This node
probing WebGPU…
waiting for server…
jobs pulled 0 · last —
tokens out 0
p50 0 · p90 0 tok/s
p50 0 · p90 0 tok/s
Trust boundary
nodes return raw model output. category mapping and ground-truth
scoring run on the coordinator — a node can't grade its own work.
accuracy — on 0 scored
The corpus — one tile per report, colored by the node that read it
0nodes
0reports read
0reports/min
0tok/s aggregate
0tokens out
0p50 tok/s · node
0p90 tok/s · node
0:00elapsed
this swarm—
ghost: single node at current pace—
Findings assemble live — probable cause
Talk to this node's model — the same weights doing the batch work
Load a model, then ask it anything. Chat takes priority over batch jobs; this node resumes pulling the moment you're done.