Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported
Python
"""Prerouter: trained cross-token routing predictor heads + step-boundary
staging.
Public API::
from edge0.prerouter import (PrerouterSpec, PrerouterState,
PrerouterHead, install_prerouter,
CrossTokenStager, LingPrerouterStager)
"""
from edge0.prerouter.heads import PrerouterHead, select_from_logits, topk_onehot
from edge0.prerouter.install import install_prerouter
from edge0.prerouter.spec import PrerouterSpec
from edge0.prerouter.stager import (
CrossTokenStager,
LingPrerouterStager,
PrerouterStager,
)
from edge0.prerouter.state import PrerouterState
__all__ = [
"PrerouterSpec",
"PrerouterState",
"PrerouterHead",
"PrerouterStager",
"CrossTokenStager",
"LingPrerouterStager",
"install_prerouter",
"select_from_logits",
"topk_onehot",
]