<ul><li>H2O.ai has released two small language models, Mississippi 2B and Mississippi 0.8B, optimized for multimodal tasks.</li><li>Mississippi 2B is designed to analyze images based on natural language instructions and can generate detailed descriptions and explanations.</li><li>Mississippi 2B and Mississippi 0.8B can both be used for text extraction tasks, with the latter outperforming other models in optical character recognition.</li><li>H2O.ai envisions these models being deployed on devices with limited processing power and for latency-sensitive use cases.</li></ul>

H2O.ai releases small language models for multimodal processing tasks

Discover more