OpenPPL / ppq

PPL Quantization Tool (PPQ) is a powerful offline neural network quantization tool.
Apache License 2.0
1.57k stars 236 forks source link

dispatch.py run error #487

Closed KKWY0909 closed 1 year ago

KKWY0909 commented 1 year ago

The code in the sample dispatch.py runs to register_network_quantizer(MyQuantizer, platform=TargetPlatform.ACADEMIC_INT4) with an error. But it runs successfully after changing the platform to PPL_CUDA_INT4. What is the reason for this?

ZhangZhiPku commented 1 year ago

TargetPlatform.ACADEMIC_INT4 is a legacy platform, i will remove it from dispatch.py soon.