H2O AI releases Danube, a super-tiny LLM for mobile applications

In the Hellaswag LLM benchmark evaluating common sense natural language inference, Danube performed with an accuracy of 69.58%, sitting just behind Stability AI’s Stable LM 2 1.6 billion parameter model.

accuracy ai applications benchmark billion common sense computer science conversational ai danube generative-ai h2o h2o ai h2o-danube-1.8b inference language llm llm benchmark microsoft phi ml and deep learning mobile mobile applications natural natural language nlp releases science sense stability stability ai stable lm