Language Tasks

• Sentence Classification
  • Zero Shot Classification: the model is not fine-tuned on the given data and label. The model only see the given label once.
• Semantic Similarity:
  • Semantic Search: Different from traditional index based search engine, Semantic Search finding related content by calculating semantic similarities between contents.
• Token Classification: Different from sentence classification, this is classification upon each individual word(token) in side a sentence. Such as grammatical components, or named entities.
• Text Generation: Generating sentence(s) using the few words(tokens) given as the beginning.
• Question Answering: Not GPT-like question answering, but extracting answer from the given input.
• Translation
• Summarization