Model Card
PG-StreetParse-M
An instance segmentation model (RF-DETR Seg Medium), trained on 6 classes, scoring 18.9% mAP.
Pictograph Research · Instance Segmentation · RF-DETR Seg Medium · FP32 · CC-BY-ND-4.0 · Public
Overview
This is an instance segmentation model built on RF-DETR Seg Medium, trained on the A2D2 dataset. It expects 432x432 input and runs in FP32 precision. It predicts car, pedestrian, truck, small vehicle, utility vehicle, bicycle.
Performance
Headline: 18.9% mAP.
| Metric | Value |
|---|---|
| mAP | 18.9% |
| mAP@50 | 36.8% |
Classes
Class index matches the model's output order.
| # | Class |
|---|---|
| 0 | car |
| 1 | pedestrian |
| 2 | truck |
| 3 | small vehicle |
| 4 | utility vehicle |
| 5 | bicycle |
Training
| Setting | Value |
|---|---|
| Base dataset | A2D2 |
| Export | A2D2-base-sample |
| Dataset size | 12,758 images |
| Input size | 432x432 |
| Epochs | 10 |
| Batch size | 4 |
| Learning rate | 0.0002 |
| Model size | medium |
| Precision | FP32 |
| Version | 1.0.0 |
Use this model
Run inference with the Pictograph SDK:
from pictograph import Client, get_model, InstanceSegmentationModel, InstanceSegmentationResult
client = Client(api_key="pk_live_...") # Settings -> API Keys
# Downloads + caches the weights on first use
model: InstanceSegmentationModel = get_model(
name="PG-StreetParse-M",
task="instance_segmentation",
format="onnx",
precision="fp32",
device="cpu",
client=client,
)
result: InstanceSegmentationResult = model.predict(image="photo.jpg")
print(result)
# Fully offline (no API key) once you have the weights + config.json
from pictograph import load_model, InstanceSegmentationModel, InstanceSegmentationResult
model: InstanceSegmentationModel = load_model(
"/path/to/weights.onnx",
"/path/to/config.json",
task="instance_segmentation",
format="onnx",
)
result: InstanceSegmentationResult = model.predict(image="photo.jpg")
- Full inference guide: local inference docs.
- REST + CLI: models API reference.
- Deploy it as an always-on endpoint: deployments.
License
Released under CC-BY-ND-4.0, the license of its training data, A2D2. Commercial use is permitted with attribution: when you use this model, please credit Audi AG (Geyer et al., 2020).
Citation
This model uses the RF-DETR Seg Medium architecture:
@inproceedings{robinson2026rfdetr,
title = {RF-DETR: Real-Time Detection Transformer},
author = {Robinson, Isaac and Robicheaux, Peter and Popov, Matvei and Ramanan, Deva and Peri, Neehar},
booktitle = {International Conference on Learning Representations (ICLR)},
year = {2026},
url = {https://arxiv.org/abs/2511.09554}
}
It was trained on A2D2, from:
@article{geyer2020a2d2,
title = {{A2D2}: {Audi Autonomous Driving Dataset}},
author = {Geyer, Jakob and Kassahun, Yohannes and Mahmudi, Mentar and Ricou, Xavier and Durgesh, Rupesh and Chung, Andrew S. and Hauswald, Lorenz and Pham, Viet Hoang and M{\"u}hlegg, Maximilian and Dorn, Sebastian and Fernandez, Tiffany and J{\"a}nicke, Martin and Mirashi, Sudesh and Savani, Chiragkumar and Sturm, Martin and Vorobiov, Oleksandr and Oelker, Martin and Garreis, Sebastian and Schuberth, Peter},
journal = {arXiv preprint arXiv:2004.06320},
year = {2020}
}Download the model files
Exports to ONNX · 130 MB · FP32. Sign in to download the weights and config, or pull it from the SDK with the Get code snippet.
Sign in to downloadAllows any use, including commercial, with credit, but adapted versions of the material may not be shared.
Read the full Creative Commons Attribution-NoDerivatives 4.0 →Build your first vision tool
$25 of free compute to start. No card required.