iamkanghyunchoi / ait

It's All In the Teacher: Zero-Shot Quantization Brought Closer to the Teacher [CVPR 2022 Oral]
GNU General Public License v3.0
30 stars 5 forks source link

Comparison with previous data-free quantization methods #3

Closed DCNSW closed 2 years ago

DCNSW commented 2 years ago

Thanks for your good work on data-free quantization and the release of the code.

Zero-shot adversarial quantization(ZAQ) is also a data-free quantization method with adversarial exploration. I notice that both IntraQ and AIT are not compared to the ZAQ method. Are they solving different problems or having different experiment settings?

Thanks for your answer.

iamkanghyunchoi commented 2 years ago

First of all, I appreciate your interest in our paper.

I believe that zero-shot quantization and data-free quantization are identical terms that aim at neural network quantization without any training dataset. Therefore, AIT, IntraQ, and ZAQ are trying to solve the same problem.

However, ZAQ uses symmetric quantization for weight and activation, while our method utilizes asymmetric quantization for both. Therefore, it is not suitable to directly compare the paper’s result with our experimental result.

In addition, we were not able to reproduce most of the results from the original ZAQ paper.