Efficient Post-Training Quantization for Referring Image Segmentation Models
An effective and efficient post-training quantization framework, PTQ4RIS, is proposed to enable the deployment of large multi-modal referring image segmentation models on resource-constrained edge devices.