π ADD AI Models Leaderboard
ADDEval evaluates AI models developed in ADD on a variety of datasets.
π Notes
-
Models are evaluated on the COCO dataset using standard object detection metrics:mAP@50 and mAP@50:95.
-
FPS (Frames Per Second) is measured on a single NVIDIA A100 GPU using batch size 1.
-
mAP scores reflect model accuracy, while FPS reflects inference speed.
-
"Size" refers to the model's weight file size in megabytes (MB), as used during inference.
-
For fairness, all models are evaluated with their default inference resolution and no additional post-processing.
-
Models with π links redirect to their official GitHub or Hugging Face pages.