Datasets:

Multilinguality:
multilingual
Size Categories:
1K<n<10K
Annotations Creators:
crowdsourced
ArXiv:
Tags:
License:
albertvillanova HF staff commited on
Commit
8a0429e
1 Parent(s): 55795e8

Add X-CODAH-zh data files

Browse files
README.md CHANGED
@@ -493,13 +493,13 @@ dataset_info:
493
  dtype: string
494
  splits:
495
  - name: test
496
- num_bytes: 394946
497
  num_examples: 1000
498
  - name: validation
499
- num_bytes: 115137
500
  num_examples: 300
501
- download_size: 7519903
502
- dataset_size: 510083
503
  - config_name: X-CSQA-ar
504
  features:
505
  - name: id
@@ -939,6 +939,12 @@ configs:
939
  path: X-CODAH-en/test-*
940
  - split: validation
941
  path: X-CODAH-en/validation-*
 
 
 
 
 
 
942
  - config_name: X-CSQA-ar
943
  data_files:
944
  - split: test
 
493
  dtype: string
494
  splits:
495
  - name: test
496
+ num_bytes: 394660
497
  num_examples: 1000
498
  - name: validation
499
+ num_bytes: 115025
500
  num_examples: 300
501
+ download_size: 237827
502
+ dataset_size: 509685
503
  - config_name: X-CSQA-ar
504
  features:
505
  - name: id
 
939
  path: X-CODAH-en/test-*
940
  - split: validation
941
  path: X-CODAH-en/validation-*
942
+ - config_name: X-CODAH-zh
943
+ data_files:
944
+ - split: test
945
+ path: X-CODAH-zh/test-*
946
+ - split: validation
947
+ path: X-CODAH-zh/validation-*
948
  - config_name: X-CSQA-ar
949
  data_files:
950
  - split: test
X-CODAH-zh/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2bcf4fdc3f7eb9726e9227a4fa66c77f97decebc9c21c2529cc18f166db06b6
3
+ size 181650
X-CODAH-zh/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6570cc4f954406e0ccf7ee22d7015a8f838fb440eb25d16dc2171317f3aa0f3d
3
+ size 56177
dataset_infos.json CHANGED
@@ -1133,53 +1133,42 @@
1133
  "features": {
1134
  "id": {
1135
  "dtype": "string",
1136
- "id": null,
1137
  "_type": "Value"
1138
  },
1139
  "lang": {
1140
  "dtype": "string",
1141
- "id": null,
1142
  "_type": "Value"
1143
  },
1144
  "question_tag": {
1145
  "dtype": "string",
1146
- "id": null,
1147
  "_type": "Value"
1148
  },
1149
  "question": {
1150
  "stem": {
1151
  "dtype": "string",
1152
- "id": null,
1153
  "_type": "Value"
1154
  },
1155
  "choices": {
1156
  "feature": {
1157
  "label": {
1158
  "dtype": "string",
1159
- "id": null,
1160
  "_type": "Value"
1161
  },
1162
  "text": {
1163
  "dtype": "string",
1164
- "id": null,
1165
  "_type": "Value"
1166
  }
1167
  },
1168
- "length": -1,
1169
- "id": null,
1170
  "_type": "Sequence"
1171
  }
1172
  },
1173
  "answerKey": {
1174
  "dtype": "string",
1175
- "id": null,
1176
  "_type": "Value"
1177
  }
1178
  },
1179
- "post_processed": null,
1180
- "supervised_keys": null,
1181
- "task_templates": null,
1182
  "builder_name": "xcsr",
 
1183
  "config_name": "X-CODAH-zh",
1184
  "version": {
1185
  "version_str": "1.1.0",
@@ -1191,27 +1180,20 @@
1191
  "splits": {
1192
  "test": {
1193
  "name": "test",
1194
- "num_bytes": 394946,
1195
  "num_examples": 1000,
1196
- "dataset_name": "xcsr"
1197
  },
1198
  "validation": {
1199
  "name": "validation",
1200
- "num_bytes": 115137,
1201
  "num_examples": 300,
1202
- "dataset_name": "xcsr"
1203
- }
1204
- },
1205
- "download_checksums": {
1206
- "https://inklab.usc.edu/XCSR/xcsr_datasets.zip": {
1207
- "num_bytes": 7519903,
1208
- "checksum": "c45b29ece740643252d5402e76be1e33f96f9d6910053f79e80d39887f10c85e"
1209
  }
1210
  },
1211
- "download_size": 7519903,
1212
- "post_processing_size": null,
1213
- "dataset_size": 510083,
1214
- "size_in_bytes": 8029986
1215
  },
1216
  "X-CODAH-de": {
1217
  "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",
 
1133
  "features": {
1134
  "id": {
1135
  "dtype": "string",
 
1136
  "_type": "Value"
1137
  },
1138
  "lang": {
1139
  "dtype": "string",
 
1140
  "_type": "Value"
1141
  },
1142
  "question_tag": {
1143
  "dtype": "string",
 
1144
  "_type": "Value"
1145
  },
1146
  "question": {
1147
  "stem": {
1148
  "dtype": "string",
 
1149
  "_type": "Value"
1150
  },
1151
  "choices": {
1152
  "feature": {
1153
  "label": {
1154
  "dtype": "string",
 
1155
  "_type": "Value"
1156
  },
1157
  "text": {
1158
  "dtype": "string",
 
1159
  "_type": "Value"
1160
  }
1161
  },
 
 
1162
  "_type": "Sequence"
1163
  }
1164
  },
1165
  "answerKey": {
1166
  "dtype": "string",
 
1167
  "_type": "Value"
1168
  }
1169
  },
 
 
 
1170
  "builder_name": "xcsr",
1171
+ "dataset_name": "xcsr",
1172
  "config_name": "X-CODAH-zh",
1173
  "version": {
1174
  "version_str": "1.1.0",
 
1180
  "splits": {
1181
  "test": {
1182
  "name": "test",
1183
+ "num_bytes": 394660,
1184
  "num_examples": 1000,
1185
+ "dataset_name": null
1186
  },
1187
  "validation": {
1188
  "name": "validation",
1189
+ "num_bytes": 115025,
1190
  "num_examples": 300,
1191
+ "dataset_name": null
 
 
 
 
 
 
1192
  }
1193
  },
1194
+ "download_size": 237827,
1195
+ "dataset_size": 509685,
1196
+ "size_in_bytes": 747512
 
1197
  },
1198
  "X-CODAH-de": {
1199
  "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",