{"Rexhaif--xnli_ru": {"description": "XNLI is a subset of a few thousand examples from MNLI which has been translated\ninto a 14 different languages (some low-ish resource). As with MNLI, the goal is\nto predict textual entailment (does sentence A imply/contradict/neither sentence\nB) and is a classification task (given two sentences, predict one of three\nlabels).\n", "citation": "@InProceedings{conneau2018xnli,\n  author = {Conneau, Alexis\n                 and Rinott, Ruty\n                 and Lample, Guillaume\n                 and Williams, Adina\n                 and Bowman, Samuel R.\n                 and Schwenk, Holger\n                 and Stoyanov, Veselin},\n  title = {XNLI: Evaluating Cross-lingual Sentence Representations},\n  booktitle = {Proceedings of the 2018 Conference on Empirical Methods\n               in Natural Language Processing},\n  year = {2018},\n  publisher = {Association for Computational Linguistics},\n  location = {Brussels, Belgium},\n}", "homepage": "https://www.nyu.edu/projects/bowman/xnli/", "license": "", "features": {"label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "id": null, "_type": "ClassLabel"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "xnli", "config_name": "ru", "version": {"version_str": "1.1.0", "description": "", "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 131037721, "num_examples": 392702, "dataset_name": "xnli_ru"}, "test": {"name": "test", "num_bytes": 1618496, "num_examples": 5010, "dataset_name": "xnli_ru"}, "validation": {"name": "validation", "num_bytes": 793912, "num_examples": 2490, "dataset_name": "xnli_ru"}}, "download_checksums": null, "download_size": 64461175, "post_processing_size": null, "dataset_size": 133450129, "size_in_bytes": 197911304}}