Scientific utility
Release material when access can support reproducibility, evaluation or useful independent research.
Open research
Open where it advances the field.
Sinuosity shares selected research artifacts where public access can accelerate progress without compromising privacy, rights, safety or evaluation integrity.
Okuafo MaizeGuard Edge v1.4 is a publicly visible research preview. Its model files and reference inference code can be inspected on Hugging Face, but repository visibility is not production approval or a license grant.
Rights and consent
Respect participant, farmer, customer and data-contributor permissions at every release boundary.
Evaluation integrity
Keep appropriate test material private where disclosure would invalidate meaningful evaluation.
Operational safety
Withhold or limit artifacts when unrestricted use could create avoidable physical or privacy risk.
Intended release areas
Research artifacts under consideration.
- Compact FieldMind research checkpoints
- FieldBench evaluation tools
- Edge-inference examples
- Reference hardware integrations
- Evaluation harness releases
- Model and dataset cards
- Training and fine-tuning recipes
These are potential release categories, not a list of available assets.
Release policy
Release requires a complete research record.
A public release should identify its version, status, intended use, limitations, evaluation basis, provenance, license, citation, changelog and responsible disclosure contact.
Private evaluation sets, sensitive field data, personal data and material whose release could undermine safety or rights may remain restricted.
Data governancePublic research register
Accessible research artifacts.
Versioned access, status and limitations
Model Release / 2026
Field vision / calibrated screening
Okuafo MaizeGuard Edge v1.5
A public development candidate for human-reviewed maize symptom screening, combining categorical and specialist FP16 TFLite models through a locked calibrator with explicit abstention.
- System
- Categorical + specialist fusion
- Runtime
- FP16 TensorFlow Lite
- Decision
- Calibrated abstention
- Status
- Public development candidate
Published for user testing with a known gray leaf spot recall limitation. It is not independently or agronomist validated and is not approved for production or autonomous diagnosis.
