Ok so this talks about "large-model splits," which sounds li...

原始 JSON

{
  "kind": 1,
  "id": "3aab228c80569358d123125b87c3d27e3033316f982fcd30719bf947ff625c5b",
  "pubkey": "d1a61498f4b1dc26df0becd1a3e9e88f355fb614b771e8b5b277d0ff99a82a23",
  "created_at": 1788431586,
  "tags": [
    [
      "e",
      "a09574a20645dc244207157883cd0b4af4661937995cbc42927353d09330484b",
      "wss://nos.lol/",
      "root",
      "199f12c94fe6a2903908c46efcb91190d9993b7380886c98321b59c513dca973"
    ],
    [
      "e",
      "d6e422c6ceb8734b9bbff2f13ec5233a451837ed791d424c75229ef800d01f14",
      "wss://relay.damus.io/",
      "reply",
      "199f12c94fe6a2903908c46efcb91190d9993b7380886c98321b59c513dca973"
    ],
    [
      "p",
      "199f12c94fe6a2903908c46efcb91190d9993b7380886c98321b59c513dca973"
    ]
  ],
  "content": "Ok so this talks about \"large-model splits,\" which sounds like some kind of parallel execution across multiple machines none of which could handle the model by itself? That would be great if it works in practice. 🤨",
  "sig": "83f2991400f0e53f5db3d33dc2da00bfef81b121e2aa23f5f96f3233a16781e6730f5b9120a04e7dbce1de538900a0b511f247955e9d6d12834efe3989415e2e"
}