micro-kiki Router v4 β€” 34-Domain Multi-Label Classifier

MiniLM-L6 embedding + MLP head. Classifies a user prompt into up to 34 domains (sigmoid, not softmax β€” domains are not mutually exclusive). Used at inference time by the micro-kiki serving layer to pick the top-K LoRA adapters to load alongside the Qwen3.6-35B-A3B base.

Numbers (2026-04-22)

Metric Value
Train / val 21 510 / 5 377
top-1 accuracy 0.464
top-3 accuracy 0.696

Files

  • router.safetensors β€” MLP head weights
  • label_map.json β€” domain ↔ index map
  • meta.json β€” training config snapshot

Source

Training & serving pipeline at hypneum-lab/micro-kiki. Trained by scripts/train_router_v4.py.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for clemsail/micro-kiki-router-v4

Finetuned
(823)
this model