Hi,
I am optimizing my model, with a npy file with 2058 samples. I want to use only 2048 of them. Using the command model_optimization_config(calibration, batch_size=32, calibset_size=64) in my alls file results in a warning about the batch_size being greater than the calibset_size, but it is not?
Moreover, the documentation mentions that it is a replacement for the Quantization API, but here, it says that the number of samples is calib_num_batch x batch_size, which the model script is a replacement of.
Am I missing something?
Thank you!