PyTorch Object Detection :: YOLOv7 TXT

YOLOv7

YOLOv7 was released in July 2020 by WongKinYiu and AlexeyAB. It achieves state of the art performance on and are trained to detect the generic 80 classes in the MS COCO dataset for real-time object detection.

YOLOv7 inference on image of horses

There are six versions of the model ranging from the namesake YOLOv7 (fastest, smallest, and least accurate) to the beefy YOLOv7-E6E (slowest, largest, and most accurate).

The differences between the different sizes of the model are:

  • The image input resolution
  • The number of anchors
  • The number of parameters
  • The number of layers

Compare YOLOv7 vs Other Models on COCO

The evaluation of YOLOv7 models show that they infer faster (x-axis) and with greater accuracy (y-axis) than comparable realtime object detection models. YOLOv7 evaluates in the upper left - faster and more accurate than its peer networks.

Evolution of layer aggregation strategies in YOLOv7