Pre-launch · design-partner pilots open for 2027Synthetic defects · edge inspection · NVIDIA stackApply for the pilot

Coreon · training and models

One defect model, adapted to each SKU in hours

Coreon fine-tunes a lightweight adapter for each part on top of a shared defect model, aligns it with a small set of real frames, and versions every model it ships.

Specification

Training
NVIDIA TAO Toolkit on DGX/HGX or cloud GPUs; DALI data loading
Tasks
Detection, segmentation, classification, anomaly and novelty
Inputs
Synvex datasets; real good and known-bad frames
Outputs
TensorRT-ready adapters for Veridex, with an evaluation report
Capabilities

What Coreon does

  1. 01

    Shared defect backbone

    Each SKU starts from a model that has seen many parts and defect types (foundation model planned).

    Planned
  2. 02

    Per-SKU adapters

    Lightweight adapters fine-tuned with NVIDIA TAO, so a new part takes hours rather than a fresh training run.

    Core
  3. 03

    Sim-to-real adaptation

    Style and lighting alignment closes the gap between rendered and real frames.

    Core
  4. 04

    Detection to anomaly

    Detection, segmentation, classification and an anomaly head for defects nobody has named yet.

    Core
  5. 05

    Versions and rollback

    Every model is versioned per SKU and station, with evaluation attached and one-step rollback.

    Core
  6. 06

    Learns from uncertainty

    Uncertain real frames are queued for review and folded into the next adapter.

    Core
Example output

What Coreon produces

Coreon adapter v7Example

AX-204 evaluation on real frames

Base model
Defect backbone (planned)
Adapter
AX-204 v7
Training data
Synvex run 18 + 120 real
Heads
Detect, segment, anomaly
Registry
Staged for Line A
Scratch0.96
Dent0.93
Burr0.88
Contamination0.91
Example. Per-class recall values are illustrative, not measured results.
How it works

Inside Coreon

  1. ST-01

    Pretrain

    The backbone learns from synthetic and consented real data.

  2. ST-02

    Adapt

    TAO fine-tunes a per-SKU adapter on Synvex data.

  3. ST-03

    Align

    A few real frames close the sim-to-real gap.

  4. ST-04

    Evaluate

    The harness reports per-class accuracy on real frames.

Specification

Specification

Items marked planned are in design with pilot partners.

ItemDetail
TrainingNVIDIA TAO Toolkit on DGX/HGX or cloud GPUs; DALI data loading
TasksDetection, segmentation, classification, anomaly and novelty
InputsSynvex datasets; real good and known-bad frames
OutputsTensorRT-ready adapters for Veridex, with an evaluation report
RegistryPer-SKU and per-station versions, rollback, approvals
Foundation modelCross-customer pretraining with consent (planned)
Questions

About Coreon

Is our data used to train other customers' models?

Not without written consent. Your images and adapters stay yours. The shared backbone only learns from data customers opt in to share.

What accuracy should we expect?

We will not quote numbers before pilots. Each pilot reports per-class accuracy measured on your real parts.

Can we export the model?

Adapters are exported as TensorRT engines for Veridex. Other export formats can be discussed for pilots.

Next step

Try Coreon on one SKU

Pilots start with one part, one defect list and accuracy measured on real frames.