raymon-tian

I put fake quantizers into arbitrary locations via pattern-based quantization, however, which has been removed in the last NNCF. I would be appreciated it if nncf could provide some flexible...

@ailias No, I didn't meet the problem. Did you set net.eval() before evaluation？

是的，一个工程demo，主要用sparkMLlib跑了跑一些简单模型

python2

@xhzcyc What is the meaning of the "my final result". Ok, I will spend a few days to check the code, and I will upload the trained model if it...

@xiangweifeng In my impression, the mask_all isn't used. For simplicity, both aren't used in my code.

oh...I am so sorry. The trained weights not work now, so I didn't upload it. I would check the code and find out the reason.

@qixianbiao The result in the demo.ipynb is based on the original caffemodel. And the keras model I trained for 5 days doesn't work, I don't upload it. Bugs may be...