COGVLM-17B OPEN VISION LANGUAGE MODEL

Summary from the TLDR AI newsletter

With 10B vision parameters and 7B language parameters, this multimodal model excels at many standard benchmarks and performs well on human evaluation.