RF-DETR Small — MLX, fp32
MLX-format conversion of Roboflow's RF-DETR Small object-detection model, for use with mlx-swift-rf-detr on Apple Silicon.
- Upstream: roboflow/rf-detr (Apache-2.0), checkpoint
rf-detr-small.pth - Task: object detection — COCO, 90 classes
- Resolution: 512 · decoder layers: 3 · queries: 300 · backbone:
dinov2_windowed_small(hidden 256) - Precision: fp32
- Files:
config.json+preprocessor_config.json+model.safetensors
Conversion
Converted from the upstream PyTorch checkpoint with
Scripts/convert_detection.py small.
Config values are read from the upstream RFDETRSmallConfig; weights keep their model. prefix and
NCHW conv layout (the Swift loader transposes conv weights to NHWC at load).
Usage (Swift)
import MLXRFDETR
let dir = URL(fileURLWithPath: "rfdetr-small-mlx-fp32") // this repo's downloaded files
let predictor = try MLXRFDETR.fromPretrained(dir, scoreThreshold: 0.3, nmsThreshold: 0.5)
let result = try predictor.predict(url: URL(fileURLWithPath: "image.jpg"))
License
Apache 2.0 — see LICENSE. Model architecture and weights © Roboflow, Inc. This
repository redistributes a format-converted copy of the upstream Apache-2.0 weights; no weights were
retrained or modified beyond the PyTorch→MLX serialization.