Datasets:

Multilinguality:
multilingual
Size Categories:
1K<n<10K
Annotations Creators:
crowdsourced
ArXiv:
Tags:
License:
albertvillanova HF staff commited on
Commit
643641c
1 Parent(s): f22aacf

Add X-CSQA-vi data files

Browse files
README.md CHANGED
@@ -898,13 +898,13 @@ dataset_info:
898
  dtype: string
899
  splits:
900
  - name: test
901
- num_bytes: 265512
902
  num_examples: 1074
903
  - name: validation
904
- num_bytes: 253784
905
  num_examples: 1000
906
- download_size: 7519903
907
- dataset_size: 519296
908
  - config_name: X-CSQA-zh
909
  features:
910
  - name: id
@@ -999,6 +999,12 @@ configs:
999
  path: X-CSQA-ru/test-*
1000
  - split: validation
1001
  path: X-CSQA-ru/validation-*
 
 
 
 
 
 
1002
  - config_name: X-CSQA-zh
1003
  data_files:
1004
  - split: test
 
898
  dtype: string
899
  splits:
900
  - name: test
901
+ num_bytes: 265210
902
  num_examples: 1074
903
  - name: validation
904
+ num_bytes: 253502
905
  num_examples: 1000
906
+ download_size: 244641
907
+ dataset_size: 518712
908
  - config_name: X-CSQA-zh
909
  features:
910
  - name: id
 
999
  path: X-CSQA-ru/test-*
1000
  - split: validation
1001
  path: X-CSQA-ru/validation-*
1002
+ - config_name: X-CSQA-vi
1003
+ data_files:
1004
+ - split: test
1005
+ path: X-CSQA-vi/test-*
1006
+ - split: validation
1007
+ path: X-CSQA-vi/validation-*
1008
  - config_name: X-CSQA-zh
1009
  data_files:
1010
  - split: test
X-CSQA-vi/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2239d72cee5c8630aacc881383bb60bb98831186583c6d6aa483fc271d18c32c
3
+ size 126459
X-CSQA-vi/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76936244047b20be2a9266f655e57d417b035dd69a574312a0aee6f0c39a3e4e
3
+ size 118182
dataset_infos.json CHANGED
@@ -799,48 +799,38 @@
799
  "features": {
800
  "id": {
801
  "dtype": "string",
802
- "id": null,
803
  "_type": "Value"
804
  },
805
  "lang": {
806
  "dtype": "string",
807
- "id": null,
808
  "_type": "Value"
809
  },
810
  "question": {
811
  "stem": {
812
  "dtype": "string",
813
- "id": null,
814
  "_type": "Value"
815
  },
816
  "choices": {
817
  "feature": {
818
  "label": {
819
  "dtype": "string",
820
- "id": null,
821
  "_type": "Value"
822
  },
823
  "text": {
824
  "dtype": "string",
825
- "id": null,
826
  "_type": "Value"
827
  }
828
  },
829
- "length": -1,
830
- "id": null,
831
  "_type": "Sequence"
832
  }
833
  },
834
  "answerKey": {
835
  "dtype": "string",
836
- "id": null,
837
  "_type": "Value"
838
  }
839
  },
840
- "post_processed": null,
841
- "supervised_keys": null,
842
- "task_templates": null,
843
  "builder_name": "xcsr",
 
844
  "config_name": "X-CSQA-vi",
845
  "version": {
846
  "version_str": "1.1.0",
@@ -852,27 +842,20 @@
852
  "splits": {
853
  "test": {
854
  "name": "test",
855
- "num_bytes": 265512,
856
  "num_examples": 1074,
857
- "dataset_name": "xcsr"
858
  },
859
  "validation": {
860
  "name": "validation",
861
- "num_bytes": 253784,
862
  "num_examples": 1000,
863
- "dataset_name": "xcsr"
864
- }
865
- },
866
- "download_checksums": {
867
- "https://inklab.usc.edu/XCSR/xcsr_datasets.zip": {
868
- "num_bytes": 7519903,
869
- "checksum": "c45b29ece740643252d5402e76be1e33f96f9d6910053f79e80d39887f10c85e"
870
  }
871
  },
872
- "download_size": 7519903,
873
- "post_processing_size": null,
874
- "dataset_size": 519296,
875
- "size_in_bytes": 8039199
876
  },
877
  "X-CSQA-hi": {
878
  "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",
 
799
  "features": {
800
  "id": {
801
  "dtype": "string",
 
802
  "_type": "Value"
803
  },
804
  "lang": {
805
  "dtype": "string",
 
806
  "_type": "Value"
807
  },
808
  "question": {
809
  "stem": {
810
  "dtype": "string",
 
811
  "_type": "Value"
812
  },
813
  "choices": {
814
  "feature": {
815
  "label": {
816
  "dtype": "string",
 
817
  "_type": "Value"
818
  },
819
  "text": {
820
  "dtype": "string",
 
821
  "_type": "Value"
822
  }
823
  },
 
 
824
  "_type": "Sequence"
825
  }
826
  },
827
  "answerKey": {
828
  "dtype": "string",
 
829
  "_type": "Value"
830
  }
831
  },
 
 
 
832
  "builder_name": "xcsr",
833
+ "dataset_name": "xcsr",
834
  "config_name": "X-CSQA-vi",
835
  "version": {
836
  "version_str": "1.1.0",
 
842
  "splits": {
843
  "test": {
844
  "name": "test",
845
+ "num_bytes": 265210,
846
  "num_examples": 1074,
847
+ "dataset_name": null
848
  },
849
  "validation": {
850
  "name": "validation",
851
+ "num_bytes": 253502,
852
  "num_examples": 1000,
853
+ "dataset_name": null
 
 
 
 
 
 
854
  }
855
  },
856
+ "download_size": 244641,
857
+ "dataset_size": 518712,
858
+ "size_in_bytes": 763353
 
859
  },
860
  "X-CSQA-hi": {
861
  "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",