OpenPPL / ppq

PPL Quantization Tool (PPQ) is a powerful offline neural network quantization tool.
Apache License 2.0
1.56k stars 236 forks source link

在读取pt模型为gragh时,会先将模型转换为onnx,再读取onnx为gragh,会对gragh进行格式化,其中会将constant转换为parameter,但若constant的value为None,导致转换为parameter后出现再次读取gragh时出现空值报错,该如何处理该问题 #501

Open KuaiQuXueX1 opened 1 year ago

ZhangZhiPku commented 1 year ago

问题将在最新一次更新中解决