne

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ne')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	219334

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_no

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_no')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3229940

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_pa

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_pa')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	87235

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_pnb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_pnb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3463

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_rm

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_rm')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	34

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sah

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sah')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	8555

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_si

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_si')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	120684

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sq

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sq')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	461598

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_sw

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_sw')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	24803

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_th

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_th')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3749826

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_tt

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_tt')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	82738

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_ur

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_ur')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	428674

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_vo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_vo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3317

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_xal

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_xal')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	36

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_deduplicated_yue

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_deduplicated_yue')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_am

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_am')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	83663

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_as

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_as')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	14985

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_azb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_azb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	15446

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_be

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_be')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	586031

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	26795

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_bxr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_bxr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	42

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ceb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ceb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	56248

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_cy

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_cy')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	157698

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_dsb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_dsb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	65

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_fr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_fr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	96742378

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_gd

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_gd')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5799

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_gu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_gu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	240691

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_hsb

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_hsb')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7959

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ia

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ia')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1040

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_io

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_io')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	694

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_jbo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_jbo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	832

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_km

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_km')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	159363

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ku

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ku')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	46535

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_la

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_la')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	94588

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_lmo

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_lmo')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1401

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_lv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_lv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1593820

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_min

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_min')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	220

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_mr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_mr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	326804

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_mwl

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_mwl')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	8

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_nah

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_nah')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	61

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_new

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_new')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	4696

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_oc

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_oc')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	10709

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_pam

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_pam')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ps

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ps')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	98216

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ro

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ro')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	9387265

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_scn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_scn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	21

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_sk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_sk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5492194

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_sr

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_sr')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1013619

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ta

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ta')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1263280

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_tk

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_tk')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6456

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_tyv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_tyv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	34

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_uz

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_uz')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	27537

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_wa

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_wa')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1001

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_xmf

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_xmf')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3783

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_it

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_it')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	46981781

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ka

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ka')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	563916

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ko

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ko')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	7345075

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_kw

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_kw')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	203

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_lez

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_lez')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	1485

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_lrc

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_lrc')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	88

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_mg

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_mg')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	17957

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ml

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ml')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	603937

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_ms

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_ms')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	534016

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_myv

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_myv')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	6

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_nds

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_nds')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	18174

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_nn

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_nn')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	185884

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_os

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_os')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	5213

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_pms

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_pms')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	3225

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}

unshuffled_original_qu

Use the following command to load this dataset in TFDS:

ds = tfds.load('huggingface:oscar/unshuffled_original_qu')

• Description:

The Open Super-large Crawled ALMAnaCH coRpus is a huge multilingual corpus obtained by language classification and filtering of the Common Crawl corpus using the goclassy architecture.

• License: These data are released under this licensing scheme We do not own any of the text from which these data has been extracted. We license the actual packaging of these data under the Creative Commons CC0 license ("no rights reserved") http://creativecommons.org/publicdomain/zero/1.0/ To the extent possible under law, Inria has waived all copyright and related or neighboring rights to OSCAR This work is published from: France.

  Should you consider that our data contains material that is owned by you and should therefore not be reproduced here, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.

  We will comply to legitimate requests by removing the affected sources from the next release of the corpus.

• Version: 1.0.0

• Splits:

Split	Examples
'train'	452

• Features:

{
    "id": {
        "dtype": "int64",
        "id": null,
        "_type": "Value"
    },
    "text": {
        "dtype": "string",