Datasets:

Multilinguality:
multilingual
Size Categories:
1K<n<10K
Annotations Creators:
crowdsourced
ArXiv:
Tags:
License:
albertvillanova HF staff commited on
Commit
ab11c86
1 Parent(s): f958747

Add X-CSQA-it data files

Browse files
README.md CHANGED
@@ -682,13 +682,13 @@ dataset_info:
682
  dtype: string
683
  splits:
684
  - name: test
685
- num_bytes: 232906
686
  num_examples: 1074
687
  - name: validation
688
- num_bytes: 221184
689
  num_examples: 1000
690
- download_size: 7519903
691
- dataset_size: 454090
692
  - config_name: X-CSQA-jap
693
  features:
694
  - name: id
@@ -957,6 +957,12 @@ configs:
957
  path: X-CSQA-fr/test-*
958
  - split: validation
959
  path: X-CSQA-fr/validation-*
 
 
 
 
 
 
960
  - config_name: X-CSQA-zh
961
  data_files:
962
  - split: test
 
682
  dtype: string
683
  splits:
684
  - name: test
685
+ num_bytes: 232604
686
  num_examples: 1074
687
  - name: validation
688
+ num_bytes: 220902
689
  num_examples: 1000
690
+ download_size: 236130
691
+ dataset_size: 453506
692
  - config_name: X-CSQA-jap
693
  features:
694
  - name: id
 
957
  path: X-CSQA-fr/test-*
958
  - split: validation
959
  path: X-CSQA-fr/validation-*
960
+ - config_name: X-CSQA-it
961
+ data_files:
962
+ - split: test
963
+ path: X-CSQA-it/test-*
964
+ - split: validation
965
+ path: X-CSQA-it/validation-*
966
  - config_name: X-CSQA-zh
967
  data_files:
968
  - split: test
X-CSQA-it/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35ed64ea041e138d3dfb003f219362f55a0f17e91dca14064570196ccb002ab3
3
+ size 122483
X-CSQA-it/validation-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ab4542e5534872ec8ac476fe31632813c163a6c624d341f2f4b91a574884f18
3
+ size 113647
dataset_infos.json CHANGED
@@ -337,48 +337,38 @@
337
  "features": {
338
  "id": {
339
  "dtype": "string",
340
- "id": null,
341
  "_type": "Value"
342
  },
343
  "lang": {
344
  "dtype": "string",
345
- "id": null,
346
  "_type": "Value"
347
  },
348
  "question": {
349
  "stem": {
350
  "dtype": "string",
351
- "id": null,
352
  "_type": "Value"
353
  },
354
  "choices": {
355
  "feature": {
356
  "label": {
357
  "dtype": "string",
358
- "id": null,
359
  "_type": "Value"
360
  },
361
  "text": {
362
  "dtype": "string",
363
- "id": null,
364
  "_type": "Value"
365
  }
366
  },
367
- "length": -1,
368
- "id": null,
369
  "_type": "Sequence"
370
  }
371
  },
372
  "answerKey": {
373
  "dtype": "string",
374
- "id": null,
375
  "_type": "Value"
376
  }
377
  },
378
- "post_processed": null,
379
- "supervised_keys": null,
380
- "task_templates": null,
381
  "builder_name": "xcsr",
 
382
  "config_name": "X-CSQA-it",
383
  "version": {
384
  "version_str": "1.1.0",
@@ -390,27 +380,20 @@
390
  "splits": {
391
  "test": {
392
  "name": "test",
393
- "num_bytes": 232906,
394
  "num_examples": 1074,
395
- "dataset_name": "xcsr"
396
  },
397
  "validation": {
398
  "name": "validation",
399
- "num_bytes": 221184,
400
  "num_examples": 1000,
401
- "dataset_name": "xcsr"
402
- }
403
- },
404
- "download_checksums": {
405
- "https://inklab.usc.edu/XCSR/xcsr_datasets.zip": {
406
- "num_bytes": 7519903,
407
- "checksum": "c45b29ece740643252d5402e76be1e33f96f9d6910053f79e80d39887f10c85e"
408
  }
409
  },
410
- "download_size": 7519903,
411
- "post_processing_size": null,
412
- "dataset_size": 454090,
413
- "size_in_bytes": 7973993
414
  },
415
  "X-CSQA-jap": {
416
  "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",
 
337
  "features": {
338
  "id": {
339
  "dtype": "string",
 
340
  "_type": "Value"
341
  },
342
  "lang": {
343
  "dtype": "string",
 
344
  "_type": "Value"
345
  },
346
  "question": {
347
  "stem": {
348
  "dtype": "string",
 
349
  "_type": "Value"
350
  },
351
  "choices": {
352
  "feature": {
353
  "label": {
354
  "dtype": "string",
 
355
  "_type": "Value"
356
  },
357
  "text": {
358
  "dtype": "string",
 
359
  "_type": "Value"
360
  }
361
  },
 
 
362
  "_type": "Sequence"
363
  }
364
  },
365
  "answerKey": {
366
  "dtype": "string",
 
367
  "_type": "Value"
368
  }
369
  },
 
 
 
370
  "builder_name": "xcsr",
371
+ "dataset_name": "xcsr",
372
  "config_name": "X-CSQA-it",
373
  "version": {
374
  "version_str": "1.1.0",
 
380
  "splits": {
381
  "test": {
382
  "name": "test",
383
+ "num_bytes": 232604,
384
  "num_examples": 1074,
385
+ "dataset_name": null
386
  },
387
  "validation": {
388
  "name": "validation",
389
+ "num_bytes": 220902,
390
  "num_examples": 1000,
391
+ "dataset_name": null
 
 
 
 
 
 
392
  }
393
  },
394
+ "download_size": 236130,
395
+ "dataset_size": 453506,
396
+ "size_in_bytes": 689636
 
397
  },
398
  "X-CSQA-jap": {
399
  "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",