Autoresearch for weather dycores. Contribute to khzhao/dynamaxx development by creating an account on GitHub.
We use a language model (LM) to aggregate the outputs of 2+ vision-language models (VLMs). Our model assemble approach is named Cola (COordinative LAnguage model or visual reasoning). Cola is most ...