Update README.md

README.md

@@ -34,9 +34,9 @@ task_ids:
 
 ### Dataset Summary
 
-XNLI-ca is a dataset of textual entailment in Catalan, which contains 21,163 pairs of premises and hypotheses, annotated according to the inference relation they have (implication, contradiction or neutral).
+Professional translation into Catalan of The Cross-lingual Natural Language Inference ([XNLI dataset])(https://github.com/facebookresearch/XNLI).
 
-This dataset was developed by the Language Technologies Unit at BSC as part of [Projecte AINA](https://politiquesdigitals.gencat.cat/ca/economia/catalonia-ai/aina/), to enrich the [Catalan Language Understanding Benchmark (CLUB)](https://club.aina.bsc.es/). 
+XNLI is an evaluation corpus for language transfer and cross-lingual sentence classification in 15 languages. It is a crowd-sourced collection of 5,000 test and 2,500 dev pairs for the MultiNLI corpus. The pairs are annotated with textual entailment and translated into 14 languages: French, Spanish, German, Greek, Bulgarian, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi, Swahili and Urdu. This results in 112.5k annotated pairs. Each premise can be associated with the corresponding hypothesis in the 15 languages, summing up to more than 1.5M combinations.
 
 ### Supported Tasks and Leaderboards
 
@@ -50,7 +50,7 @@ The dataset is in Catalan (`ca-CA`).
 
 ### Data Instances
 
-Three JSON files, one for each split.
+Two JSON files, one for each split.
 
 ### Example:
 
@@ -85,47 +85,46 @@ Three JSON files, one for each split.
     
 ### Data Splits
 
-* dev.json: 2116 examples
-* test.json: 2117 examples
-* train.json: 16930 examples
+* dev.json: 2490 examples
+* test.json: 5010 examples
 
 ## Dataset Creation
 
 ### Curation Rationale
+
 We created this dataset to contribute to the development of language models in Catalan, a low-resource language.
 
 ### Source Data
 
-Source sentences are extracted from the [Catalan Textual Corpus](https://doi.org/10.5281/zenodo.4519349) and from [VilaWeb](https://www.vilaweb.cat) newswire.
+[XNLI](https://github.com/facebookresearch/XNLI).
 
 #### Initial Data Collection and Normalization
 
-12000 sentences from the BSC [Catalan Textual Corpus](https://doi.org/10.5281/zenodo.4519349), together with 6200 headers from the Catalan news site [VilaWeb](https://www.vilaweb.cat), were chosen randomly. We filtered them by different criteria, such as length and stand-alone intelligibility. For each selected text, we commissioned 3 hypotheses (one for each entailment category) to be written by a team of native annotators. 
+This dataset is a professional translation of XNLI into Catalan, commissioned by BSC LangTech Unit within Projecte AINA.
 
-Some sentence pairs were excluded because of inconsistencies. 
+For more information on how XNLI was created, refer to the paper [XNLI: Evaluating Cross-lingual Sentence Representations](https://arxiv.org/abs/1809.05053), or 
+visit the [XNLI's webpage](https://github.com/facebookresearch/XNLI).
 
 #### Who are the source language producers?
 
-The Catalan Textual Corpus corpus consists of several corpora gathered from web crawling and public corpora. More information can be found [here](https://doi.org/10.5281/zenodo.4519349).
 
-[VilaWeb](https://www.vilaweb.cat) is a Catalan newswire.
 
 ### Annotations
 
 #### Annotation process
 
-We commissioned 3 hypotheses (one for each entailment category) to be written by a team of annotators. 
+
 
 #### Who are the annotators?
 
-Annotators are a team of native language collaborators from two independent companies.
+[N/A]
 
 ### Personal and Sensitive Information
 
 No personal or sensitive information included.
   
 ## Considerations for Using the Data
-
+The Cross-lingual Natural Language Inference (XNLI) corpus is a crowd-sourced collection of 5,000 test and 2,500 dev pairs for the MultiNLI corpus. The pairs are annotated with textual entailment and translated into 14 languages: French, Spanish, German, Greek, Bulgarian, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi, Swahili and Urdu. This results in 112.5k annotated pairs. Each premise can be associated with the corresponding hypothesis in the 15 languages, summing up to more than 1.5M combinations.
 ### Social Impact of Dataset
 
 We hope this dataset contributes to the development of language models in Catalan, a low-resource language.
@@ -142,7 +141,7 @@ We hope this dataset contributes to the development of language models in Catala
 
 ### Dataset Curators
 
-Languge Technologies Unit at the Barcelona Supercomputing Center (langtech@bsc.es)
+Language Technologies Unit at the Barcelona Supercomputing Center (langtech@bsc.es)
 
 This work was funded by the [Departament de la Vicepresidència i de Polítiques Digitals i Territori de la Generalitat de Catalunya](https://politiquesdigitals.gencat.cat/ca/inici/index.html#googtrans(ca|en) within the framework of [Projecte AINA](https://politiquesdigitals.gencat.cat/ca/economia/catalonia-ai/aina).