‹ BackHN Continuity

Thread

MiMo v2.6

1130 points · 483 comments · volf_

  1. stymaar · · focus · HN ↗
    Flash[1]: 309B total / 15B activated parameters

    Pro [2]:, 1.02T total / 42B activated parameters

    [1]: <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;XiaomiMiMo&#x2F;MiMo-V2.6-Flash-RL" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;XiaomiMiMo&#x2F;MiMo-V2.6-Flash-RL

    [2]: <a href="https:&#x2F;&#x2F;huggingface.co&#x2F;XiaomiMiMo&#x2F;MiMo-V2.6-Pro-RL" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;XiaomiMiMo&#x2F;MiMo-V2.6-Pro-RL

    1. verdverm · · focus · HN ↗
      curious why the HF pill (on the right) always has inaccurate values
      1. stymaar · · focus · HN ↗
        I noticed the same, and I wonder as well.
        1. verdverm · · focus · HN ↗
          I suspect they are calculating something in the weights or config, I see it pretty consistently with quants
      2. bopbop9876 · · focus · HN ↗
        I believe it&#x27;s because this model is natively fp8 (for the most part), and that display struggles native quants.
      3. wren6991 · · focus · HN ↗
        Packed 4-bit weights are often identified as u8 byte arrays in the safetensors metadata, so the HF UI counts them as half a weight each.
        1. verdverm · · focus · HN ↗
          I would think there is sufficient information in the various config files to work this out, based what I&#x27;ve seen in my own quant artifacts.
          1. wren6991 · · focus · HN ↗
            Yeah, I agree it&#x27;s probably fixable, but I think a naive interpretation of the metadata is probably the source of this bug (which HF has had for as long as I can remember)
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.