NLU Datasets

5.3 million Chinese-German parallel sentence pairs stored in text format, covering multiple domains such as tourism, medical treatment, daily life, news, etc. The data desensitization and quality checking had been done. It can be used for machine translation, NLP research, and bilingual text analysis.

This dataset contains 84,516 English sentences annotated with intent classes, slot labels, and slot values. The intent field includes music, weather, date, schedule, and smart home equipment, etc. It is applied to intent recognition, intent classification, and slot filling tasks.

This dataset contains 1.08 million English-Russian sentence pairs, the corpus consists of aligned English and Russian text data covering diverse topics and general language usage scenarios. Sensitive content, including political, adult, and personally identifiable information (PII), has been filtered and removed. it can be a base corpus for machine translation, natural language processing, and multilingual AI model development.

This dataset contains 1.34 million English-Korean sentence pairs, providing high-quality bilingual text resources for machine translation, natural language processing, and multilingual AI model training. The corpus consists of aligned English and Korean text data covering diverse topics and general language usage scenarios. Sensitive content, including political, adult, and personally identifiable information (PII), has been filtered and removed. It can be a base corpus for text-based data analysis, used in machine translation and other fields.

This dataset contains 380,000 aligned Japanese-English sentence pairs. Sensitive content such as political, pornographic, and personal information has been excluded. It can be a base corpus for translation systems, cross-lingual information retrieval, and text-based language processing applications.

This dataset contains 47,811 English sentences annotated with intent classes, slot labels, and slot values. The intent field includes music, weather, date, schedule, smart home equipment, etc. it is applied to intent recognition, intent classification, and slot filling tasks.

This dataset contains 80 million Chinese-English parallel sentences, covering domains such as travel, medicine, daily conversation. It is stored in txt format, cleaned, desensitized, and quality-checked. It can be used as a fundamental dataset for machine translation, bilingual NLP tasks, and other text processing applications.

This dataset contains 100,000 Chinese-Uyghur parallel sentence pairs stored in TXT format. The data achieves over 80% in fluency and translation adequacy. The dataset has undergone cleaning, anonymization, and quality inspection. The corpus is suitable for machine translation and translation model training.

This dataset contains 380,000 Uyghur-Chinese parallel sentence pairs, with Uyghur text aligned with corresponding Chinese translations. The data has been cleaned, anonymized, and quality-checked. It can be used as base corpus for text data analysis in machine translation and related fields.