Faster Rcnn Resnet 50 Fpn

--- library_name: lucid license: bsd-3-clause tags: - object-detection - faster - lucid datasets: - coco pipeline_tag: object-detection model-index: - name: faster-rcnn-resnet50-fpn results: - task: { type: object-detection } dataset: { name: COCO, type: coco } metrics: - { type: box mAP, value: 37.0 } --- # Faster R-CNN (ResNet-50-FPN) > Ren et al., 2015 — *Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks* (arXiv:1506.01497) [Lucid](https://github.com/ChanLumerico/lucid) port of `torchvision/FasterRCNN_ResNet50_FPN_Weights.COCO_V1`, converted to Lucid-native safetensors. ## Available weights | Tag | box mAP | Params | GFLOPs | Size | Source | |---|---|---|---|---|---| | `COCO_V1` *(default)* | 37.0 | 41.8M | 134.38 | 159.72 MB | torchvision | ## Usage ```python import lucid.models as models from lucid.models.weights import FasterRCNNResNet50FPNWeights # default tag model = models.faster_rcnn_resnet50_fpn(pretrained=True) # explicit tag (enum or string) model = models.faster_rcnn_resnet50_fpn(weights=FasterRCNNResNet50FPNWeights.COCO_V1) model = models.faster_rcnn_resnet50_fpn(pretrained="COCO_V1") # preprocessing travels with the weights weights = FasterRCNNResNet50FPNWeights.COCO_V1 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 `torchvision/FasterRCNN_ResNet50_FPN_Weights.COCO_V1` via `python -m tools.convert_weights faster_rcnn_resnet50_fpn --tag COCO_V1`. Key mapping + numerical parity verified against the source. ## License `bsd-3-clause` — inherited from the original weights. ## Citation ``` @inproceedings{ren2015faster, title={Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks}, author={Ren, Shaoqing and He, Kaiming and Girshick, Ross and Sun, Jian}, booktitle={Advances in Neural Information Processing Systems (NeurIPS)}, year={2015} } ```