the quality of GLM 5.2 in just 155 or 162 GB?

  • Multiplexer@discuss.tchncs.de
    link
    fedilink
    English
    arrow-up
    3
    ·
    9 days ago

    Anyone knows, why 4bit quant is only marginally smaller than 8bit quant, though?
    And shouldn’t 8bit be roughly equal to the number of parameters anyway, so ~300GB?

    • BeefAndPoultry@lemmus.orgOP
      link
      fedilink
      English
      arrow-up
      4
      ·
      9 days ago

      the original model has a lot of parts that were natively trained in 4 bit, so those layers can’t go higher