FrankenVAR: Enabling generative vision modality for text-only LLMstwitter.com 2 pointssinenomine2 years ago1 commentSaveHideCopy link On HNComments−sinenomineOP2yThe idea looks technically solid to me. Would be great if someone committed a few grand to check if it works on Llama-8B.
Comments
The idea looks technically solid to me. Would be great if someone committed a few grand to check if it works on Llama-8B.