LLMs are surprisingly great at compressing images and audio, DeepMind researchers find - eviltoast
Despite being primarily trained on text, these models achieved remarkable compression rates on image and audio data.
    • enkers@sh.itjust.works
      15·
      1 year ago

      That was my first thought as well, but it doesn’t seem to be the case:

      In their study, the Google DeepMind researchers repurposed open-source LLMs to perform arithmetic coding, a type of lossless compression algorithm.