yolof

Use Case and High-Level Description

YOLOF is a simple, fast, and efficient object detector without FPN. Model based on “You Only Look One-level Feature” paper. It was implemented in PyTorch* framework. Model used DarkNet-53 with Cross Stage Partial blocks as backbone. For details see repository. This model was pre-trained on Common Objects in Context (COCO) dataset with 80 classes. Mapping of class IDs to label names provided in <omz_dir>/data/dataset_classes/coco_80cl.txt file.

Specification

Metric

Value

Type

Detection

GFLOPs

175.37942

MParams

48.228

Source framework

PyTorch*

Accuracy

Accuracy metrics obtained on Common Objects in Context (COCO) validation dataset for converted model.

Metric

Value

mAP

60.69%

COCO mAP (0.5)

66.23%

COCO mAP (0.5:0.05:0.95)

43.63%

Input

Original model

Image, name - image_input, shape - 1, 3, 608, 608, format is B, C, H, W, where:

  • B - batch size

  • C - channel

  • H - height

  • W - width

Channel order is BGR. Mean values - [103.53, 116.28, 123.675].

Converted model

Image, name - image_input, shape - 1, 3, 608, 608, format is B, C, H, W, where:

  • B - batch size

  • C - channel

  • H - height

  • W - width

Channel order is BGR.

Output

Original model

The list of instances. The postprocessing is implemented inside the model and is performed while inference of model. So each instance it is a object with next fields:

  • detection box

  • label - predicted class ID

  • score - onfidence for the predicted class

Detection box has format [x_min, y_min, x_max, y_max], where:

  • (x_min, y_min) - coordinates of the top left bounding box corner

  • (x_max, y_max) - coordinates of the bottom right bounding box corner

Converted model

The array of detection summary info, name - boxes, shape - 1, 504, 38, 38. The anchor values are 16,16, 32,32, 64,64, 128,128, 256,256, 512,512.

For each case format is B, N\*84, Cx, Cy, where

  • B - batch size

  • Cx, Cy - cell index

  • N - number of detection boxes for cell

Detection box has format [x, y, h, w, class_id_1, …, class_id_80], where:

  • (x, y) - raw coordinates of box center, multiply by corresponding anchors to get relative to the cell coordinates

  • h, w - raw height and width of box, apply exponential function and multiply by corresponding anchors to get absolute height and width values

  • class_id_1,…, class_id_80 - probability distribution over the classes in logits format, apply sigmoid function to get confidence of each class

Download a Model and Convert it into OpenVINO™ IR Format

You can download models and if necessary convert them into OpenVINO™ IR format using the Model Downloader and other automation tools as shown in the examples below.

An example of using the Model Downloader:

omz_downloader --name <model_name>

An example of using the Model Converter:

omz_converter --name <model_name>

Demo usage

The model can be used in the following demos provided by the Open Model Zoo to show its capabilities: