Datasets:

Multilinguality:
multilingual
Size Categories:
1K<n<10K
Annotations Creators:
crowdsourced
ArXiv:
Tags:
License:
albertvillanova HF staff commited on
Commit
89977ca
1 Parent(s): aa7aa32

Add X-CODAH-vi data files

Browse files
README.md CHANGED
@@ -464,13 +464,13 @@ dataset_info:
464
  dtype: string
465
  splits:
466
  - name: test
467
- num_bytes: 543375
468
  num_examples: 1000
469
  - name: validation
470
- num_bytes: 157000
471
  num_examples: 300
472
- download_size: 7519903
473
- dataset_size: 700375
474
  - config_name: X-CODAH-zh
475
  features:
476
  - name: id
@@ -999,6 +999,12 @@ configs:
999
  path: X-CODAH-ru/test-*
1000
  - split: validation
1001
  path: X-CODAH-ru/validation-*
 
 
 
 
 
 
1002
  - config_name: X-CODAH-zh
1003
  data_files:
1004
  - split: test
 
464
  dtype: string
465
  splits:
466
  - name: test
467
+ num_bytes: 543089
468
  num_examples: 1000
469
  - name: validation
470
+ num_bytes: 156888
471
  num_examples: 300
472
+ download_size: 251390
473
+ dataset_size: 699977
474
  - config_name: X-CODAH-zh
475
  features:
476
  - name: id
 
999
  path: X-CODAH-ru/test-*
1000
  - split: validation
1001
  path: X-CODAH-ru/validation-*
1002
+ - config_name: X-CODAH-vi
1003
+ data_files:
1004
+ - split: test
1005
+ path: X-CODAH-vi/test-*
1006
+ - split: validation
1007
+ path: X-CODAH-vi/validation-*
1008
  - config_name: X-CODAH-zh
1009
  data_files:
1010
  - split: test
X-CODAH-vi/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c17c93515e64d619de5aaeb508c5a04b85ed7434952c67a7c9afb5f3ec2a1c9
3
+ size 192251
X-CODAH-vi/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec45f6022ab728806e28445d9373f2ab5faee95f8b5a4b3bd745c042f2bd41bf
3
+ size 59139
dataset_infos.json CHANGED
@@ -1903,53 +1903,42 @@
1903
  "features": {
1904
  "id": {
1905
  "dtype": "string",
1906
- "id": null,
1907
  "_type": "Value"
1908
  },
1909
  "lang": {
1910
  "dtype": "string",
1911
- "id": null,
1912
  "_type": "Value"
1913
  },
1914
  "question_tag": {
1915
  "dtype": "string",
1916
- "id": null,
1917
  "_type": "Value"
1918
  },
1919
  "question": {
1920
  "stem": {
1921
  "dtype": "string",
1922
- "id": null,
1923
  "_type": "Value"
1924
  },
1925
  "choices": {
1926
  "feature": {
1927
  "label": {
1928
  "dtype": "string",
1929
- "id": null,
1930
  "_type": "Value"
1931
  },
1932
  "text": {
1933
  "dtype": "string",
1934
- "id": null,
1935
  "_type": "Value"
1936
  }
1937
  },
1938
- "length": -1,
1939
- "id": null,
1940
  "_type": "Sequence"
1941
  }
1942
  },
1943
  "answerKey": {
1944
  "dtype": "string",
1945
- "id": null,
1946
  "_type": "Value"
1947
  }
1948
  },
1949
- "post_processed": null,
1950
- "supervised_keys": null,
1951
- "task_templates": null,
1952
  "builder_name": "xcsr",
 
1953
  "config_name": "X-CODAH-vi",
1954
  "version": {
1955
  "version_str": "1.1.0",
@@ -1961,27 +1950,20 @@
1961
  "splits": {
1962
  "test": {
1963
  "name": "test",
1964
- "num_bytes": 543375,
1965
  "num_examples": 1000,
1966
- "dataset_name": "xcsr"
1967
  },
1968
  "validation": {
1969
  "name": "validation",
1970
- "num_bytes": 157000,
1971
  "num_examples": 300,
1972
- "dataset_name": "xcsr"
1973
- }
1974
- },
1975
- "download_checksums": {
1976
- "https://inklab.usc.edu/XCSR/xcsr_datasets.zip": {
1977
- "num_bytes": 7519903,
1978
- "checksum": "c45b29ece740643252d5402e76be1e33f96f9d6910053f79e80d39887f10c85e"
1979
  }
1980
  },
1981
- "download_size": 7519903,
1982
- "post_processing_size": null,
1983
- "dataset_size": 700375,
1984
- "size_in_bytes": 8220278
1985
  },
1986
  "X-CODAH-hi": {
1987
  "description": "To evaluate multi-lingual language models (ML-LMs) for commonsense reasoning in a cross-lingual zero-shot transfer setting (X-CSR), i.e., training in English and test in other languages, we create two benchmark datasets, namely X-CSQA and X-CODAH. Specifically, we automatically translate the original CSQA and CODAH datasets, which only have English versions, to 15 other languages, forming development and test sets for studying X-CSR. As our goal is to evaluate different ML-LMs in a unified evaluation protocol for X-CSR, we argue that such translated examples, although might contain noise, can serve as a starting benchmark for us to obtain meaningful analysis, before more human-translated datasets will be available in the future.\n",
 
1903
  "features": {
1904
  "id": {
1905
  "dtype": "string",
 
1906
  "_type": "Value"
1907
  },
1908
  "lang": {
1909
  "dtype": "string",
 
1910
  "_type": "Value"
1911
  },
1912
  "question_tag": {
1913
  "dtype": "string",
 
1914
  "_type": "Value"
1915
  },
1916
  "question": {
1917
  "stem": {
1918
  "dtype": "string",
 
1919
  "_type": "Value"
1920
  },
1921
  "choices": {
1922
  "feature": {
1923
  "label": {
1924
  "dtype": "string",
 
1925
  "_type": "Value"
1926
  },
1927
  "text": {
1928
  "dtype": "string",
 
1929
  "_type": "Value"
1930
  }
1931
  },
 
 
1932
  "_type": "Sequence"
1933
  }
1934
  },
1935
  "answerKey": {
1936
  "dtype": "string",
 
1937
  "_type": "Value"
1938
  }
1939
  },
 
 
 
1940
  "builder_name": "xcsr",
1941
+ "dataset_name": "xcsr",
1942
  "config_name": "X-CODAH-vi",
1943
  "version": {
1944
  "version_str": "1.1.0",
 
1950
  "splits": {
1951
  "test": {
1952
  "name": "test",
1953
+ "num_bytes": 543089,
1954
  "num_examples": 1000,
1955
+ "dataset_name": null
1956
  },
1957
  "validation": {
1958
  "name": "validation",
1959
+ "num_bytes": 156888,
1960
  "num_examples": 300,
1961
+ "dataset_name": null
 
 
 
 
 
 
1962
  }
1963
  },
1964
+ "download_size": 251390,
1965
+ "dataset_size": 699977,
1966
+ "size_in_bytes": 951367
 
1967
  },
1968
  "X-CODAH-hi": {
1969
  "description": "To evaluate multi-lingual language models (ML-LMs) for commonsense reasoning in a cross-lingual zero-shot transfer setting (X-CSR), i.e., training in English and test in other languages, we create two benchmark datasets, namely X-CSQA and X-CODAH. Specifically, we automatically translate the original CSQA and CODAH datasets, which only have English versions, to 15 other languages, forming development and test sets for studying X-CSR. As our goal is to evaluate different ML-LMs in a unified evaluation protocol for X-CSR, we argue that such translated examples, although might contain noise, can serve as a starting benchmark for us to obtain meaningful analysis, before more human-translated datasets will be available in the future.\n",