wissamantoun/orca_parquet

--- pretty_name: ORCA – Arabic Language Understanding Benchmark language: - ar configs: - config_name: abusive data_files: - split: train path: data/abusive/train.parquet - split: validation path: data/abusive/validation.parquet - split: test path: data/abusive/test.parquet - config_name: adult data_files: - split: train path: data/adult/train.parquet - split: validation path: data/adult/validation.parquet - split: test path: data/adult/test.parquet - config_name: age data_files: - split: train path: data/age/train.parquet - split: validation path: data/age/validation.parquet - split: test path: data/age/test.parquet - config_name: ans-claim data_files: - split: train path: data/ans-claim/train.parquet - split: validation path: data/ans-claim/validation.parquet - split: test path: data/ans-claim/test.parquet - config_name: ans-stance data_files: - split: train path: data/ans-stance/train.parquet - split: validation path: data/ans-stance/validation.parquet - split: test path: data/ans-stance/test.parquet - config_name: aqmar-ner data_files: - split: train path: data/aqmar-ner/train.parquet - split: validation path: data/aqmar-ner/validation.parquet - split: test path: data/aqmar-ner/test.parquet - config_name: arabic-ner data_files: - split: train path: data/arabic-ner/train.parquet - split: validation path: data/arabic-ner/validation.parquet - split: test path: data/arabic-ner/test.parquet - config_name: baly-stance data_files: - split: train path: data/baly-stance/train.parquet - split: validation path: data/baly-stance/validation.parquet - split: test path: data/baly-stance/test.parquet - config_name: dangerous data_files: - split: train path: data/dangerous/train.parquet - split: validation path: data/dangerous/validation.parquet - split: test path: data/dangerous/test.parquet - config_name: dialect-binary data_files: - split: train path: data/dialect-binary/train.parquet - split: validation path: data/dialect-binary/validation.parquet - split: test path: data/dialect-binary/test.parquet - config_name: dialect-country data_files: - split: train path: data/dialect-country/train.parquet - split: validation path: data/dialect-country/validation.parquet - split: test path: data/dialect-country/test.parquet - config_name: dialect-pos data_files: - split: train path: data/dialect-pos/train.parquet - split: validation path: data/dialect-pos/validation.parquet - split: test path: data/dialect-pos/test.parquet - config_name: dialect-region data_files: - split: train path: data/dialect-region/train.parquet - split: validation path: data/dialect-region/validation.parquet - split: test path: data/dialect-region/test.parquet - config_name: emotion data_files: - split: train path: data/emotion/train.parquet - split: validation path: data/emotion/validation.parquet - split: test path: data/emotion/test.parquet - config_name: emotion-reg data_files: - split: train path: data/emotion-reg/train.parquet - split: validation path: data/emotion-reg/validation.parquet - split: test path: data/emotion-reg/test.parquet - config_name: gender data_files: - split: train path: data/gender/train.parquet - split: validation path: data/gender/validation.parquet - split: test path: data/gender/test.parquet - config_name: hate-speech data_files: - split: train path: data/hate-speech/train.parquet - split: validation path: data/hate-speech/validation.parquet - split: test path: data/hate-speech/test.parquet - config_name: irony data_files: - split: train path: data/irony/train.parquet - split: validation path: data/irony/validation.parquet - split: test path: data/irony/test.parquet - config_name: machine-generation data_files: - split: train path: data/machine-generation/train.parquet - split: validation path: data/machine-generation/validation.parquet - split: test path: data/machine-generation/test.parquet - config_name: mq2q data_files: - split: train path: data/mq2q/train.parquet - split: validation path: data/mq2q/validation.parquet - split: test path: data/mq2q/test.parquet - config_name: msa-pos data_files: - split: train path: data/msa-pos/train.parquet - split: validation path: data/msa-pos/validation.parquet - split: test path: data/msa-pos/test.parquet - config_name: offensive data_files: - split: train path: data/offensive/train.parquet - split: validation path: data/offensive/validation.parquet - split: test path: data/offensive/test.parquet - config_name: qa data_files: - split: train path: data/qa/train.parquet - split: validation path: data/qa/validation.parquet - split: test path: data/qa/test.parquet - config_name: sarcasm data_files: - split: train path: data/sarcasm/train.parquet - split: validation path: data/sarcasm/validation.parquet - split: test path: data/sarcasm/test.parquet - config_name: sentiment data_files: - split: train path: data/sentiment/train.parquet - split: validation path: data/sentiment/validation.parquet - split: test path: data/sentiment/test.parquet - config_name: sts data_files: - split: train path: data/sts/train.parquet - split: validation path: data/sts/validation.parquet - split: test path: data/sts/test.parquet - config_name: topic data_files: - split: train path: data/topic/train.parquet - split: validation path: data/topic/validation.parquet - split: test path: data/topic/test.parquet - config_name: wsd data_files: - split: train path: data/wsd/train.parquet - split: validation path: data/wsd/validation.parquet - split: test path: data/wsd/test.parquet - config_name: xlni data_files: - split: train path: data/xlni/train.parquet - split: validation path: data/xlni/validation.parquet - split: test path: data/xlni/test.parquet --- # ORCA – Arabic Language Understanding Benchmark Converted from the original [`wissamantoun/orca_hf`](https://huggingface.co/datasets/wissamantoun/orca_hf) dataset script to parquet files. ## Usage ```python from datasets import load_dataset # Load a specific config ds = load_dataset("PATH/orca_parquet", name="sentiment") # Load all configs ds = load_dataset("PATH/orca_parquet") ```展示名称：ORCA – 阿拉伯语言理解基准测试集（ORCA – Arabic Language Understanding Benchmark） 语言： - 阿拉伯语（ar） 配置项： - 配置名称：辱骂性文本分类（abusive） 数据文件： - 划分集：训练集（train），路径：data/abusive/train.parquet - 划分集：验证集（validation），路径：data/abusive/validation.parquet - 划分集：测试集（test），路径：data/abusive/test.parquet - 配置名称：成人内容检测（adult） 数据文件： - 划分集：训练集，路径：data/adult/train.parquet - 划分集：验证集，路径：data/adult/validation.parquet - 划分集：测试集，路径：data/adult/test.parquet - 配置名称：年龄属性理解（age） 数据文件： - 划分集：训练集，路径：data/age/train.parquet - 划分集：验证集，路径：data/age/validation.parquet - 划分集：测试集，路径：data/age/test.parquet - 配置名称：回答-主张分类（ans-claim） 数据文件： - 划分集：训练集，路径：data/ans-claim/train.parquet - 划分集：验证集，路径：data/ans-claim/validation.parquet - 划分集：测试集，路径：data/ans-claim/test.parquet - 配置名称：回答-立场分类（ans-stance） 数据文件： - 划分集：训练集，路径：data/ans-stance/train.parquet - 划分集：验证集，路径：data/ans-stance/validation.parquet - 划分集：测试集，路径：data/ans-stance/test.parquet - 配置名称：阿克马尔命名实体识别（aqmar-ner） 数据文件： - 划分集：训练集，路径：data/aqmar-ner/train.parquet - 划分集：验证集，路径：data/aqmar-ner/validation.parquet - 划分集：测试集，路径：data/aqmar-ner/test.parquet - 配置名称：阿拉伯语命名实体识别（arabic-ner） 数据文件： - 划分集：训练集，路径：data/arabic-ner/train.parquet - 划分集：验证集，路径：data/arabic-ner/validation.parquet - 划分集：测试集，路径：data/arabic-ner/test.parquet - 配置名称：巴利立场分类（baly-stance） 数据文件： - 划分集：训练集，路径：data/baly-stance/train.parquet - 划分集：验证集，路径：data/baly-stance/validation.parquet - 划分集：测试集，路径：data/baly-stance/test.parquet - 配置名称：危险内容检测（dangerous） 数据文件： - 划分集：训练集，路径：data/dangerous/train.parquet - 划分集：验证集，路径：data/dangerous/validation.parquet - 划分集：测试集，路径：data/dangerous/test.parquet - 配置名称：二元方言分类（dialect-binary） 数据文件： - 划分集：训练集，路径：data/dialect-binary/train.parquet - 划分集：验证集，路径：data/dialect-binary/validation.parquet - 划分集：测试集，路径：data/dialect-binary/test.parquet - 配置名称：方言国别分类（dialect-country） 数据文件： - 划分集：训练集，路径：data/dialect-country/train.parquet - 划分集：验证集，路径：data/dialect-country/validation.parquet - 划分集：测试集，路径：data/dialect-country/test.parquet - 配置名称：方言词性标注（dialect-pos） 数据文件： - 划分集：训练集，路径：data/dialect-pos/train.parquet - 划分集：验证集，路径：data/dialect-pos/validation.parquet - 划分集：测试集，路径：data/dialect-pos/test.parquet - 配置名称：方言地域分类（dialect-region） 数据文件： - 划分集：训练集，路径：data/dialect-region/train.parquet - 划分集：验证集，路径：data/dialect-region/validation.parquet - 划分集：测试集，路径：data/dialect-region/test.parquet - 配置名称：情感分类（emotion） 数据文件： - 划分集：训练集，路径：data/emotion/train.parquet - 划分集：验证集，路径：data/emotion/validation.parquet - 划分集：测试集，路径：data/emotion/test.parquet - 配置名称：情感回归（emotion-reg） 数据文件： - 划分集：训练集，路径：data/emotion-reg/train.parquet - 划分集：验证集，路径：data/emotion-reg/validation.parquet - 划分集：测试集，路径：data/emotion-reg/test.parquet - 配置名称：性别属性理解（gender） 数据文件： - 划分集：训练集，路径：data/gender/train.parquet - 划分集：验证集，路径：data/gender/validation.parquet - 划分集：测试集，路径：data/gender/test.parquet - 配置名称：仇恨言论检测（hate-speech） 数据文件： - 划分集：训练集，路径：data/hate-speech/train.parquet - 划分集：验证集，路径：data/hate-speech/validation.parquet - 划分集：测试集，路径：data/hate-speech/test.parquet - 配置名称：反语识别（irony） 数据文件： - 划分集：训练集，路径：data/irony/train.parquet - 划分集：验证集，路径：data/irony/validation.parquet - 划分集：测试集，路径：data/irony/test.parquet - 配置名称：机器生成文本检测（machine-generation） 数据文件： - 划分集：训练集，路径：data/machine-generation/train.parquet - 划分集：验证集，路径：data/machine-generation/validation.parquet - 划分集：测试集，路径：data/machine-generation/test.parquet - 配置名称：问题对问题匹配（mq2q） 数据文件： - 划分集：训练集，路径：data/mq2q/train.parquet - 划分集：验证集，路径：data/mq2q/validation.parquet - 划分集：测试集，路径：data/mq2q/test.parquet - 配置名称：现代标准阿拉伯语词性标注（msa-pos） 数据文件： - 划分集：训练集，路径：data/msa-pos/train.parquet - 划分集：验证集，路径：data/msa-pos/validation.parquet - 划分集：测试集，路径：data/msa-pos/test.parquet - 配置名称：冒犯性言论检测（offensive） 数据文件： - 划分集：训练集，路径：data/offensive/train.parquet - 划分集：验证集，路径：data/offensive/validation.parquet - 划分集：测试集，路径：data/offensive/test.parquet - 配置名称：问答任务（qa） 数据文件： - 划分集：训练集，路径：data/qa/train.parquet - 划分集：验证集，路径：data/qa/validation.parquet - 划分集：测试集，路径：data/qa/test.parquet - 配置名称：讽刺识别（sarcasm） 数据文件： - 划分集：训练集，路径：data/sarcasm/train.parquet - 划分集：验证集，路径：data/sarcasm/validation.parquet - 划分集：测试集，路径：data/sarcasm/test.parquet - 配置名称：情感分析（sentiment） 数据文件： - 划分集：训练集，路径：data/sentiment/train.parquet - 划分集：验证集，路径：data/sentiment/validation.parquet - 划分集：测试集，路径：data/sentiment/test.parquet - 配置名称：语义文本相似度（sts） 数据文件： - 划分集：训练集，路径：data/sts/train.parquet - 划分集：验证集，路径：data/sts/validation.parquet - 划分集：测试集，路径：data/sts/test.parquet - 配置名称：主题分类（topic） 数据文件： - 划分集：训练集，路径：data/topic/train.parquet - 划分集：验证集，路径：data/topic/validation.parquet - 划分集：测试集，路径：data/topic/test.parquet - 配置名称：词义消歧（wsd） 数据文件： - 划分集：训练集，路径：data/wsd/train.parquet - 划分集：验证集，路径：data/wsd/validation.parquet - 划分集：测试集，路径：data/wsd/test.parquet - 配置名称：跨语言自然语言推理（xlni） 数据文件： - 划分集：训练集，路径：data/xlni/train.parquet - 划分集：验证集，路径：data/xlni/validation.parquet - 划分集：测试集，路径：data/xlni/test.parquet # ORCA – 阿拉伯语言理解基准测试集 本数据集由原始 [`wissamantoun/orca_hf`](https://huggingface.co/datasets/wissamantoun/orca_hf) 数据集脚本转换为Parquet格式文件而来。 ## 使用方法 python from datasets import load_dataset # 加载指定配置的数据集 ds = load_dataset("PATH/orca_parquet", name="sentiment") # 加载全部配置的数据集 ds = load_dataset("PATH/orca_parquet")