Q4Q4 post-train Resnet-50 cannot get accuracy 73.3.

[haichaoyu]

Hello,

Thanks for the great codebase. I tried Experiment W4A4 + ACIQ + Bit Alloc(A) + Bit Alloc(W) + Bias correction by python inference/inference_sim.py -a resnet50 -b 512 -pcq_w -pcq_a -sh --qtype int4 -qw int4 -c laplace -baa -baw -bcw but only get accuracy 71.6. Is there any particular setting that should be followed? Thanks.

My environment: Pytorch 1.3 Torchvision 0.4.2

[haichaoyu]

Hi @ynahshan, any advice? Thanks.

[nfrumkin]

I got the same results here. Reproducing using explicit commands: 4A4W naive: 63.68% 4A4W ACIQ + Bias Correction + BitAlloc (weights + acts): 71.72%