Tutorial — Multi-modal AI

Detect objects with RF-DETR

RF-DETR is a permissively-licensed, NMS-free DETR family covering 90 COCO classes. The detection runtime also serves D-FINE for the smaller 80-class set.

Level: Beginner
Time: ~10 min
Prerequisites: Tenzro CLI installed, sample image
Stack: CLI · JSON-RPC

Load the detection model

The catalog has RF-DETR in six sizes (nano, small, medium, base, large, 2xl) — small is a good starting point. Passing catalog_id inherits the input resolution, class count and decoder ABI from the catalog entry, and applies its license tier. path is where the ONNX graph already sits on the node.

tenzro detect catalog

tenzro detect load \
  --model det \
  --path /models/rf-detr-small.onnx \
  --catalog-id rf-detr-small

Run detection on an image

The CLI returns xyxy boxes in pixel coordinates with class labels and post-sigmoid scores.

tenzro detect run \
  --model det \
  --image image.jpg \
  --score-threshold 0.3

Switch to D-FINE for closed-class COCO

D-FINE returns post-sigmoid sorted boxes already in pixel space — fewer client-side steps.

tenzro detect load \
  --model dfine \
  --path /models/d-fine-s.onnx \
  --catalog-id d-fine-s

tenzro detect run --model dfine --image image.jpg --score-threshold 0.4

Call from JSON-RPC

The RPC returns the same shape: an array of {bbox, label_id, score}.

curl -s https://rpc.tenzro.xyz -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tenzro_detect","params":{"model_id":"det","image_base64":"…","score_threshold":0.3}}'

Open the docs →Browse all tutorials

← All tutorials