What is the difference between calling torch.ao.quantize and torch.quantize

quantization
1

I am new to PyTorch Quantization some documentation use torch.ao.quantize, and another uses torch. quantization.

2

please use torch.ao.quantization.quantize since torch.quantization.quantize is going to be deprecated

3

What did pytorch quantization become deprecated? And, is the torch.ao is the replacement for PyTorch.quantization?

4

yes, torch.quantization is going to be replaced by torch.ao.quantization migration is in progress, we’ll make an announcement when it is done

5

What does ao mean in torch.ao ?

6

it means architecture optimization, it will include quantization, sparsity and pruning and other ao techniques

3 Likes