Is It Possible to Use a Random Dataset to Generate the INT8 Post-training Quantization Model?

Unable to determine whether random dataset can be input to generate INT8 Post Training Quantization models.

In general, a random input dataset can be used to apply 8-bit post-training quantization. NNCF API has two main capabilities, Basic quantization and Quantization with accuracy control where calibration datasets and validation datasets need to be prepared. The minimum dataset should be 300 samples.

Refer to this Dataset Preparation for Object Detection Sample to prepare the dataset.