Skip to content

d3:lite

6 TagsUpdated 0.8B params8192 contextEnglishApache-2.0by vLLM Semantic Router

Qwen3.5-0.8B: 0.526 on typed decisions, zero-shot.

0.8b
ollaya run d3:lite --preset triage "I was charged twice for my subscription this month and want a refund."

Details

  • graph77e49ecae62b · 9 MBonnx · d3 · 0.8B · fp32
  • weightsca6222dbf2b7 · 1.7 GBhuggingface.co/vllm-sr/d3-lite/resolve/03560c2…/model-00001-of-00001.safetensors
  • weightsfe11cfac5c79 · 1 MBhuggingface.co/vllm-sr/d3-lite/resolve/03560c2…/readout.safetensors
  • tokenizer06b9509352d2 · 20 MBhuggingface.co/bespokelabs/Bespoke-Nimble-9B-v2/resolve/4b8c04d…/tokenizer.json
  • decision930eba275f57 · 5 KB{"engine": "onnx", "family": "d3", "encoder": "", "layout": "d3-code-readout-v1", …}
  • calibration0f89e51fde90 · 133 B{"temperature": [1.0, 1.0, 1.0]}
  • license3d7ec11263ac · 10 KBDecision 3.0 (d3) by the vLLM Semantic Router contributors (https://huggingface.co/vllm-sr)

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 3.0 (d3) readme and all models