Skip to content

decision:eos

2 TagsUpdated 0.75B params16384 contextEnglish, ChineseApache-2.0by vLLM Semantic Router

Decision 1.0 Eos, fully fine-tuned Qwen3.5-0.8B: 17.49 on Decision Index 0.2.

0.75b
ollaya run decision:eos --preset triage "I was charged twice for my subscription this month and want a refund."

Details

  • graph7917445d40c7 · 8 MBonnx · decision · 0.75B · fp32
  • weights613f491da279 · 1.5 GBhuggingface.co/llm-semantic-router/Decision-1.0-Eos-0.8B/resolve/3c2d632…/backbone/model.safetensors
  • weights75f4ee8b512a · 4 MBhuggingface.co/llm-semantic-router/Decision-1.0-Eos-0.8B/resolve/3c2d632…/decision_head.safetensors
  • tokenizer06b9509352d2 · 20 MBhuggingface.co/llm-semantic-router/Decision-1.0-Eos-0.8B/resolve/3c2d632…/tokenizer.json
  • decision2a6f1ef91f03 · 3 KB{"engine": "onnx", "family": "decision", "encoder": "", "layout": "decision-endpoint-v1", …}
  • calibration9cb229b1316a · 217 B{"temperature": [1.039, 1.039, 1.039]}
  • license51ba79094bb8 · 10 KBDecision 1.0 by the vLLM Semantic Router contributors (https://huggingface.co/llm-semantic-router)

Every layer is checked against its sha256 when it is pulled. Weights and tokenizers download from the model author's Hugging Face repository at a pinned commit; Ollaya never re-hosts them.

Decision 1.0 readme and all models