Efficient Continual Calibration of Quantized Classification Models for Edge Devices

A framework called QCore that enables efficient continual calibration of quantized classification models on resource-limited edge devices without requiring full training data or expensive back-propagation.