_normalize_distr function bug in entropy calibrator

[PT0X0E]

Description

In pytorch-quantization package, the _normalize_distr(distr) function in entropy calibrator is not right.

def _normalize_distr(distr):
    summ = np.sum(distr)
    if summ != 0:
        distr = distr / summ

This won't actually normalize the input distr as expected. The distr = is a new array. To achieve in-place, it should be distr /= summ.

The reason this calibrator still works is that in scipy's entropy function, scipy also performs normalization. But we better fix it here in case scipy changes its implementation.

[lix19937]

see #3752

[lix19937]

BTW, Pytorch Quantization no longer receive further development.