vllm-sr/Vela-2.0-0.3B

zero-shot-classificationUpstream checksumsonnxsafetensorsvela2-unifieddecision-modelsystem-onemodernbert

License: apache-2.0
Size: not mirrored yet
Downloads in the last 30 days: 0 / 30 days

No official torrent yet — requesting it tells the flock what to host next. You cannot seed a catalog-only page.

Cataloged from vllm-sr/Vela-2.0-0.3B

All models by vllm-sr →

Vela 2.0: Towards Open Foundation Routing Models. vLLM Semantic Router x KR Labs

Docs | Blog | GitHub | Vela 2.0 collection

Vela 2.0 0.3B

Open Foundation Routing Models

Routing decisions. Safety checks. Precise text spans.

Vela 2.0 brings routing questions and span-level decisions into one model interface. Supply your options, labels and rubrics at request time; ask about the request, its context and the answer together.

The 0.3B is the family's compact multilingual encoder, with an 8,192-token input budget and Torch or ONNX inference on CPU or GPU.

Route with your own criteria. Choose a route, check a policy condition, assign a score or select multiple labels.
Locate the text behind a signal. Return spans of personal information and unsupported claims with character offsets and probabilities.
Ask across the whole interaction. Use typed request, context and answer parts in one request, through Python or a SystemOne-compatible HTTP API.

Model specification	Vela 2.0 0.3B
Backbone	22-layer bidirectional ModernBERT encoder
Hidden width	768
Input budget	8,192 tokens, including schema and state
Questions	Choice, Yes/no (Noul), Score, Span, Set
Inference	Torch fp32; ONNX fp32 or fp16 encoder with fp32 heads
Deployment	CPU or GPU; Python API and HTTP server

Quickstart

Ask for personal information, unsupported answer text and a route in one request:

For focused examples, see Span questions and Set questions below.

python

from transformers import AutoModel
m = AutoModel.from_pretrained("vllm-sr/Vela-2.0-0.3B", trust_remote_code=True)# Runs on CPU by default; optionally use m = m.to("cuda"). Keep Torch loading in fp32.PII_LABELS = m.vela2_engine.cal["pii_schema"]["labels"]
result = m.system_one(    state={"request": "Hi, I'm Tom Baker (tom.baker@example.com). What is the maximum daily dose of paracetamol for an adult?",           "source": "For adults, the maximum dose of paracetamol is 4 grams in 24 hours, taken as 500 mg to 1 g every 4 to 6 hours.",           "answer": "Adults can take up to 6 grams of paracetamol in 24 hours, in doses of 500 mg to 1 g every 4 to 6 hours."},    questions={        "pii": {"type": "span", "instructions": "Which spans are personal information?", "criteria": PII_LABELS,                "over": "request"},        "halu": {"type": "span", "instructions": "Which spans of the answer are not supported by the context?",                 "criteria": {"unsupported": "a claim not supported by the context"}},   # over the answer by default        "domain": {"type": "choice", "instructions": "Which subject area is this request about?", "over": "request",                   "criteria": {"health": "medicine, clinical practice, nutrition, ageing or sexual health",                                "math": "arithmetic, algebra, geometry, statistics or other mathematics",                                "other": "a subject that fits none of the listed areas"}},    })

Selected fields from the recorded Torch CPU response, with probabilities rounded to three decimals:

json

{  "answers": {    "pii": {"type": "noul", "noul": 1.0},    "halu": {"type": "noul", "noul": 0.984},    "domain": {"type": "choice", "choice": "health", "confidence": 0.51, "probabilities": {"health": 0.727, "math": 0.217, "other": 0.056}}  },  "spans": {    "pii": [{"label": "PERSON", "start": 8, "end": 17, "text": "Tom Baker", "probability": 1.0}, {"label": "EMAIL_ADDRESS", "start": 19, "end": 40, "text": "tom.baker@example.com", "probability": 0.983}],    "halu": [{"label": "unsupported", "start": 16, "end": 29, "text": "up to 6 grams", "probability": 0.895}]  }}

Offsets are Unicode code-point offsets into the targeted field. Full response, HTTP, SDK and ONNX examples [blocked].

Span questions: locate text

Use "type": "span" to extract text with labels and character offsets. Set criteria to a label-to-description mapping and over to the state field to inspect. This PII example reuses m and the trained PII_LABELS from above:

python

span_result = m.system_one(    state={"request": "Hi, I'm Tom Baker (tom.baker@example.com)."},    questions={        "pii": {            "type": "span",            "instructions": "Which spans are personal information?",            "criteria": PII_LABELS,            "over": "request",        }    },)print(span_result["spans"]["pii"])

Each returned span includes label, text, start, end and probability. Offsets use Unicode code points, with start inclusive and end exclusive, in the selected field. The first example also shows hallucination detection: use the unsupported label, target over="answer", and supply grounding text in state["source"].

Set questions: select multiple labels

Use "type": "set" when zero, one or several labels can apply. Supply your labels and descriptions in criteria; each label is scored independently. This example reuses m from above:

python

set_result = m.system_one(    state="My card was charged twice and the parcel never arrived.",    questions={        "issues": {            "type": "set",            "instructions": "Which issues does the customer report?",            "criteria": {                "billing": "payments, charges, refunds or invoices",                "shipping": "delivery of an order or a parcel",                "login": "signing in, passwords or account access",            },        }    },)print(set_result["sets"]["issues"]["selected"])print(set_result["sets"]["issues"]["probabilities"])

Read the selected labels and all per-label probabilities from sets["issues"]. Probabilities do not need to sum to one. Labels are selected when their probability exceeds the question's threshold; omit threshold to use shipped calibration, or add it to the question to set your own cutoff. The applied value is returned in set_result["thresholds"]["issues"].

Questions and outputs

Question	Use it for	Output
Choice	Route a request or pick one of 2–255 supplied options.	Selected key and distribution
Yes/no (Noul)	Check a condition against the state.	P(yes)
Score	Rate against 2–10 ordered levels.	Expected level and distribution
Span	Locate personal information or claims unsupported by the source.	Labels, text, offsets and probabilities
Set	Select any number of supplied labels.	Selected labels and per-label probabilities

All five types share one interface. Span and Set preserve SystemOne compatibility through additional response fields and Noul views. Request and response reference [blocked].

Results

Four sizes, one family: 0.3B, 0.8B, 4B, 9B

Task	Vela 2.0 0.3B	Reference
Router safety, macro AUC over 14 sets	0.871	GLiNER2.5-Decide 0.704
Prompt attacks, unseen families, AUC	0.882	Vela 1.0 Guard 0.792
Short PII, exact micro-F1	0.995	Vela 1.0 PII 0.976

Selection: checkpoint selection used dev splits; final release selection also considered test results.
Safety: these task families were part of Vela's training, unlike Decide's. This comparison measures router safety performance, rather than zero-shot generalisation.

Selected results above use identical rows for each task comparison. Full results and evaluation protocol [blocked] include every benchmark, confidence intervals and the separate shipped, test-blind and research PII calibration paths.

Architecture

A 22-layer bidirectional ModernBERT encoder with schema-conditioned decision and word–label span readouts. The first attention pre-norm is Identity; this diagram shows the SDPA path. Editable SVG [blocked].

Operator and readout diagrams

Attention and GEGLU

The SDPA attention path uses 12 heads of width 64 and applies ×4 YaRN RoPE to Q/K. GEGLU has two 1,152-wide branches. Editable SVG [blocked].

Decision and span readouts

Vela 2.0 0.3B: option cosine plus MLP and word-label cosine readouts

The decision and span readouts use 256-dimensional projections and shared readout LayerNorm parameters. Learned τ and inference calibration temperatures T are distinct. Editable SVG [blocked].

The backbone reads schema and typed state bidirectionally. Decision heads combine option cosine scores with an option MLP; the span head compares word and label states. Model, calibration and training details [blocked].

Reference

Usage [blocked]: Python, HTTP, SDK, Span/Set, typed parts, shortcuts and ONNX.
Evaluation [blocked]: full tables, scoring paths, references and selection disclosures.
Training and model details [blocked]: data, training recipe, input schema, calibration and repository files.
Export parity [blocked]: Torch/research-scorer and ONNX comparisons on the measured rows.

Credit and citation

Vela 2.0 is led by KR Labs and vLLM Semantic Router.

Read the Vela 2.0 technical overview.

bibtex

@misc{vela2_unified_2026,  title        = {Vela 2.0: Towards Open Foundation Routing Models},  author       = {{KR Labs} and {vLLM Semantic Router}},  year         = {2026},  note         = {Blog post. Model: vllm-sr/Vela-2.0-0.3B},  howpublished = {\url{https://vllm-sr.ai/blog/vela-2-0-open-foundation-routing-models/}}}

Licence

Apache-2.0 for this model's weights, code and documentation. It is derived from Decision-1.0-Kai-0.6B (Apache-2.0) and Vela-1.0-Encoder-307M (MIT, from mmBERT); their licences and notices are passed on in LICENSE, NOTICE, NOTICE_SOURCES.json and LICENSES/. The tokenizer carries the Gemma Terms of Use: DISTRIBUTION_TERMS.md [blocked] · License scope [blocked]. Changes against Kai and Vela, including the eight added tokenizer entries: MODIFICATIONS.md [blocked]. Training data keep their own licences, some of them CC-BY-SA share-alike (see NOTICE_SOURCES.json).

Explore the Vela model collection