lucid-dl/detr-resnet-101

DETR (ResNet-101)

Carion et al., 2020 — End-to-End Object Detection with Transformers (arXiv:2005.12872)

Lucid port of facebookresearch/detr/detr_resnet101, converted to Lucid-native safetensors.

Available weights

Tag	box mAP	Params	GFLOPs	Size	Source
COCO_2017 (default)	43.5	60.5M	—	231.53 MB	facebookresearch

Usage

import lucid.models as models
from lucid.models.weights import DETRResNet101Weights

# default tag
model = models.detr_resnet101(pretrained=True)

# explicit tag (enum or string)
model = models.detr_resnet101(weights=DETRResNet101Weights.COCO_2017)
model = models.detr_resnet101(pretrained="COCO_2017")

# preprocessing travels with the weights
weights = DETRResNet101Weights.COCO_2017
preprocess = weights.transforms()
out = model(preprocess(image)[None])
# ObjectDetectionOutput: per-query/proposal class logits + boxes
logits, boxes = out.logits, out.pred_boxes

Conversion

Converted from facebookresearch/detr/detr_resnet101 via python -m tools.convert_weights detr_resnet101 --tag COCO_2017. Key mapping + numerical parity verified against the source.

License

apache-2.0 — inherited from the original weights.

Citation

@inproceedings{carion2020detr,
  title={End-to-End Object Detection with Transformers},
  author={Carion, Nicolas and Massa, Francisco and Synnaeve, Gabriel and Usunier, Nicolas and Kirillov, Alexander and Zagoruyko, Sergey},
  booktitle={European Conference on Computer Vision (ECCV)},
  year={2020}
}