If training quantification supports in pytorch now?

In tensorflow, there is a training quantificationtensorflow_quantize
So I wonder if anyone has any ideas to support training quantification in pytorch, or has samples?
Now it is quite hard for me to add op(such as quant weight) or modify model struct(such as fold conv and norm op), So if anyone can help me?

Quantization is coming to PyTorch. It won’t be long before it is fully operational, but you can already try it out on master or a nightly release.