---
license: apache-2.0
library_name: rfdetr.cpp
tags:
  - object-detection
  - rfdetr
  - gguf
  - ggml
  - cpp-inference
  - image-segmentation
  - instance-segmentation
pipeline_tag: image-segmentation
base_model: roboflow/rfdetr
---

# RF-DETR Seg-Nano — GGUF for rfdetr.cpp

GGUF-format weights of [Roboflow RF-DETR Seg-Nano](https://github.com/roboflow/rf-detr) (segmentation variant) for use with [rfdetr.cpp](https://github.com/mudler/rf-detr.cpp), a C++/ggml implementation that matches the upstream PyTorch model on CPU.

This repo contains all four standard quantizations of this variant. **F16 is the recommended default** — same accuracy as F32, 1.85× smaller, and typically the fastest on modern CPUs thanks to ggml's F32×F16 matmul fast path.

## Available files

| File | Quant | Size (MB) | Recall @ IoU 0.5 | Recall @ IoU 0.95 | Mean mask IoU | Pixel agreement | Latency (median ms, T=8) |
|---|---|---:|---:|---:|---:|---:|---:|
| `rfdetr-seg-nano-f32.gguf` | F32 | 127.1 | 0.9553 | 0.9553 | 0.9913 | 0.9998 | 114.7 |
| `rfdetr-seg-nano-f16.gguf` ← **recommended** | F16 | 67.8 | 0.9267 | 0.9267 | 0.9911 | 0.9998 | 108.6 |
| `rfdetr-seg-nano-q8_0.gguf` | Q8_0 | 39.9 | 0.9553 | 0.9553 | 0.9901 | 0.9998 | 119.0 |
| `rfdetr-seg-nano-q4_K.gguf` | Q4_K | 31.8 | 0.8949 | 0.6126 | 0.9636 | 0.9990 | 151.6 |

All accuracy numbers are computed against the upstream PyTorch reference (`rfdetr 1.7.0`) on 7 COCO val2017 images at threshold 0.5. Latency is measured with `rfdetr-cli bench` (8 iters + 3 warmup) at T=8 threads on a single AMD Ryzen 9 9950X3D image (`coco_kitchen.jpg`, 640x427).

## Architecture

- Backbone: DINOv2-small
- Input resolution: 312×312
- Patch size: 12
- Decoder layers: 4
- Object queries: 100
- Task: instance segmentation (boxes + per-query masks)
- Mask resolution: 78×78 per query (image_size / 4)

## Quantization notes

- **F32** — full-precision reference, ~120 MB. Bit-exact PyTorch parity.
- **F16** — matmul-multiplicand weights only; LayerNorms, conv kernels, embeddings, biases, and layer-scale gammas stay F32. Lossless on this model and consistently the fastest variant on CPU.
- **Q8_0** — best size/accuracy tradeoff under F16; ~3× smaller than F32 with effectively identical detections.
- **Q4_K** — smallest practical quant. Rows with `ne[0] % 256 != 0` (the decoder's 128-dim MLP halves, 60 tensors) silently fall back to Q8_0 per ggml's quantizer logic — net compression is still ~3.8× over F32. Use only when the size budget is tight; expect a measurable Recall@0.95 drop relative to F16/Q8_0 (see file table above).

## Usage

```bash
# 1. Clone + build rfdetr.cpp
git clone https://github.com/mudler/rf-detr.cpp
cd rt-detr.cpp
cmake -B build -DRFDETR_BUILD_CLI=ON && cmake --build build -j

# 2. Download a quant (F16 recommended)
hf download mudler/rfdetr-cpp-seg-nano rfdetr-seg-nano-f16.gguf --local-dir models/

# 3. Run segmentation (writes per-detection PNG masks to /tmp/seg_masks/)
build/bin/rfdetr-cli detect \
    --model models/rfdetr-seg-nano-f16.gguf \
    --input my_image.jpg \
    --threshold 0.5 --threads 8 \
    --masks /tmp/seg_masks \
    --output detections.json
```

## Accuracy methodology

All accuracy metrics are computed against the upstream PyTorch reference (rfdetr 1.7.0) on 7 COCO val2017 images at threshold 0.5. Each detection match uses greedy Hungarian-style assignment by IoU (≥ 0.5 lenient, ≥ 0.95 strict) with class equality required.

Mask metrics are pixel-wise IoU between binary masks at the **original** image resolution (not the network's working resolution), after sigmoid + bicubic upsample of the per-query mask logits. **Pixel agreement** is the fraction of pixels where the C++ and PyTorch binary masks match.

See [BENCHMARK.md](https://github.com/mudler/rf-detr.cpp/blob/main/BENCHMARK.md) and [`benchmarks/results/accuracy_sweep.json`](https://github.com/mudler/rf-detr.cpp/blob/main/benchmarks/results/accuracy_sweep.json) for the full sweep across all 32 (variant × quant) cells.

## License

Apache-2.0 — matches the upstream [rfdetr](https://github.com/roboflow/rf-detr) license.