Statements (58)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:cosmic_ray_detector
|
| gptkbp:application |
gptkb:Autonomous_Vehicles
gptkb:robotics Surveillance Systems |
| gptkbp:architecture |
19 layers
|
| gptkbp:based_on |
gptkb:YOLOv3
|
| gptkbp:community_support |
Active open-source community
|
| gptkbp:designed_for |
Real-time Object Detection
|
| gptkbp:developed_by |
gptkb:Joseph_Redmon
|
| gptkbp:evaluates |
Frames Per Second (FPS)
Mean Average Precision (m AP) |
| gptkbp:feature |
Low latency
Data augmentation Real-time processing Model compression techniques Real-time video processing Batch normalization Skip connections Visualization tools available Easy to deploy Lightweight model Anchor boxes Multi-scale detection Open CV integration Class prediction Export to ONNX format Feature pyramid networks Keras compatibility Localization accuracy Low computational cost Multiple object detection Non-max suppression Pre-trained weights available Pruning support Py Torch compatibility Quantization support Tensor Flow compatibility Compatible with CPUs Compatible with GPUs |
| https://www.w3.org/2000/01/rdf-schema#label |
Tiny YOLOv3-19l
|
| gptkbp:input_output |
416x416 pixels
Bounding boxes and class probabilities |
| gptkbp:is_a_framework_for |
gptkb:Darknet
|
| gptkbp:license |
GPL-3.0
|
| gptkbp:performance |
High precision and recall
Faster than YOLOv3 |
| gptkbp:platform |
Cross-platform compatibility
|
| gptkbp:provides_information_on |
gptkb:COCO_dataset
|
| gptkbp:related_model |
gptkb:YOLOv2
gptkb:YOLOv3 Tiny YOLOv2 |
| gptkbp:release_date |
gptkb:2018
|
| gptkbp:successor |
gptkb:YOLOv4
Tiny YOLOv4 |
| gptkbp:training |
gptkb:stage_adaptation
|
| gptkbp:uses |
gptkb:neural_networks
|
| gptkbp:bfsParent |
gptkb:Tiny_YOLOv3
|
| gptkbp:bfsLayer |
7
|