Comments

Hacker News

may I ask why you named it Arda? haha

by ardatasci

That's kinda neat. Do you do any decoding tricks to handle varying processing capabilities across device models and ipads? Like shortening ctx length or maybe looking into the LFM models that have MoEs with like 1b active?

by sango_ta

Right now I am just restricting it to devices with 8GB RAM. I plan to start testing older phones soon to see if I can get it running well on them. MoE with low active params is interesting. I wonder if there is smaller models based on specific pairs (En/Es, ect.)

by cjjuice

Join the discussion

Write your take first — we'll ask for email only when you're ready to publish.

  • Hacker News
  • may I ask why you named it Arda? haha
    by ardatasci
  • That's kinda neat. Do you do any decoding tricks to handle varying processing capabilities across device models and ipads? Like shortening ctx length or maybe looking into the LFM models that have MoEs with like 1b active?
    by sango_ta
  • Right now I am just restricting it to devices with 8GB RAM. I plan to start testing older phones soon to see if I can get it running well on them. MoE with low active params is interesting. I wonder if there is smaller models based on specific pairs (En/Es, ect.)
    by cjjuice

Related stories