Skip to content

Latest commit

 

History

History
 
 

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 

Tiny YOLOv2

Description

This model is a real-time neural network for object detection that detects 20 different classes. It is made up of 9 convolutional layers and 6 max-pooling layers and is a smaller version of the more complex full YOLOv2 network.

Model

Model Checksum Download (with sample test data) ONNX version Opset version
Tiny YOLOv2 MD5 58 MB 1.0 1
MD5 58 MB 1.2 7
MD5 58 MB 1.3 8

Paper

"YOLO9000: Better, Faster, Stronger" arXiv:1612.08242

Dataset

The Tiny YOLO model was trained on the Pascal VOC dataset.

Source

The model was converted from a Core ML version of Tiny YOLO using ONNXMLTools. The source code can be found here. The Core ML model in turn was converted from the original network implemented in Darknet (via intermediate conversion through Keras).

Inference

Input

shape (1x3x416x416)

Preprocessing

Output

shape (1x125x13x13)

Postprocessing

The output is a (125x13x13) tensor where 13x13 is the number of grid cells that the image gets divided into. Each grid cell corresponds to 125 channels, made up of the 5 bounding boxes predicted by the grid cell and the 25 data elements that describe each bounding box (5x25=125). For more information on how to derive the final bounding boxes and their corresponding confidence scores, refer to this post.

Sample test data

Sets of sample input and output files are provided in

  • serialized protobuf TensorProtos (.pb), which are stored in the folders test_data_set_*/.

License

MIT