🦦 Otter, a multi-modal model based on OpenFlamingo (open-sourced version of DeepMind's Flamingo), trained on MIMIC-IT and showcasing improved instruction-following and in-context learning ability.
Hi. I see your evaluation results on Twitter. That is outstanding.
Can you share "OTTER-Image-MPT7B" and your prompt templates for different MME subtasks on the GitHub repo?
Thanks!
Hi. I see your evaluation results on Twitter. That is outstanding. Can you share "OTTER-Image-MPT7B" and your prompt templates for different MME subtasks on the GitHub repo? Thanks!