Integrated into the neck or head of the network to capture global context without the heavy computational cost of standard transformers.

In these "Lightweight" (LS) models, the following components are typically highlighted in the full papers:

Based on your search for "LS Models (10)", there are several recent publications that match this technical profile:

Reduces parameters and FLOPs while maintaining feature extraction quality.

While based on YOLOv8, this "LS" (Lightweight and Scalable) variant is highly cited for its use of Multi-Scale Ghost Convolution (MSGConv) and efficiency gains of up to 55% FPS. Full Text Access: View the full paper on ResearchGate . Key Technical Features of LS-Models